Cytokines a REVIEW:
"Those small but potent messenger proteins, cytokines, have become the drug targets of choice in rheumatoid arthritis following the extravagant success of Enbrel and Remicade, which both inhibit the action of tumor necrosis factor. Cytokine blockade, once considered a sure loser, is now wildly popular and biotech and pharmaceutical companies alike are testing intriguing new therapies in the clinic – including an array of interleukin-blockers, more than a few p38 MAP kinase inhibitors and a handful of small molecules intended to block cytokine production per se. Each has potential, but cytokine blockade, as these companies are discovering, can have unexpected and unpleasant consequences. Only a lucky few of these drugs will survive the perils to emerge winners.
At a scientific meeting in 1992 in Arad, Israel, Ravinder Maini presented the first clinical results for a new monoclonal antibody. Maini, a rheumatologist at the Kennedy Institute in London, England, had treated twenty rheumatoid arthritis patients -- with incredible results. Pain symptoms decreased by 73 percent, the number of swollen joints fell 72 percent, and the median duration of patients' morning stiffness dropped from three hours to five minutes. "The [audience] response was, 'It sounds too good to be true,'" recalls immunologist Marc Feldmann, Maini's Kennedy collaborator. Their anti-cytokine antibody was an instant sensation.
The Arad talk sent shock waves through the biotech community. Stunned Immunex scientists in the audience went home to rush their own molecule into development to compete with Maini's antibody, later dubbed Remicade. Other companies launched their own anti-cytokine programs. Now, a decade later, over 400,000 rheumatoid arthritis patients have been treated with Immunex Corp.'s Enbrel and with Remicade, made by Centocor Inc. (now part of Johnson & Johnson). Together they generated over $2 billion in sales last year (Immunex was completely unable to meet Enbrel demand), and should approach $3 billion this year. In 2001 Amgen Inc.'s Kineret hit the market, and in January 2003 Abbott Laboratories' Humira became the fourth member of the anti-cytokine family of blockbusters.
More are coming. The astounding success of Enbrel and Remicade, both of them TNF (tumor necrosis factor) inhibitors, has led the pharmaceutical and biotech industries to embrace cytokine therapy with unbridled enthusiasm. Rheumatoid arthritis (RA), a joint-crippling autoimmune disease that afflicts roughly two million Americans, is the preferred testing ground. At least sixteen cytokine-targeted agents are in clinical development for RA (see the table below) and many more are in the preclinical pipeline. The first decade of the 21st century is poised to become the decade of the cytokine.
But pharmacologically suppressing key elements of the body's immune system can lead to unexpected consequences. As more patients take cytokine blockers, more serious side effects are reported, and the new therapies must navigate uncharted toxic minefields if they hope to reap the financial rewards of the industry's vast collective investment.
Even ten years ago, few would have predicted the cytokine's ascendance to its current heights. "Most rheumatologists would not have believed that blocking one cytokine would do the job and could be done relatively safely," says David Fox, a rheumatologist at the University of Michigan. "It was a big leap."
Cytokines -- small, potent proteins that communicate between cells -- are relative latecomers to biology. The prototype inflammatory cytokine, tumor necrosis factor, was discovered and named in 1971 by tumor immunologist Lloyd Old of the Memorial Sloan-Kettering Cancer Center in New York. Old injected cancerous mice with two bacterial toxins, shrinking the tumors. He correctly deduced that the animal's immune system had responded to the toxins by producing some tumor-shrinking factor. He called the unknown factor TNF. The same year, interleukin-1 was serendipitously discovered by Yale University's Igal Gery, who was trying to show that immune tolerance to antigens could be transferred from one mouse to another. Gery unexpectedly observed immune system stimulation in his control cultures and called the mysterious activating substance "lymphocyte-activating factor." (In 1979 it became interleukin-1.) Both cytokines were cloned in the mid-80s; today, the cytokine family has over a hundred members.
Throughout the 1980s rheumatoid arthritis was viewed as a disease sustained by out-of-control T cells, so most researchers ignored cytokines. But work by the Kennedy Institute's Feldmann and by Gary Firestein of the University of California, San Diego, among others, demonstrated in the late 1980s that inflammatory cytokines were present in synovial fluid (the lubricating fluid in the joint) in patients with rheumatoid arthritis. In 1990, Firestein proposed a major role for cytokine "networks" in the disease. "We were really considered heretics at the time, and took a lot of heat for it," Firestein recalls.
