The booming field of cancer immunotherapy is predicated on the notion that the human body is built to fight off disease—all we need to do is figure out how to unleash it against tumors. But that seemingly simple proposition turns out to be quite complex. Give the immune system too much gas and it can kill its host; don’t ease up on the brakes enough and it sputters out. Cancer wins.
Happily, researchers have made amazing progress in finding the delicate balance between safety and effectiveness while turning immune cells against cancer. Drugs that, by various means, release the brakes on the immune system are one of the more promising avenues being pursued.
Already, three antibodies that block proteins that dampen the immune response have reached the market. These drugs, known as checkpoint inhibitors, are powerfully effective in some people with skin and lung cancer. For advanced melanoma patients, a combination of two checkpoint inhibitors has changed the outlook for two-year survival from just 5% to as much as 79%.
But there’s plenty of room for improvement. The side effects from that antibody cocktail are harsh. And everyone wants to move the needle for survival in other types of cancer.
Recent studies in animals—and early data from human trials—suggest one way to boost the activity of checkpoint inhibitors is to also target an enzyme called indoleamine 2,3-dioxygenase, or IDO1. Researchers have spent years understanding its role in normal immune surveillance and unraveling how cancer cells use it to evade an immune system attack.
Today, a growing number of small-molecule IDO1 inhibitors are in or poised to enter the clinic. Even though much still needs to be learned about how, and whether, these molecules work in humans, big pharma firms have collectively spent nearly $1 billion in the past six months to add IDO1 inhibitors to their immuno-oncology portfolios.
This landgrab has been more than 15 years in the making. The first inkling of IDO1’s potential as a cancer drug target came in 1998 when researchers from Georgia Regents University uncovered its role in immune system regulation.
The immune system regularly encounters all sorts of innocuous antigens that it knows not to attack, be they in the food we eat or in the pollen wafting out of our gardens. And throughout time, women have carried a giant bundle of foreign antigens—otherwise known as a baby—that the immune system tolerates for nine long months.
Companies with IDO1 inhibitors have struck lucrative deals with big pharma firms in the past two years.
Flexus Biosciences: Bristol-Myers Squibb pays $800 million up front and up to $450 million in milestones for IDO1 and TDO2 assets.
iTeos Therapeutics: Pfizer pays $30 million up front, plus equity investment and milestone payments, for IDO1 and TDO2 assets.
NewLink Genetics: Genentech pays $150 million up front, with more than $1 billion in potential milestones, to license NLG919 and develop IDO1/TDO2 molecules.
Incyte: Multiple nonexclusive pacts to study INCB24360 with other checkpoint inhibitors, including the following:
◾ Merck & Co.’s anti-PD1 antibody pembrolizumab in lung and other cancers
◾ AstraZeneca’s anti-PDL1 antibody MEDI4736 in multiple cancers
◾ BMS’s anti-PD1 antibody nivolumab in multiple cancers
◾ Genentech’s anti-PDL1 antibody MPDL3280A in lung cancer
But for decades, immunologists couldn’t fully explain the phenomenon. It couldn’t be that the entire immune system is suppressed—after all, pregnant women can still fight off a cold. They wondered: “How does the immune system decide, ‘Here are these foreign antigens, whether it’s food or pollen or a baby, that aren’t self, and yet you shouldn’t respond to them’?” says David H. Munn, a cancer immunology expert at Georgia Regents.
By studying pregnant mice, Munn and his colleague Andrew L. Mellor helped unravel part of the mystery. It turns out the immune system isn’t just ignoring that foreign tissue. Rather, fetal cells expressing IDO1 signal T cells, important immune cells, to stand off. IDO1 initiates this signal by kicking off the breakdown of tryptophan via the kynurenine pathway, one of the most critical routes for degrading the amino acid. The resulting tryptophan metabolites keep T cells from bearing down on foreign antigens.
In a pregnancy, this immunosuppression is a good thing; in cancer, it’s deadly. “That was a phenomenon we were very interested in as oncologists because our hypothesis was that tumors hijacked those normal processes for creating immune tolerance,” Munn says.
