June 20, 2018
Volume 29, No. 119

The news source of record covering the development of innovative human therapies
Actionable Intelligence • Incisive Analysis

Adenosine dreams fueling new approaches in cancer; Arcus embarks on phase I

By Randy Osborne, Staff Writer

Although Merck & Co. Inc. continues to chalk up investor-pleasing sales with Keytruda (pembrolizumab) and Bristol-Myers Squibb Co. (BMS) keeps doing well with Opdivo (nivolumab), researchers have yet to strike upon the best combinations with other agents.

The prospect of using an oral indoleamine 2,3-dioxygenase 1 (IDO1) inhibitor took a hit in April when Incyte Corp. and Kenilworth, N.J.-based Merck reported that the pivotal phase III Echo-301/Keynote-252 study testing Incyte’s epacadostat paired with Keytruda in individuals with unresectable or metastatic melanoma failed a review by the external data monitoring committee (DMC). The experiment missed the primary endpoint of improving progression-free survival in the overall population compared to Keytruda monotherapy and was expected to fall short of statistical significance on the co-primary endpoint of overall survival, prompting the companies – following the DMC’s advice – to stop the study. Merck and Incyte, of Wilmington, Del., said the safety profile turned out consistent with previous studies of the Keytruda/epacadostat combo. (See BioWorld, April 9, 2018.)

BMS “has now followed other players in this field by suspending its phase III programs of the compound BMS-986205 from the same class,” Leerink Partners LLC analyst Geoffrey Porges noted in a May 2 report, adding that the firm “appears to be retreating from the previous aggressive development plan, and taking a more cautious look at the data from the remaining early stage studies.” IDO1 stumbling blocks “could also have broad implications for other immuno-oncology (I-O) agents under development that do not have single agent activity and are built on the promise of potential synergy with checkpoint blockade,” he said. The recent annual meeting of the American Society of Clinical Oncology (ASCO) in Chicago included research aplenty on how to boost I-O therapies, especially in the PD-1 class.

“Given the challenges of this endeavor, the meeting was disappointingly marked by more failures and post-mortems, than by new breakthroughs,” he said.

This might be where Hayward, Calif.-based Arcus Biosciences Inc. comes in. The company’s approach involves inhibiting the adenosine pathway, and the FDA this month cleared IND applications for its two most advanced candidates: AB-928 (an oral antagonist of the adenosine 2a and 2b receptors) and AB-122 (an anti-PD-1 antibody). The latter was in-licensed development and commercialization in North America, Europe, Japan and elsewhere from Wuxi Biologics Co. Ltd. and partner Harbin Gloria Pharmaceuticals Co. Ltd., in an $816 million deal. This means Arcus can proceed with its planned phase I/Ib trial measuring the safety, tolerability and preliminary efficacy of AB-928 in combo with AB-122 and other drugs plus chemotherapy in patients with breast and gynecologic malignancies. Two more applications are due this month which, if cleared, will let the firm go ahead with trials of AB-928 combos in gastrointestinal malignancies, non-small-cell lung cancer and renal cell carcinoma. At the same time, Arcus has been wrapping up the regulatory process to evaluate the AB-928 and AB-122 duo in Australia. Adenosine is highly concentrated in tumors and their microenvironment, and apparently sends a strong immunosuppressive signal to lymphocytes. Following close behind in the pipeline are a small-molecule inhibitor of CD73, called AB-680, and an anti-TIGIT antibody, AB-154. TIGIT stands for “T cell immunoreceptor with Ig and ITIM domains.” (See BioWorld Today, Aug. 18, 2017.)