*This product incorporates Nektar Therapeutics' PEGylation technology
Meanwhile, Feldmann's lab (to many's disbelief) was showing amazing effects from TNF-blocking antibodies in culture and later in animals. "A whole lot of different cytokines were downregulated when you gave anti-TNF in culture," says Feldmann. "We then thought we'd hit the jackpot."
Almost everyone else viewed targeting cytokines as a waste of time. "The thought was, if you just block one of them, then all the others will just keep on driving the biological processes," Feldmann remembers. "Therefore the whole thing would be pretty useless." Even if TNF blockade worked, critics reasoned, it would surely be too toxic. "There was great concern that, if you blocked TNF, the potential for side effects would be unlimited," Firestein says.
Almost alone in his conviction that anti-TNF would work in rheumatoid arthritis, Feldmann approached several biotech companies, with no luck. He finally convinced Centocor, though: The company, which already had made an anti-TNF antibody to try in sepsis, agreed to move it ahead in arthritis.
The Arad meeting in 1992 quieted the skeptics, and clinical trials moved steadily forward, with Immunex soon overtaking Centocor (which got diverted into Crohn's Disease). Enbrel won the race, garnering FDA approval in November 1998; Centocor's Remicade was approved for RA the following year.
Cytokine therapy, against all odds, had arrived. It succeeded where monoclonal antibodies against T cells, including anti-CD4, anti-CD5, and anti-CD52 (Campath), all failed miserably as RA therapies. Most of these antibodies provoked serious infections while barely showing a therapeutic effect, probably because they depleted "suppressor" T cells while failing to eliminate helper T cells in the synovium. Matrix metalloproteinase inhibitors, most infamously Roche's Trocade, also fell by the wayside, victims of the redundant effects of proteases active in joint destruction. "It would be extremely difficult to block enough metalloproteinases to really make a major difference," says Raphaela Goldbach-Mansky, a senior clinical investigator at the National Institutes of Health (NIH).
TNF blockers won out in the end, but they're hardly perfect drugs. Although about 70 percent of patients show some benefit, only 30 percent experience a dramatic response. And while these drugs halt joint destruction in responding patients, "they are not cured of rheumatoid arthritis," notes UC San Diego's Firestein. "If you stop the drug their disease will return." Since a year's worth of Enbrel or Remicade can run $15,000, this is a problem. Even more worrisome, for long-term treatment, is the threat of serious infections, especially fungal infections and tuberculosis. Blocking TNF seems to lead to the breakdown of granulomas, nodes of T cells and macrophages that contain the spread of latent tuberculosis, which is harbored by several million Americans. In addition, a few patients on TNF blockade have developed a transient disease resembling lupus erythematosus, and some others display neurological side effects like confusion and loss of coordination.
Tumor necrosis factor-alpha
(Research Collaboratory for Structural Bioinformatics; Protein Data Base)
Finally, there is the potential for cancer, since TNF suppression might theoretically allow the activation of latent Epstein-Barr virus, which can transform B cells and lead to lymphoma. So far the cancer fear has not materialized, but an FDA meeting to review cancer risk, scheduled for early March, suggests that the agency may have new data implicating TNF blockade in cancer.
And TNF inhibitors may have maxed out in terms of efficacy. Second-generation agents in the pipeline, like Celltech Group plc's CDP870, have much longer half-lives and thus offer greater dosing convenience. But they're not likely to work better than Enbrel, Remicade and Humira. "I don't believe that anybody's going to come up with another pure TNF inhibitor that is somehow more effective than the ones that are already out there," says Firestein. "The same is not true with IL-1. I'm optimistic that we will see better results for it using newer agents."
At least three companies have novel IL-1-targeted therapies in the clinic. But an IL-6 antagonist instead stole the show at last October's American College of Rheumatology (ACR) meeting in New Orleans. In a randomized, placebo-controlled trial, a stunning 78 percent of patients receiving a certain dose of an anti-IL-6 receptor monoclonal antibody achieved an "ACR20" response, compared to 11 percent on placebo. (ACR20 is a standardized mix of subjective and objective measures of RA patient improvement, and has become the benchmark for FDA drug approval.) "It looks like it's going to be, at least in the preliminary studies, as effective as the TNF inhibitors," says Firestein. "I was surprised… at least in animal models, it is not really one of the key cytokines that regulates inflammatory arthritis."
