Articles

A Universal Anti-Cancer Drug?

We frequently deal with fraud and quackery on this blog, because part of our mission is to inform the public about such things, and also they are great examples for explaining the difference between legitimate and dubious medical claims. It is always our goal not just to give a pronouncement about this or that therapy, but to work through the logic and evidence so that or readers will learn how to analyze claims for themselves, or at least know when to be skeptical.

One skepticism-inducing red flag is any treatment that claims to treat a wide range of ailments, especially if those ailments are known to have difference causes and pathophysiologies. Even claiming that one treatment might be effective against all cancer is dubious, because cancer is not one disease, but a category of disease. We are fond of pointing out that there are many types and stages of cancer, and each one requires individualized treatments. As an aside, it is ironic that CAM proponents often simultaneously tout how individualized their treatment approach is, but then claim that one product or treatment can cure all cancer. Meanwhile they criticize the alleged cookie-cutter approach of mainstream medicine, which is actually producing a more and more individualized (and evidence-based) approach to such things as cancer.

In any case – my immediate response to any article or website claiming to treat most or all cancer is to be highly skeptical, but I reserve final judgment until after I read through the details. What kinds of evidence are being presented to support the claims, and what are the alleged mechanisms of action? Are those making the claims being cautious like a scientist should, or are they being promotional like a used-car salesman?

A recent study claiming a potential treatment for many types of cancer has been making the rounds. The title of the article being circulated is, One Drug to Shrink All Tumors. What made me take immediate interest in this article was that it was not on a dubious website, sensational tabloid, or even mainstream news outlet, but on the news section of the American Academy for the Advancement of Science (AAAS) website. This is a report of serious medical research. The title, I suspect, is perhaps a bit more sensational than it otherwise would have been because of a geeky nod to the “one ring to bind them all” Lord of the Rings quote. Regardless of the source and the headline – what is the science here?

What scientists have discovered over the last decade is that some blood cancer cells (leukemia and lymphoma) express on their surface a protein called CD47. This protein acts as a signal to the immune system and prevents it from attacking the cancer cells. The same CD47 protein is expressed on normal blood cells and keeps them safe from the immune system as they circulate around the body. It makes sense that those neoplastic cells that by chance express CD47 are the ones that will survive and develop into a cancer. Those that do not express the protein are probably weeded out by the immune system. Cancer cells are therefore the Darwinian survivors, although not beneficial for the host.

The AAAS is now reporting on a new finding of Dr. Irving Weissman, who made the original CD47 discovery:

“What we’ve shown is that CD47 isn’t just important on leukemias and lymphomas,” says Weissman. “It’s on every single human primary tumor that we tested.” Moreover, Weissman’s lab found that cancer cells always had higher levels of CD47 than did healthy cells. How much CD47 a tumor made could predict the survival odds of a patient.

Not exactly all tumors, but every one tested so far. Still, this is an amazing discovery. It seems that expression of CD47 is a universal or near universal property of all cancer cells, perhaps required for their ability to evade the immune system long enough to become clinically detectable. Weissman and his team then took the next logical step, they created anti-CD47 antibodies and then tested their activity against tumors. They transplanted “human breast, ovary, colon, bladder, brain, liver, and prostate tumors” into mice, and then compared untreated mice to those treated with anti-CD47 antibodies. What they found is that those mice that were treated generally experienced shrinkage of their tumor and a marked decrease in the risk of the tumor spreading. This is not the same thing as a cure, but any treatment that shrinks tumor and prevents metastasis is a useful treatment against cancer and is likely to prolong survival, decrease recurrence, and increase the probability of a cure.

Our experience with past treatments that show similar promise is that they turned out not to be outright cures, but those that worked became useful additional treatments that improve outcome in some cancers. If anti-CD47 antibodies pan out as an effective treatment, it is likely that they too will be added to the anti-cancer armamentarium, but will not be cures by themselves.

So far, so good – we have what sounds like a very plausible mechanism, and reasonable animal data. Further, the scientists express the proper degree of caution and skepticism toward their own research. They point out that anti-CD47 antibodies also allow the immune system to attack healthy blood cells, causing anemia. The decrease in blood counts was short-lived in the mice, but obviously this will be something to watch carefully in human trials. The also report:

“The microenvironment of a real tumor is quite a bit more complicated than the microenvironment of a transplanted tumor,” he notes, “and it’s possible that a real tumor has additional immune suppressing effects.”