Arcus won’t have clinical data until next year. But Basel, Switzerland-based Novartis AG offered early results at ASCO with its adenosine 2a receptor antagonist NIR-178/PBF-509, and Astrazeneca plc, of London, rolled out findings with the anti-ecto-5′-nucleotidase (CD73) candidate MEDI-9447/oleclumab. A February 2016 paper in Trends in Cancer called CD73 “a key molecule, because the degradation of adenosine monophosphate [AMP] into adenosine results in the generation of an immunosuppressed and pro-angiogenic niche within the tumor microenvironment that promotes the onset and progression of cancer. Targeting CD73 has resulted in favorable antitumor effects in preclinical models” and combos with other immune-modulating therapies represent “a particularly attractive therapeutic option.” The same month, a paper in Immunotherapy said “the CD73-adenosine axis constitutes one of the most promising pathways” in I-O. Corvus Pharmaceuticals Inc., of Burlingame, Calif., has CPI-444, an early-stage adenosine 2a receptor inhibitor, as well as CPI-006, an anti-CD73 antibody. BMS is working with an antibody in the latter category as well, BMS-986179.

“We’ve never seen a molecule like this”

Analyst Porges, with an “outperform” rating on Arcus, said AB-928 “has potentially the best profile of all the drugs in development in this class (so far) and we continue to believe that the program represents a worthwhile ‘shot on goal’ in this challenging arena. The low response rates that have been
observed to competitors’ drugs so far highlight the need to identify the optimal match between the right disease (or patient subset) and the right combination strategy.” Arcus’ deep study of biomarkers – including PD-L1, tumor mutational burden, CD73, and other immune checkpoints – in the planned trials could give the company a leg up, regarding the discovery of the mostresponsive patient populations.

Arcus CEO Terry Rosen told BioWorld that, although drugs targeting the adenosine pathway have been around for many years in the central nervous system (CNS) space, “what’s interesting is that a lot of the features that you would design to make a good CNS drug don’t match what you would do if you were trying to design one for the cancer setting, in particular the tumor microenvironment. In the brain, you’re antagonizing adenosine that’s at the level of a neurotransmitter, small little bursts. In the tumor microenvironment, you essentially get tumor cells that are bathed in adenosine. We decided, let’s keep the molecule out of the brain so that we can hit the tumor as hard as possible without running into any CNS pharmacology or toxicity.”

What’s more, the tumor is “loaded with albumin, among other proteins,” Rosen said. With small molecules, “you run into the issue of non-specific protein binding that diminishes their intrinsic activity. In the brain, you don’t have that issue. In designing our molecule, since we were starting from scratch, we set out to incorporate assays” to make certain that “potency wasn’t dramatically shifted. We came up with a molecule with a blood to brain ratio is about one percent, so it doesn’t get into the brain substantially.” Taking aim at the 2b receptor brings an “incremental advantage,” he said. The candidate boasts a long half-life that could allow for once-a-day dosing. Arcus found, “in an assay that measures adenosine receptor activation, we could essentially block that effect to greater than 90 percent, even at 24 hours post-dosing.” The upshot is that, “if there is any benefit to be gained in the right combination and the right setting, we’ll definitely be able to answer the question [of whether] the molecule has the horsepower.”

Adenosine can be blocked at the receptor or at the site of its formation by targeting CD73. “We decided to go after a small molecule approach,” rather than shoot for an antibody, Rosen said. “We were fortunate enough that we set up a collaboration with a professor at the University of Leipzig named Norbert Sträter. He had done essentially the totality of the structural biology that’s out there for this particular [CD73] target,” and solved a lot of problems for Arcus. “We’ve never seen a molecule like this, any of us,” though the team specializes in the area.

Eying potential triplet therapy

Juan Jaen, Arcus’ president, said “there’s an evolving picture, that at least in some patients [who] don’t respond to PD-1 therapy, there’s an increased level of the machinery that produces adenosine, as measured by the expression level of CD73. In those patients, the therapeutic rationale is that combining a PD-1 blocker with AB-928 could rescue some of those patients, sensitize them to the benefit of PD-1 blockade.

I think the science is deeper in terms of how long it’s been understood that, in the presence of adenosine, PD-1 inhibition is going to be impaired, if you want, as far as its ability to stimulate T cells [goes].” The company is especially keen on investigating AB-928 with chemotherapy, and such effort “exceeds the amount of attention that we’re going to be paying to the PD-1 combination,” Jaen said. “There’s a growing notion that [immunogenic cell deathinducing therapies] rely, at least in part, on recruitment of a T-cell immune response against the tumor. There’s no doubt that they’re killing cancer cells in a frontal attack” as well, he said. “The question becomes, [what] if you have a tumor that’s completely bathed in adenosine, where those T cells are infiltrating in response to the chemotherapy and either are unable to help eradicate the tumor or could do a lot better if the adenosine fog wasn’t present. That’s essentially the therapeutic argument.”