The emergence of a new target in arthritis has jolted the industry. The anti-IL-6 antibody, called MRA, belongs to Chugai Pharmaceutical Co. Ltd. MRA "is certainly something that has caused a major impact and great interest," says the NIH's Goldbach-Mansky, who nevertheless urges caution. "This is short-term data, and the observation period in the ACR paper was only three months. The published data covered only a treatment period of two weeks." MRA investigators, at the ACR meeting, verbally reported at least one case of active Epstein-Barr virus (EBV) infection in their patient cohort. EBV can infect the liver, lymph nodes and spleen, with possible organ damage, and can lead to lymphoma. "One really needs to be careful, because the side effect profile of these agents cannot be predicted," says Goldbach-Mansky.
This patient died two months after receiving a single dose of MRA. The apparent cause of death: hemophagocytic syndrome (the destruction of red blood cells by the immune system) following EBV activation. According to trial investigator Norihiro Nishimoto of Osaka University, it's still unclear that MRA caused the death. "There are two possibilities," he writes in an e-mail to Signals. "MRA might suppress the immune system of the patient and result in the activation of EBV, [or] MRA might suppress the EBV activation and disappearance of MRA might have caused the restoration of IL-6 activity and this IL-6 might activate EBV. But we do not know the exact mechanism at this moment."
The patient had detectable DNA from Epstein-Barr virus in her plasma, but was still given the drug because she "could not be excluded by the attending doctor," writes Nishimoto. Nevertheless, he adds, "we can exclude such a patient" from future treatment. Whether or not MRA caused the death, this incident "raises a red flag indicating that we need to be aware of the role that these [immune system] mediators play in host defense," writes Firestein.
IL-1 blockade may be a safer bet, because Kineret, Amgen's marketed IL-1 receptor antagonist, has not yet caused any exotic side effects. While it's just as likely as Enbrel and Remicade to lead to serious infections, Kineret so far has not been implicated in tuberculosis, lupus, EBV activation or cancer. Numerous companies are placing their bets on direct or indirect IL-1 inhibitors, not least because IL-1 has long been strongly linked to joint destruction in rheumatoid arthritis, and because IL-1 blockade may work for people who don't respond to TNF inhibitors.
But IL-1 is a more elusive target. Blocking IL-1 receptors with antibodies requires saturation to prevent IL-1 signaling. "Blocking half or three quarters of the receptors doesn't seem to be good enough," says the University of Michigan's Fox. "You might have to block more than 95 percent. So this is just a tough hill to climb." This challenge may explain why Kineret is not as widely effective as TNF inhibitors.
Or it could be that IL-1 is just an inferior target. "The question is, is [Kineret] not as good as the TNF blockers because IL-1 isn't as good a target, or is it because their blocker just isn't blocking IL-1 as well as you potentially could?" says Neil Stahl, senior vice president of preclinical development and biomolecular science at Regeneron Pharmaceuticals Inc. Stahl believes IL-1 is an excellent target, and argues that targeting the cytokines themselves makes more sense than trying to saturate their receptors. It's a compelling argument, but this approach has its own pitfalls: Immunex tried a soluble IL-1 receptor as a drug, along the lines of Enbrel for TNF, but it didn't work in human trials. That's probably because Immunex's molecule bound the endogenous IL-1 receptor antagonist more than IL-1 itself, cancelling out any beneficial effect.
Interleukin 1 beta
(Swiss Institute of Bioinformatics, SWISS-3DIMAGE)
The great challenge is to target the IL-1 receptor agonist cytokine without taking out the antagonist. Regeneron's solution is its IL-1 Trap fusion protein. The drug grew out of research in the early '90s at Regeneron showing that cytokine signaling requires binding to dual receptor components in a two-step process. "If we could make a soluble blocker that encompassed both cytokine receptors," explains Stahl, "it would have the potential to bind the cytokine with very high affinity." Regeneron's resulting molecule has about 25 times greater affinity for IL-1 than for IL-1 receptor antagonist, thus promising specificity. IL-1 Trap, in Phase I trials, showed good safety, and it appeared to work well at the higher doses. A randomized Phase II trial is ongoing.
