Cancer immunotherapy is, in a word, tantalizing. It might save the life of someone at death’s door and keep the cancer at bay for years. Or it might not work at all. There’s no telling what a patient will experience.
“There’s a real poverty of understanding of how the machine that is us works,” says Roger Perlmutter, the executive vice president of pharmaceutical giant Merck (NYSE: MRK) and the head of its research division, Merck Research Laboratories.
Perlmutter (pictured) lives this dilemma. He and his team shape Merck’s immunotherapy strategy and its sprawling network of trials, and pick which drugs to pair with the company’s flagship pembrolizumab (Keytruda). More than half of the 900 or so ongoing trials involving pembrolizumab combine the Merck drug with another treatment.
Since he left Amgen (NASDAQ: AMGN) to return to Merck in 2013, Perlmutter has seen pembrolizumab come to market and become the most lucrative cancer immunotherapy in the world. It is on track to generate more than $6 billion in sales in 2018, thanks to approvals in 10 different types of cancer, and has surpassed a rival drug from Bristol-Myers Squibb (NYSE: BMY) to become the standard of care, alongside chemotherapy, for a majority of people with advanced lung cancer. “The impact that it’s having on patients with advanced malignant diseases is remarkable,” Perlmutter says.
But Perlmutter is dealing with the same problem as the biopharma industry writ large. Immunotherapy, for all of its promise, works for only 20 to 30 percent of patients whose immune systems are revved up enough to wipe out their cancers. Merck is racing Bristol, Roche/Genentech, and others to learn why most patients don’t respond, and to boost their results by pairing the right drugs and technologies with immunotherapies like pembrolizumab.
The race has had its share of critics. In interviews with Xconomy, many oncologists have chided companies for moving too fast, running redundant studies, and launching trials of combination therapies with inadequate grasp of the effects the combinations might have in the body. Some high-profile failures have ensued, most notably last year’s Phase 3 flop in melanoma, which combined pembrolizumab and a so-called IDO inhibitor, epacadostat, from Incyte (NASDAQ: INCY). The ripple effects are still being felt.
To hear what might be next for the field and for Merck, Xconomy caught up with Perlmutter at the J.P. Morgan Healthcare Conference in San Francisco last week. The following conversation has been edited and condensed.
Xconomy: Why don’t most people respond to immunotherapy, and what is Merck doing to boost those response rates?
Roger Perlmutter: There are three explanations. The first is a cancer that no immune system can see, period. In that circumstance, you’re going to have to try alkylating agents [a type of chemotherapy] or other things to change the nature of the cancer so that it becomes more recognizable to the immune system.
The second is the tumor is recognizable by an immune system, just not the patient’s immune system. In that case, immunization might be helpful. We’ve embarked upon a strategy that says, what can we immunize with, and how can we identify patients who have those characteristics?
The third is a patient’s immune system can see the tumor, but for one reason or another, just [one type of immunotherapy] isn’t enough. There’s something else you need. A lot of focus has gone into that.
Xconomy: Given all the potential options, how do you focus?
RP: Anything that is pro-inflammatory is beneficial alongside Keytruda. That includes a lot of things that you wouldn’t think of, ordinarily, as being pro-inflammatory. Radiation therapy is pro-inflammatory, and you can see it work in combination. Traditional platinum-based cytotoxic chemotherapy is pro-inflammatory. While most people said, ‘Oh, you shouldn’t be giving these things together, because you’re going to suppress the immune system,’ the reality is the pro-inflammatory character wins out. I had a lot of people who said, ‘Why are you doing these chemo combos?’ Yet look at the results.
Surgery is also a pro-inflammatory stimulus, and that’s one reason the neo-adjuvant approach [as in, giving a patient Keytruda before surgery] has worked. I think we’ve made enormous progress in pursuing these different paths.
Xconomy: Several oncologists believe there isn’t enough basic research and too much spaghetti-against-the-wall when it comes to clinical trials for combinations. What is your response?
RP: The philosophical underpinnings of our approach have integrity. It’s not like we do everything. To those who are concerned that we’re just saying, “Oh, what the hell, let’s try anything, orange juice, mineral oil, whatever”—we are thinking this through very carefully. Among the literally hundreds of things we could do, we select just a few. Although we have more than 450 combination studies ongoing, it’s the product of a very thoughtful, precise selection.
Xconomy: What would help produce a more rationalized, refined approach?
RP: There are big gaps in our understanding of how Keytruda works. I would like a much more thorough molecular, cellular description of what happens in a cancer patient who responds to Keytruda. What are the T cells doing that are responsible for this good work? An understanding of those cells could be enormously helpful in getting better combinations with Keytruda. Thus far it’s an embarrassment to say—and I’m embarrassed for the entire scientific community—none of us have been able to achieve this.
