publication date: Aug. 9, 2019
Learning from suramin: A case study of NCI’s much-hyped cancer drug that crashed and burned—35 years ago
By Cy Stein, MD, PHD
City of Hope National Medical Center
Almost 35 years ago, while the nation suffered in the vicious grip of the HIV epidemic, a young man from South Carolina with AIDS named Boyd Helton found his way to the NIH Clinical Center in Bethesda. While there, he was recruited into a clinical research protocol designed to lower the expression of viral proteins in his blood, and, ideally, to increase the numbers of his circulating CD4+ T-cells.
The drug treatment examined in the clinical research protocol, an inhibitor of multiple viral and mammalian DNA and RNA polymerases, had been chosen because of work performed in the lab of Eric DeClerq, a noted Belgian scientist. The principal investigator on the treatment protocol was Sam Broder, in the years before he successfully investigated AZT in patients with HIV, and later became NCI director.
The name of the drug was suramin.
Helton was Patient Zero in what I call the suramin saga, which culminated in the drug eventually entering clinical trials as an anticancer drug. After some initial apparent success in prostate cancer, suramin—for a radiant instant—became the cynosure of all eyes in solid tumor oncology.
It’s a story that has made its way into books and—nearly—into the movies. I collaborated with my cousin Harry Stein, a professional writer, to produce “The Magic Bullet,” a heavily fictionalized take on the development of suramin.
So why isn’t my name listed as an author?
Because I was a young junior faculty member at the time, with a strong sense that publishing novels would be frowned upon by my then-current institution.
The “Magic Bullet” was published in 1995 by Delacorte Press. The work briefly became a USA Today best-seller and was optioned to New Regency Pictures. The late Penny Marshall signed on as director, and a screenplay was commissioned from Paul Schrader, whose screen credits include “Raging Bull,” “Bringing Out the Dead,” and “First Reformed.” Sadly, the movie was never made. Penny soon backed out to do “The Preacher’s Wife,” rejecting the suramin saga in favor of Denzel Washington and Whitney Houston. Who could imagine? The studio brought in a second director and screenwriter, but after investing two million bucks, canned the entire project.
The brief hullabaloo surrounding suramin was a false dawn, the fool’s gold of cancer drug discovery: suramin could also produce mighty serious toxicity.
The approach to the pharmacologic challenge presented by this drug has also been mocked and derided over the past several decades. To add insult to injury, both the drug and the PI on all the clinical suramin trials were bashed in a 2012 book by Otis Brawley and Paul Goldberg titled “How We Do Harm,” published by St. Martin’s Press.
“If it didn’t work, double the dose”
I contend that the story surrounding the development of this drug is far more nuanced than most are aware of. I know, because I was responsible for much of it.
Now, almost 35 years later, with the benefit of the objectivity of time and distance, I believe it’s worthwhile to undertake take a dispassionate reappraisal of the suramin saga, both to set the story straight and to see what lessons can be learned.
In 1985, when the tale begins, the world of cancer treatment looked very different. In that world, chemo was king. The clinical wisdom about chemo treatment, expressed by the highest authority at the NCI was, “if it didn’t work, double the dose.”
We were also ominously informed that “you can’t make omelets without breaking eggs.” Perhaps the NCI doctor who so instructed me, a decent chap otherwise, was unaware this was Lenin’s justification for the beastly treatment he meted out to anyone who disagreed with his inhuman policies during the Bolshevik Revolution of 1917.
In 1985, there were no cell phones and no Internet. Fax machines were still in development. There were no monoclonal antibody-based treatments for cancer, no tyrosine kinase inhibitors, and no immuno-oncology agents (except for perhaps BCG). 5-HT3 receptor antagonists, such as Zofran or Kytril for emesis, were unknown. Oral hormonal agents for prostate cancer remained well beyond the far horizon, the successful marketing of taxanes was still years in the future, and the term “personalized medicine” was a non sequitur.
Though attitudes were changing, in 1985 the physician still retained much of his or her god-like aura. Patients heard, believed in, and accepted experimental treatments that could produce toxicities that today would be unacceptable. Cancer was a tough disease that required tough doctors!
The attending physicians who were our mentors, we were told, were the best of the best, the cream of the crop. Each was doing God’s work. As first-year fellows at the Medicine Branch of the NCI in 1985, this was the ambiance, the culture we young, eager physicians were expected to absorb without question or doubt. And if doubts you dared to have, it was best to leave them at the door or keep them to yourself. When consorting with the lords of cancer (the “oncocrats”, so named by Eli Glatstein, chief of radiation oncology at the NCI), be mindful you were a serf.