Importantly for drug developers, Munn also showed the effect could be harnessed: When he gave pregnant mice 1-methyltryptophan, an inhibitor of IDO1 and closely related enzymes, the immune system attacked the fetus.
By the early 2000s, researchers at various institutions had shown that IDO1 is expressed at varying levels in many kinds of tumors. They hypothesized that cancer cells cleverly use the enzyme to hide from the immune system.
IDO1 is one of several molecules that keep the immune system from attacking our own antigens or innocuous invaders. Around the time that IDO1 was being discovered, researchers were uncovering other so-called checkpoint proteins, including CTLA-4, PD-1, and PD-L1.
During the immune system’s constant surveillance for foreign invaders, these molecules play a role in putting the brakes on T cells: CTLA-4 and PD-1 are both proteins on T cells that, at different stages, suppress an immune response; PD-L1 is the ligand for PD-1 and is often expressed by cancer cells to turn off the immune response.
But whereasother checkpoint proteins are critical to everyday immune surveillance, IDO1 isn’t turned on in most normal tissues. The hope is that this limited activity will make it a safer drug target. Mice engineered to lack IDO1 don’t have autoimmune problems, explains George C. Prendergast, a molecular biologist who leads the Lankenau Institute for Medical Research. “Tumors love to turn IDO1 on, but they rely on it in a way that other normal tissues don’t seem to,” he says.
Those findings inspired the first efforts to invent drugs to block the enzyme. Although several big drug firms over the years have quietly been looking at the target, until recently, the only companies publicly in the game were the biotechs Incyte and NewLink Genetics.
The first potent and selective IDO1 inhibitor on the industry radar came from Incyte, which started working on it in 2004. The drug discovery campaign was tricky because the company was working without knowing what an ideal inhibitor should look or act like, says Reid M. Huber, Incyte’s chief scientific officer.
Because the drug would be first in class, Incyte scientists had to build a data set to understand the pharmacokinetic and pharmacodynamic properties that would be important for activity. In addition, the biology around the target was not completely understood, so they had to figure out how an IDO1 inhibitor would be useful both on its own and in combination with other emerging immunotherapies.
Few if any tool, or probe, compounds existed against the target, so Incyte had to screen about 300,000 compounds to find a lead with enough activity to merit a medicinal chemistry campaign, recalls Andrew Combs, the biotech’s vice president of discovery chemistry.
Within a year and a half, the company’s chemists had generated about 2,000 compounds based on that lead. Incyte gradually winnowed them down to INCB24360. Incyte’s compound entered the clinic in 2010 and now is in a range of studies that combine it with immuno-oncology treatments being developed by other firms.
Around the time Incyte was starting to develop IDO1 inhibitors, NewLink Genetics signed two deals to enter the field. In 2005, NewLink licensed indoximod, the compound used by Munn and Mellor in their pregnant-mouse studies. That same year, NewLink also bought OncoRx, a biotech firm started by Lankenau’s Prendergast to develop small molecules that block the enzyme.
Indoximod, which is simply the d-isomer of 1-methyltryptophan, has gone through a range of clinical studies, both on its own and in combination with other cancer-fighting compounds. But over time, researchers have come to acknowledge it is not a true IDO1 inhibitor and instead appears to hit multiple enzymes.
Chemists at NewLink, which declined to comment for this story, went on to screen compounds against the isolated enzyme and found a more selective IDO1 inhibitor. The molecule, NLG919, was so appealing that last October Genentech paid $150 million for access to it. The firm has since renamed it GDC919.
Genentech shopped around before picking NewLink’s compound, which it chose because of its potency, selectivity, and good oral bioavailability, says Wendy Young, vice president of small-molecule discovery at Genentech. “It’s a fantastic tool, and we’ve invested quite a bit in going forward with it in combination with other molecules in our portfolio.”
Indeed, very soon after the discovery of IDO1’s role in cancer, researchers knew that blocking it wouldn’t be enough to coax the immune system into attacking cancer. “If you take your foot off that brake and then don’t step on the accelerator, nothing happens. You just create a situation in which something could happen,” Georgia Regents’ Munn explains.