X: What’s holding this back? A lack of available technologies?
RP: A little bit. But the clinical studies are hard to do. They’re prospective studies with multiple biopsies. Not every cancer patient is appropriate for that, and not everyone will give informed consent. Of course, you would like to compare [immunotherapy] to other therapies. But these days everybody says, ‘I’ll take the [immunotherapy], thank you very much.’ So it’s very hard to do a randomized study.
X: One of the most high-profile combination failures so far was the Phase 3 test of Keytruda and Incyte’s IDO inhibitor epacadostat in melanoma. What happened?
RP: IDO was pretty speculative to begin with. We did single-arm studies with our colleagues at Incyte, and if we squinted really hard, it looked like, gosh, those responses seemed a little bit broader and a little bit deeper [than with Keytruda alone]. Not changing night into day, but a little bit better. So we said, OK, let’s do a randomized, controlled clinical trial.
Incyte was so convinced they were going to see something, they wanted to do five [Phase 3 trials]. We said, let’s start with one. In essence, we found that the glimmer of hope [in the single-arm studies] was just a play of chance, which is so often the case. The IDO inhibitor really doesn’t have an effect. It was an important test to do, but I always had reservations, and that’s the reason I wasn’t willing to pile into six Phase 3 studies. After the first one, we changed another to a Phase 2 study and got rid of the others.
[Editor’s note: Incyte didn’t respond to requests for comment.]
X: Several of your peers have made big bets on CAR-T cell therapy [a cutting edge form of immunotherapy that uses live T cells, modified to find and kill cancer]. Bristol, for instance, just agreed to buy Celgene in part to bet on the future of CAR-T. Why hasn’t Merck jumped in?
RP: I love the CAR-T data in hematological malignancies. People’s lives have been saved, so that’s great. But to make this work, you’re going to have to treat solid tumors. And if you’re going to do solid tumors, you have to find a tumor-specific antigen to wipe out. Guess what? People have been looking for tumor-specific antigens for the last 50 years. We don’t have any.
The second problem is the industrialization of the process. Right now, with autologous transplant—we remove a blood sample from you, engineer it to kill your tumor cells, and in 14 days return the sample to you—the chain-of-custody issue becomes a big deal. Anybody who’s practiced medicine knows how often blood transfusions are mismatched. I hate to bet against American engineers, or any engineers for that matter, but the full-scale throughput for CAR-T is very small.
David Chang, who developed Yescarta [one of two FDA-approved CAR-T therapies] at Kite Pharma, and who used to work for me at Amgen, spent a lot of time talking to me about this. When Kite was acquired by Gilead [Sciences], the maximum throughput in their new facility was six patients a week. I’m not interested in treating six patients a week. I need to treat 100,000 patients. How am I going to do that?
The future for those kinds of therapies is tumor-specific antigens and allogeneic [‘off the shelf’] treatments, so you’ve got something in a bottle.
[Editors note: A Gilead spokeseman told Xconomy that its El Segundo, CA, manufacturing facility currently has the capacity, at full operation, to produce up to 4,000 CAR-T treatments per year, which would equate to as many as 77 per week.]
X: Are you waiting for some type of proof-of-concept from the allogeneic players before jumping into that business?
RP: Or for the improvement in bi-specific [antibodies], which we and others are doing. Those are good, too. The bi-specifics, in principle, could do everything that the cell-based therapies can do.
X: By comparison, you were in on mRNA technology early through a relationship with Moderna. [Messenger RNA, or mRNA, is an unproven drug-making method that aims to coax the body into producing its own disease-fighting proteins.] How do you think it will help you?
RP: We’ve known Moderna since before they started. Shortly after I returned to Merck in 2013, they said they’d found something surprising—in principle they could make any vaccine component just from the mRNA. We have a very big vaccine business, and I’m interested in improving it. Maybe we could use this to get at some things that are otherwise extremely hard to engineer—a common flu vaccine, for example, so you don’t have to make a new one every year based on what shows up in Hong Kong. Or a vaccine for respiratory syncytial virus, for which we’ve never been able to develop a satisfactory immunization regimen.
Also, it became clear that mRNA seemed like a natural way of immunizing patients [who might otherwise not respond to immunotherapy]. Keytruda doesn’t prime the immune system; mRNA would prime it.
X: Where do you see cancer care going in the near-term?
RP: You’re not going to see much monotherapy anymore, except perhaps in an adjuvant setting. In first-line [treatment of] lung cancer, which is obviously the biggest market, 70 percent of patients are [eligible for] Keytruda plus chemo, and it’ll be more later. My expectation is that Keytruda will be foundational in the treatment of the majority of malignancies, and the question will be Keytruda plus what? It’s going to be Keytruda plus X, where X is going to be a large set of different things, and it will be much more personalized, depending on the tumor type.