CEO Rosen said standard-of-care chemos are “way sub-optimal,” and they – platinum agents and anthracycline – fit into the category referenced by Jaen. The hallmark of immunogenic cell death is the massive spill-out of adenosine triphosphate [ATP] and its conversion to adenosine. We’re looking to really play big, and we think we’ll be the first ones looking at this pathway together with chemotherapy.”

Rosen broke down the science. “What you have is a normal cell loaded with ATP, which you normally think of, in biology, as providing an energy source to everything. But it also has another role. When the cell breaks open, ATP induces an immunostimulatory response.” At this point, CD39 enzyme clips of the first two phosphates in ATP; CD73, the “rate-limiting enzyme, clips off that final phosphate. That produces adenosine, and then not only are you diminishing the immunostimulatory effects of ATP, but you’re forming adenosine. It’s not just a waste product. It’s actually an immunosuppressant. If you think about that from the standpoint of the tumor – we’ll be a little anthropomorphic here – it has managed to highjack that biology, such that, if you picture that you were cracking open this tumor cell and it spills out ATP, it’s calling in the immune system.” Tumors overexpress CD73, “so it’s a protective mechanism,” he said. “ATP is produced, the tumor is fighting that by converting it into adenosine, and turning a signal that would cause its demise into a signal that’s putting the soldiers coming in there into a quiescent state. T cells are attacking, and adenosine is now binding to adenosine receptors on the T cells, causing them to become much less effective.”

Jaen summed up. “You need to think about the convergence of immunogenic cell death in the context of a tumor that’s rich in CD73. Otherwise, you’re going to say, ‘Why did you need to do anything if the chemotherapeutic agent is engendering an immune response in its own right?’” That’s why AB-928 goes into the mix. Although all players in the space are pursuing combos with PD-1 or PD-L1 combos, he finds “almost a surprising overlap in the indications” of interest: NSCLC, renal cell carcinoma, colorectal cancer, and triple-negative breast cancer. “Everybody seems to agree that those are areas worth exploring,” he said. We are the only company at this point that is about to start those combinations with immunogenic cell death inducing chemotherapy.”

Some firms “seem to be moving in the direction of combining CD73 antibodies with small-molecule adenosine 2a receptor antagonists, which is an interesting concept,” Jaen said. “It’s trying to inhibit the axis both upstream and downstream at the same time.” Arcus has chosen a different route, since “we think that the quality of our individual molecules is such that we’ll be able to shut down the pathway either upstream or downstream without necessarily having to double up.” Going for both at once sometimes “reflects that the agents that you’re using for each one of those points of intervention are not all that great,” he said.

CEO Rosen said that his firm’s approach can be seen as replacing the IDO1 bid only at a high level, because the chemistry is so unlike from one to the other. “If IDO had been wildly successful as of today, we’d still be just as excited about the adenosine pathway,” he said. “In the IDO world, while there’s a fair amount understood about its immunosuppressive activity, it’s really not understood completely how [the mechanism] mediates all of its effects downstream. The consequence of that is that there was not really, to the best of our knowledge, a good biomarker strategy for that program.” By contrast, Arcus’ biomarkers emerge as a “natural consequence of the biology” involved.

“Our real strength is around small molecules,” Rosen said, with a nod to the TIGIT and PD-1 candidates. One is “a backbone therapy that already exists. With TIGIT, for a number of reasons, we think it could be the next important backbone therapy,” he said. “A priori, AB-928 is going to get the most attention” as leader of the pack. “But if it doesn’t merit that attention, we’ll diminish its place in our portfolio. We don’t feel like we have this thing, and we have to just bet and bet and bet on it.

We’re not married to it,” and will adjust resources as the data dictate. There’s even talk of a potential triplet, incorporating an adenosine drug, a PD-1 antibody, and TIGIT, but the company plans to take the research step by step. “We’ve got to be careful,” he said.