To hit the gas while releasing the brake, drug companies are pairing IDO1 inhibitors with antibodies targeting other checkpoint proteins—some of which are already on the market. In 2001, Bristol-Myers Squibb’s anti-CTLA-4 antibody Yervoy was approved to treat melanoma, and last year, Merck & Co. and BMS both won the Food & Drug Administration’s okay for anti-PD1 antibodies. Now the focus is on combinations that increase their efficacy.
A 2013 study by University of Texas M. D. Anderson Cancer Center immunologist James P. Allison provided a strong rationale for combining IDO1 inhibitors with antibodies targeting CTLA-4, PD-1, or PD-L1. Allison, who discovered Yervoy, was trying to understand why only a subset of patients responds to the melanoma treatment.
His study showed that cancer cells exposed to Yervoy use IDO1 to hide out from the immune system. But when mice with melanoma were given drug combinations that block both CTLA-4 and IDO1, their tumors shrank, and they lived longer. Allison reported similar findings for combinations of IDO1 inhibitors and antibodies targeting PD-1 and PD-L1.
A few months later, University of Chicago professor Thomas F. Gajewski came out with “a mechanistic tour de force” detailing how pairs of inhibitors of CTLA-4, PD-L1, and IDO1 were able to reactivate T cells, says Eric H. Rubin, Merck’s vice president of oncology clinical research. He says the study provided “a nice mechanistic rationale for why you might expect greater response in combination than with either drug alone.”
As the idea of combining IDO1 inhibitors with other immuno-oncology treatments gains validation, several new players have emerged to challenge Incyte and NewLink. The goal has been to improve upon the pharmacokinetic properties of the pioneering compounds, ultimately making them better partners for other checkpoint inhibitors.
Flexus Biosciences, which made its debut in 2013, looked at the public data on Incyte’s INCB24360 and saw an opportunity for improvement. The data suggested that IDO1 needs to be fully blocked to elicit a response. To achieve this, patients must take 600 mg of Incyte’s compound every 12 hours, says Juan Jaen, head of R&D at Flexus. By the time a patient was on the verge of taking his or her next pill, concentrations of the drug in the blood had dropped precipitously.
“That led us to believe we needed a flatter profile—a longer half-life to avoid huge peaks in order to cover valleys,” Jaen says, “but also a molecule that is significantly more potent.”
Flexus chemists believed they should be aiming for molecules 10 times more potent than the Incyte compound. They ended up with a new class of compounds that are anywhere from 20 to 100 times more potent. By improving both the projected half-life and potency in humans, Flexus believes its lead compound can be dosed once daily in amounts as low as 10 to 20 mg.
“These are all models, and it’s like predicting the weather seven weeks from today,” Jaen cautions. “It’s a fairly educated guess, but at the end of the day, data need to be generated.” Still, Flexus’s preclinical findings looked good enough for Bristol-Myers Squibb to agree to pay a whopping $800 million up front for the firm’s lead IDO1 inhibitor and a series of follow-up compounds.
Although BMS won’t discuss the attraction of the Flexus molecule until the deal has closed, “I think you can surmise that we obviously felt it was very good because we spent so much money,” says Carl P. Decicco, BMS’s head of drug discovery.
Decicco notes that BMS scientists also had been searching for potent but safe IDO1 inhibitors that could be administered once a day. “We think low dose is important, and that’s been a primary driver of our internal program,” Decicco says, adding that because an IDO1 inhibitor will be used in combination with other drugs, putting the least strain on the liver is essential.
BMS’s confidence in its purchase comes not just from Flexus’s compounds but from the potential of IDO1 as a drug target. BMS and Incyte have a clinical pact to test INCB24360 with the PD-1 inhibitor Opdivo (nivolumab).
Like Flexus, the Belgian biotech iTeos Therapeutics is benchmarking its IDO1 inhibitor program against Incyte’s compound with an eye on finding a safe molecule that will stick around the active site for 24 hours. The privately held biotech was spun off of the Brussels branch of the Ludwig Institute for Cancer Researchand the de Duve Institute at Belgium’s Catholic University of Louvain in 2012 to develop molecules to block IDO1 and tryptophan 2,3-dioxygenase (TDO2), an enzyme found in the liver and brain that, like IDO1, breaks down tryptophan.