Sometime in the second half of 1985, when I was a first-year fellow, Boyd Helton became my patient. I was responsible for ensuring his adherence to the suramin for HIV protocol, making certain he had his blood drawn and his treatments on time. Some doctoring was involved, but steering patients through clinical protocols at the NCI was mostly a clerical job for the highly educated.
Sam Broder was the fellowship director. He was cheerful, helpful, encouraging and usually respectful of the folks in the trenches who were doing their best to manage some very sick patients. I remember Sam as having a quick wit and a razor-sharp sense of humor. He wielded both like a stiletto in a knife fight, eviscerating an incautious or unbalanced opponent. Anyone, including cancer lords, who tried to match wits with Sam would come off second best. But for all his bonhomie, Broder set a high bar for the NCI fellows. He expected only the best from us; nothing less was acceptable.
One day, Boyd called me and said that no matter how much water he was drinking, he was still dehydrated and needed to drink more. He also complained of dizziness on standing. I brought him into the NIH Clinical Center for a blood draw and an exam. My first thought was that he had developed adult-onset diabetes. But that turned out not to be the case as his sugar was normal. Perplexed, I sent him home.
Three days later, Boyd called again, complaining of the same problem. Another trip to the Clinical Center and an in-depth investigation revealed that Boyd suffered from adrenocortical insufficiency, or Addison’s Disease. His adrenal cortexes were not producing sufficient steroid hormones to allow him to retain sufficient salt and water for him to support his blood pressure. This adrenal gland dysfunction was a side-effect of his suramin treatment (C. Stein, et al., Ann. Intern. Med., 1986 104:286-287). The problem was corrected and Boyd soon recovered.
From African trypanosomiasis to adrenal cancer
Several days later, I dispatched my medical student, Wayne Saville, to the library to gather published material on suramin. He returned with a treasure trove of information. Most significant were the observations that suramin could bind to and inhibit the functioning of various heparin-binding growth factors (e.g., platelet derived growth factor; information about basic fibroblast growth factor came a little later) that had been implicated in tumor growth.
Little or no information was found about any other toxicities of the drug, which had been used for many years in the treatment of African trypanosomiasis, as well as other tropical diseases. I learned that suramin had first been synthesized in 1916 by students of Paul Ehrlich, and was referred to as Bayer 205 by Paul de Kruif in “Microbe Hunters,” a book I’d read over and over as a youngster.
The drug intrigued me—perhaps it was an anti-HIV drug, and perhaps even more. For by this time, I learned, Alexandra Levine (who coincidentally more than 25 years later was to become my supervisor at the City of Hope) and her group at USC had seen a tumor response in a suramin-treated lymphoma patient with HIV (A. Levine et al., Ann. Intern. Med.,1986 105:32-37).
If suramin could ablate the steroid-producing function of the adrenal cortex by causing the destruction of the zona fasciculata, could it perhaps also destroy malignant adrenocortical cells? This thought paralleled the history of mitotane, a DDT relative which also destroyed the normal adrenal cortex. In 1985, mitotane was the only active agent for the treatment of adrenocortical cancer.
Suramin needed to prove itself in the clinic as an anti-cancer drug. A trial in metastatic adrenocortical cancer required strong backing from one of the attending physicians. So, I brought my ideas to Charles E. “Snuffy” Myers, Jr., head of the Pharmacology Branch of the NCI. (The moniker “Snuffy” was his father’s, who was thought to resemble an old cartoon character named Snuffy Smith).
Sam Broder was also very encouraging and remained so almost throughout.
Compared to the titanic egos found in the Medicine Branch of 1985, Snuffy seemed mild-mannered, affable and relatively non-judgmental. He was also open to new ideas. With the reputation of super-strength in clinical cancer pharmacology, Snuffy had recently won an award for his work in intraperitoneal chemotherapy from the Milken Family Foundation. On the other hand, with respect to the practice of clinical medicine and oncology, he was perhaps less strong.
In “How We Do Harm,” Brawley refers to Snuffy Myer’s “beliefs” about suramin. He claims that science has demonstrated they were “without exception, wrong.” But Brawley’s Snuffy is not the Snuffy with whom I worked closely in the late 1980s.
He adhered to the same rules of evidence we all did. For example, the drug was not taken directly from Boyd Helton directly into patients with cancer. We first determined that the drug also wiped out the adrenal cortex in monkeys (P. Feullian, et al. J. Clin. Endocrinol. Metab., 1987 65:153-158). Had Snuffy gone rogue, I would have ceased working with him. And I’m no reflexive defender of Dr. Myers. I’ve spoken to him perhaps once or twice in the past thirty years. But I never knew him to be a “nuke it, napalm it and damn the consequences” buccaneer as Brawley pictures him. He was far more thoughtful than that, except perhaps for a tendency to fly by the seat of his pants. Like everyone else, Myers was ambitious and accepted the prevailing orthodoxy, which had produced active treatments and even cures for some tumors. At the time, there simply was no other orthodoxy to believe in.