Meanwhile, according to Amgen spokeswoman Rebecca Hamm, Amgen has completed phase I enrollment for its IL-1 receptor type II, which appeared to have stalled in development in the wake of the July 2002 Immunex acquisition. The type II "decoy" receptor, in theory, should bind IL-1 but not IL-1 receptor antagonist, thus avoiding Immunex's earlier problems. "Perhaps that could be the Enbrel equivalent…for the IL-1 system," says the University of Michigan's Fox. But until anti-IL-1 biologicals work in large controlled trials, their ability to compete with TNF blockade will remain in doubt.
Vertex Pharmaceuticals Inc. is taking a radically different approach. Instead of going after IL-1 or its receptor, Vertex chose to block the cytokine indirectly by targeting the upstream enzyme responsible for its processing. The result is Pralnacasan, a small molecule ICE inhibitor. ICE, or interleukin-beta converting enzyme, is an enzyme that cleaves IL-1, releasing the active form of IL-1 beta. Without the clip, IL-1 cannot get out of cells (and so cannot deliver inflammatory signals or damage tissue). A few years into its ICE program, Vertex discovered that Pralnacasan also blocked IL-18 in the same way. Although less is known about IL-18, "it's a very important cytokine," says John Alam, Vertex's senior vice president of drug evaluation and approval. "If you knock out IL-18, or affect IL-18, you affect interferon-gamma, plus TNF and IL-1 and some other cytokines."
Vertex's upstream approach has obvious advantages. Blocking cytokine production, instead of blocking its activity, leads (in theory) to more complete inhibition and greater potency. And targeting an enzyme's active site allows a small molecule approach, instead of using an antibody or soluble receptor to prevent protein-protein interactions a la Enbrel and Remicade. A pill is not only far more convenient (and marketable) than an injectable, but it's also likely to be safer than the once-a-month or twice-a-month doses that will eventually become standard in cytokine blockade for RA. If patients on a twice-monthly therapy develop an infection, "you really can't reverse the blocking of the immune system," says Alam. A once-a-day drug, in contrast, he points out, should allow quick immune system recovery once the patient stops taking it.
Interleukin-1 beta converting enzyme
(Swiss Institute of Bioinformatics, SWISS-3DIMAGE)
But "upstream" small molecule inhibitors carry risks. "There is inherently less specificity than with a biologic," Fox points out. "For none [of the target enzymes] do we know all the substrates. So there are going to be some other molecules that are converted to their active or inactive forms by the same enzymes, and there could be other modes of action that could be helpful, or could be harmful, or could be counterproductive…I'd be sort of cautious until I really see the clinical data."
So far, the Pralnacasan clinical data look good. Vertex and its partner, Aventis, recently completed a 285-patient Phase IIa clinical trial. At its high dose Pralnacasan achieved a good-but-not-great 44 percent ACR20 response overall, with few side effects. A subset of patients already taking methotrexate (the standard first-line RA therapy) before adding Pralnacasan achieved an impressive 54 percent response -- "right in the range of the anti-TNF drugs in that population," Alam points out. Vertex and Aventis plan to hike the dose further in a Phase IIb study to begin in the second quarter of this year.
Several other companies, in the same mode, are working on upstream blockade of TNF using small molecules. These drugs, including one from Bristol-Myers Squibb Co. now in Phase II, target TACE (TNF-alpha converting enzyme). But TACE inhibition, because it shuts down TNF production, may be riskier than the downstream Enbrel/Remicade approach. "TNF is needed for many [normal] functions," says the NIH's Goldbach-Mansky. "So being more efficient at blocking TNF might lead to a higher toxicity profile. And I'm not sure we would gain a lot from doing that."
Some raise the same concern for drugs targeting p38 MAP kinase in rheumatoid arthritis. "I just don't see how you can block anything in the MAP kinase pathways without affecting cells throughout the body," says the University of Michigan's Fox. "Those are such fundamental cellular processes."
But p38 inhibitors should be tried, says Firestein of UC San Diego. "We said the same thing about cytokines ten years ago," he said. "But it turns out, actually, that humans are fairly adaptive creatures, and blocking IL-1 or TNF or some of these other very important mechanisms can have therapeutic benefit, with a manageable risk." Firestein thinks blocking p38 might work with manageable side effects, as long as specific isoforms of the enzyme -- such as p38-alpha -- are targeted.