Developing an assay that keeps IDO1, a heme-containing redox enzyme, catalytically active is a challenge, acknowledges Michel Detheux, iTeos’s chief executive officer. “It’s not an easy enzyme to handle,” he says, which might explain why so few screening efforts have yielded good starting points for drug discovery.
iTeos was able to come up with a low-molecular-weight—under 250 daltons—molecule that still manages to be potent and selective, according to Stefano Crosignani, the company’s head of medicinal chemistry.
In December, Pfizer paid nearly $30 million and took an equity investment in iTeos to gain access to the biotech’s preclinical IDO1 and TDO2 compounds. The pact “goes beyond the asset” and is also about tapping the expertise of the iTeos team, notes Valeria Fantin, who leads the tumor cell biology group within Pfizer’s oncology research unit.
While perusing the IDO1 landscape, Pfizer scientists have seen a variety of molecular flavors but believe they still don’t know what properties are the most important for a successful drug. “I still think a big unknown in this area of IDO1 drug discovery is, how well do we even need to cover this target?” Fantin says. “Is this a target that needs to be covered 24/7, and if that is the case, is that type of coverage going to be tolerated? Or maybe 50% is enough.”
The winning profile for an IDO1 inhibitor will emerge only when human testing on the various compounds begins. Researchers also hope to learn the magnitude of the benefit—if any—an IDO1 inhibitor can add to the various checkpoint antibodies.
So far, few human data exist to support the concept. Last year at the annual meeting of the American Society of Clinical Oncology, preliminary results from a study pairing Incyte’s INCB24360 with BMS’s CTLA-4 blocker Yervoy suggested the two drugs improve response compared with Yervoy on its own for melanoma patients whose cancer had metastasized.
But in a less encouraging sign, five out of seven patients given a higher dose of INCB24360 had to stop treatment because of elevated liver enzymes. Indeed, another challenge for companies will be understanding when and how to safely use IDO1 inhibitors.
Unlike a mutation in a tumor that is always present, IDO1 is only expressed in certain circumstances. The enzyme might show up in a biopsy when cancer is first detected, but it also might not appear until after a patient is treated with a checkpoint inhibitor.
iTeos’s Detheux compares figuring out how to harness the immune system with watching a movie rather than looking at a snapshot. “It’s not a fixed situation; it’s something which can evolve over time,” he says. “That is more challenging.”
Answers should start to emerge next year, when data from more clinical studies are revealed. Meanwhile, the number of experiments is swiftly multiplying. In addition to pairing IDO1 inhibitors with antibodies that target CTLA-4, PD-1, and PD-L1, big pharma firms are excited about doing the same with antibodies against other immune regulators, including OX40, LAG3, and TIM-3.
Later this year or early next year, Genentech plans clinical studies of GDC919 combined with its anti-PDL1antibody MPDL3280A or its anti-OX40 antibody MOXR0916. And Incyte, which has chosen not to partner its IDO1 inhibitor, has nonexclusive partnerships to pair its compound with checkpoint inhibitors from AstraZeneca, BMS, Genentech, and Merck. BMS sees the potential for IDO1 inhibitors to be combined with a range of molecules in its pipeline.
When considering what success might look like for IDO1 inhibitors, researchers come back to the astounding results delivered by the first checkpoint inhibitors. Before Yervoy, only about 5% of patients with advanced melanoma were alive two years after their diagnosis. Yervoy increased survival to roughly a quarter of melanoma patients. Combining Yervoy with the PD-1 inhibitor Opdivo brought two-year survival figures up to 79%.
In lung cancer, PD-1 inhibitors have brought two-year survival up to the 40% range. The question now is whether responses will jump to melanoma-like levels with the addition of an IDO1 inhibitor. “Will we get that same effect with IDO?” Decicco asks. “If we get that, I will be dancing in the streets.”