The other person I convinced to join the suramin team was Renato LaRocca. Renato and I had a long history together. We were interns and residents at the same medical center in New York City, and were now first-year fellows at the NCI. In our second year, we even joined the same molecular biology research laboratory. Renato was bright and articulate, and, in everyone’s opinion, a superb physician. I thought him an excellent addition to the team, and one sorely needed.
That was the way Snuffy first became interested in the use of suramin as an anticancer agent. Other assertions by Brawley in his book were written many years after the fact. In other areas, he also appeared to have relied on incorrect or second-hand information. For example, Brawley stated that he and the other first year fellows in 1988-9 saw “half of all adrenal cancers diagnosed in the United States.” This is an exaggeration. In the late 1980s, there were about 3-400 cases of adrenocortical cancer diagnosed yearly in the U.S. Our publication (RV La Rocca, et al., J. Clin. Endo. Metab., 1990 71:497-504) reported on only 17 of them, who were seen over a period of several years. Anyone can do the math.
After I wrote the treatment protocol, LaRocca and I shepherded it through the NCI IRB. In mid-1986 or thereabouts, we began accruing patients with metastatic adrenocortical cancer. These were patients with an aggressive disease that carried a grim prognosis. Most patients were willing to try just about anything. Within a short time, to our amazement, we saw our first objective disease responses. (Unfortunately, there were only 2PRs in 16 evaluable patients, and two other mixed responses in this trial. Five patients had stable disease for up to 10 months).
In the early going, it was all very exhilarating, until we also began to observe some of the novel toxicities of the drug. Suramin is a potent inhibitor of several lysosomal enzymes that are responsible for the digestion of glycosaminoglycans. These heparin-like molecules tend to accumulate in tissues, particularly the cornea, causing suramin keratopathy. This resulted in eye irritation that could require artificial tears, patching, and might require cessation of suramin dosing.
Glycosaminoglycans, particularly heparan and dermatan sulfates, can also circulate, functioning as inhibitors of thrombin (Horne, et al., Am J. Hematol., 1988 71:273-279) to produce anticoagulation. Thus, it’s difficult to understand Brawley’s comments about Silvio Conte, a former member of the U.S. House of Representatives who had prostate cancer and was treated with suramin. I never knew the man, because as of June, 1989 I was no longer a member of the suramin team. (Yep, the craziness finally caught up with me. You can get a sense of how crazy it really was from the pages of “The Magic Bullet.”)
Brawley states that Conte had a blood clot in his brain and linked it to his suramin treatment. At same time, Brawley correctly stated suramin is an anti-coagulant, which prevents clotting. It seems more likely to me that the well-known pro-coagulant effects of uncontrolled prostate cancer led to the thrombosis in Mr. Conte’s brain.
Ultimately, three patients with adrenocortical cancer and one other patient with a lymphoma also developed a “reversible” polyneuropathy that was reminiscent of Guillain-Barre syndrome (R. LaRocca, et al., Neurology, 1990 40:954-960). However, by no means were all these recoveries complete, and rehab could be difficult and time-consuming.
Should the trial have been stopped after the first two patients with adrenocortical cancer developed the Guillain-Barre-like syndrome? Some thought so at the time, because one day an FDA inspector wearing a bright, shiny silver badge came to have a chat with us.
Somehow, Snuffy contrived not to be present, so two edgy fellows, LaRocca and myself, defended the suramin trials before one badass FDA inspector. As I recall, we managed to convince him that with Myers’s new dosing regimen, which required continuous infusion of drug as opposed to bolus dosing, peak plasma suramin levels would remain below 300 µg/mL. This seemed to be the plasma level above which most suramin toxicity developed. The suramin trials were allowed to continue.
I was badly affected by the suramin-induced polyneuropathy, and still remember the patients afflicted by it. But Myers, one of the best of the best, was the protocol PI, and neither he nor Sam Broder supported ending the trials. LaRocca and I trusted their judgment. Besides, it was only a question of breaking a few eggs, right? So, despite the polyneuropathy, not only were the trials continued, but treatment with suramin was opened-up to a general phase II trial.
Early on, we found a response in a patient with prostate cancer, and began a new trial in that disease, in which Mr. Conte participated. This is when the NCI made what was, in my opinion, a serious error. Myers, LaRocca and I were well-aware of the flaws in the prostate cancer trial. We knew we’d need some time to make sense of the data we had and to fix the problems we’d run into. For instance, the assay required to determine serum suramin levels was cumbersome and beyond the resources of many centers.