Specificity is crucial for success. Mitogen-activated protein (MAP) kinases control a vast array of physiological processes by phosphorylating their target proteins. The p38 family was discovered in 1994 at SmithKline Beecham plc, after chemists there screened a compound library for TNF inhibitors. Finding one especially potent compound, they then worked backwards to identify its target: p38 MAP kinase. The company's early drug candidates foundered, however, due of lack of p38 specificity or because of toxicity. GlaxoSmithKline may have overcome these obstacles, though, because it has a new agent in Phase I trials.
Vertex, Scios Inc. and Boehringer Ingelheim GmbH also have p38 inhibitors in the clinic, all of them further along. P38 remains an attractive target, Alam says, because blocking it seems to inhibit TNF, IL-1 and IL-6 at once. "It's a master switch, or kind of a 'dimmer' switch for inflammation," Alam says. "When you shut off p38, you get a dampening effect against a number of different cytokines that are at the heart of the inflammatory response." The way p38 activates cytokines is not fully understood, but it seems to involve inducing the expression of some cytokines and stabilizing the messenger RNA of others. But it's clearly a key regulator.
And clearly a key component of Johnson & Johnson's long-term strategy to increase its presence in the RA field. The big pharma is paying $2.4 billion to acquire Scios – especially its p38 drug candidate, which is an orally delivered compound.
A dampened cytokine response "is, from a clinical standpoint, exactly what you want," says Vertex's Alam. "You don't want to shut off TNF completely, because there will be times when the body needs TNF. And you don't want to shut off IL-1 and IL-6 completely."
But no one has yet proven the importance of p38 in rheumatoid arthritis. "We know p38 is expressed in the synovium," says UC's Firestein. "We even know that it is activated in the synovium and that it is phosphorylated. But what we don't know is what percentage of the cytokines that are produced in the joint are regulated by p38, in rheumatoid arthritis."
Alam thinks those doubts are overblown. "I would have agreed maybe three years ago," he says, "but with some of the data that we have, we think we've addressed that -- obviously not fully." Vertex's first-generation p38 inhibitor, VX-745, was abandoned in September 2001, but not before demonstrating a good 40 percent ACR20 response in a 59-patient trial. (Vertex dropped the drug because it penetrated the blood-brain barrier in animals, raising the possibility of nervous system toxicity.) The biotech industry views the VX-745 results as important proof of principle for p38 inhibitors. Phase II trials on Vertex's current drug candidate VX-702 – which does not breach the blood-brain barrier -- are scheduled to begin by mid-year.
These drugs are just the most obvious candidates for the next rheumatoid arthritis blockbuster. Other cytokines are now emerging as drug targets, especially IL-12, IL-15 and IL-18. Some experts think NF-kB, a transcription factor central to immunity and inflammation, can be targeted in RA, although blocking such a prolific regulatory molecule has obvious risks. But several drugs already in the clinic target chemokines, which recruit leukocytes to the joint, and adhesion molecules, which promote T-cell migration, aggregation and activation. A Bristol-Myers Squibb molecule blocks T-cell costimulatory signals, and even B cells are emerging as a drug target now that IDEC Pharmaceuticals Corp.'s cancer drug Rituxan has shown some worth in rheumatoid arthritis patients. "The question is not whether there are potential targets, but how are we going to prioritize them and move them rapidly into the clinic?" says Firestein. "Because there are not enough patients, there's not enough money, there are not enough doctors, to test all of these targets."
Firestein is pushing drug companies to biopsy the joints of their early patients to check for molecular drug effects, rather than waiting for efficacy results. This would help triage new drug candidates. But the bigger need is to thoroughly understand what causes and perpetuates rheumatoid arthritis, a quest that's only beginning. "This is truly one of the greatest mysteries left in medicine," says Fox. Firestein and Feldmann, who challenged the orthodoxy a dozen years ago with their brash, contrarian cytokine theories, have been vindicated by success of the new biologics. A cure for rheumatoid arthritis must await an equally revolutionary breakthrough."
signalsmag.com
Enjoy
Jeff |