No one had cared about our work in adrenocortical cancer. But once those first few PSA and objective responses were seen in metastatic castrate resistant prostate cancer (mCRPC), all hell broke loose. We were told to provide information about the trial to anyone who asked. And so, investigators, like Steve Cvitkovic, would barge into my office with a set of demands that had to be filled immediately. The inevitable result would be disappointment, which often happens when drugs are pushed ahead too rapidly in trials.
Who made the choice to disseminate the trial? I’m not certain, though I doubt whoever did so had much of a choice. In those days, we had few good treatments for mCRPC. Rich and powerful men were dying from it in pain, in misery, and in droves. I think the oncocrats wanted the answer quickly and didn’t want to wait around for years until we figured it out.
Magic bullet for prostate cancer?
In our original prostate cancer trial, a total of 38 patients were accrued. Six of 17 had eventual objective disease PRs, and 21/38 (55%) experienced a decrease of greater than or equal to 50% of their PSA. Ten patients had a decrease of greater than or equal to 75%. This translated into dramatically increased overall survival. These numbers, if validated, compare favorably with contemporary prostate cancer treatments.
Unfortunately, Brawley misstates the reasons for suramin’s apparent early clinical success. The observed responses were not due to the simultaneous withdrawal of non-steroidal anti-androgens (e.g., flutamide). These drugs can indeed produce both PSA declines and objective responses in patients with mCRPC. The withdrawal phenomenon was first brought to my attention by Jayne Gurtler, of Metairie, LA, who noted a six-month response in one of her patients after withdrawal of bicalutamide. We were thus aware of this possibility before the trial in prostate cancer began.
Also, despite Brawley’s claims, we were also aware of the extremely long plasma half-life of suramin, and the correlation between toxicity and peak plasma suramin levels. This is what Myers’s dosing scheme (see below) tried to modulate. Regardless, just what were these men with advanced prostate cancer trying to accomplish in a clinical trial at the NCI, if their cancer was still sensitive to standard flutamide treatment?
I passed the information from Gurtler on to Myers, who I was told sent it along to Howard Scher at Memorial Sloan Kettering. Scher and his group performed a large trial that validated Gurtler’s eagle-eyed finding. However, it is possible that the hydrocortisone that men received because of the damage suramin caused to the normal adrenal cortex produced some of the responses.
71% of those with prostate cancer who had pain (21 patients) experienced sufficient pain relief to stop or reduce their use of opiate analgesics by one-half (C. Myers, et al., J Clin. Oncol., 1992, 10:881-889). Pain relief was the indication that Warner-Lambert sought FDA approval for in 1998.
However, two patients (5%) died of drug-related toxicity, one of DIC, the other of sepsis. Under Myers’s dosing scheme, peak suramin plasma levels did not exceed 300 µg/mL, and neurologic toxicity was limited to grade 1 or 2 peripheral sensory neuropathy. Nevertheless, many cancer pharmacologists found this dosing scheme peculiar, as Brawley points out; I know at least one who said it made no sense at all. Perhaps in an academic sense it didn’t. But the results speak for themselves.
As Brawley notes, the FDA turned Warner-Lambert down. They were correct to do so—the drug was too toxic. After that, suramin pretty much disappeared. Long before, the members of suramin team had gone their own ways. It was an ugly divorce caused by the usual typical human emotions—greed, envy and fear.
“The Magic Bullet” is dated now, though you can still purchase it used on Amazon. But it’s still kind of funny for such a dismal subject as cancer, if you read it as a character study.
As a final swipe at Myers, Brawley contended that “the change in PSA—which Snuffy believed to be an indicator of clinical improvement—isn’t worth much. It’s a measurement made in a lab. It doesn’t necessarily mean anything to an actual patient. Your PSA could go down, and you may not know the difference.”
True, the FDA does not accept PSA as an endpoint for the evaluation of treatment efficacy in prostate cancer. Nevertheless, we now know that in mCRPC, PSA declines produced by drug treatments are predictive for prolonged overall survival. Patients whose PSAs decline >50% after treatment will survive significantly longer than those who do not reach that threshold; patients whose PSAs decline >75% will survive longer yet. And those patients whose PSAs decline >90% do the best of all. These patients may survive many years. As any medical genitourinary oncologist can attest, in clinical practice, PSA matters a hell of a lot.
Looking backwards three decades, I wonder if the light suramin cast was worth the candle. In my opinion, probably not. But then, I remember the patient I was introduced to at the FDA ODAC hearing in 1998. He was treated with suramin at Johns Hopkins in 1990 with no toxicity and no evidence of disease since, a span of what was then eight years. When my role in the development of this drug was related to him, he couldn’t have been more gracious and appreciative.
I was delighted to see that at least someone had benefitted from our hard work and stress, and of the willing sacrifices made by so many others.