Another important question, Jacks says, is how CD47 antibodies would complement existing treatments. “In what ways might they work together and in what ways might they be antagonistic?” Using anti-CD47 in addition to chemotherapy, for example, could be counterproductive if the stress from chemotherapy causes normal cells to produce more CD47 than usual.

Exactly – animal models are great preludes to human studies, but they are not the same as human data. Biological systems are complex, and we cannot predict what factors may affect the response of a native human tumor to this treatment based upon evidence from transplanted tumors. Further, there are other factors that may impact the net effect of this treatment in humans with cancer, such as interactions with chemotherapy or other host responses.

Therefore, there is no substitute for controlled human trials. We cannot and should not base claims of efficacy, or even safety, on a plausible mechanism and encouraging animal data. What we have so far is sufficient justification for phase I clinical trials in humans, and nothing more.

This is a point worth emphasizing, as it is very common for dubious treatments and products to base clinical claims on preliminary basic science or animal data, often backed up by nothing but anecdotes. Together this kind of evidence can make a compelling story to the non-scientist, and that’s what the marketers are counting on. It is helpful to see how real scientists approach the evidence, and the process that is necessary before we can know that a treatment is safe and effective.

Anti-CD47 antibodies are definitely a promising discovery and a treatment to watch. I hope the research pans out and we can add this to our available treatments for cancer. It is possible, even probable, that (if the treatment proves safe and effective) in the future we may test cancers for the presence of CD47, and then then give the treatment only to those who express this protein – further individualizing cancer treatment. Human studies are likely to take 5-10 years, however, and that’s if everything goes well.

Posted in: Cancer

Leave a Comment (21) ↓

21 thoughts on “A Universal Anti-Cancer Drug?

  1. cervantes says:

    We’ve been through this sort of thing many times. An advance in understanding, and maybe some encouraging results in animal models, gets translated by the mass media into a world-altering breakthrough. Remember Judah Folkman and angiogenesis? Yes, he labored in the wilderness for a long time until people agreed, oh yeah, tumors have to recruit a blood supply, whereupon the New York Times reported that the cure for cancer was just a few years away. Then it turned out that angiogenesis inhibition is of some use as an adjunct therapy in some circumstances, but it’s really just a small incremental advance in treatment.

    The media love the medical superhero stories, but they’re pretty rare in reality and there’s a very long way to go between a laboratory observation and a useful medical intervention. When it does arrive, it’s seldom what it was cracked up to be. Sorry for being a downer, but that’s how it is.

  2. David Gorski says:

    Yep. Having been in the biz since the 1990s, you’ll forgive me if I’m a bit jaded. This reporting is a bit too much hype. cervantes nails it with the comparison to Judah Folkman. Back in the mid to late 1990s, the argument was that, because nearly all solid tumors induce angiogenesis, antiangiogenic therapy would be an amazing advance that would allow us to turn cancer into a chronic disease by treating the tumor blood vessels, which, because they are normal cells, do not develop resistance to antiangiogenic therapy. Of course, it was a lot more complicated than that, and, as we all know, antiangiogenic therapy has not turned out to be a magic bullet of any sort. And, to be fair, Judah Folkman never claimed that it would be, his astonishing results in mice notwithstanding. Before angiogenesis, it was immunotherapy (that’s what targeting CD47 is, by the way), which has waxed and waned in popularity as the new “cure for cancer” since at least the 1980s, which was when I was in medical school and started paying attention to such things.

    Here’s the problem I see. CD47 isn’t just expressed by cancer cells, but by a lot of normal cells too. True, it’s expressed at higher levels on cancer cells, but it’s still expressed on normal cells. Here’s the critical thing from the article:

    Although macrophages also attacked blood cells expressing CD47 when mice were given the antibody, the researchers found that the decrease in blood cells was short-lived; the animals turned up production of new blood cells to replace those they lost from the treatment, the team reports online today in the Proceedings of the National Academy of Sciences.

    That will be the key question as targeting CD47 is moved from mice to humans: Will it result in widespread hemolysis in humans and will human bone marrows be able to turn up the production of various blood cells lost due to treatment? In other words, what will the toxicity be? It might well be more toxic in humans.

    Be that as it may, it’s an interesting finding and quite promising. By fortuitous coincidence, I’ll be attending the meeting of the American Association for Cancer Research (AACR) in Chicago beginning this weekend. I’ll definitely keep a look out for talks and posters on CD47, to see what the latest findings are. Also by bizarre coincidence, for Monday I was thinking of updating my post about “why haven’t we cured cancer yet?” given that there have been some developments that suggest that “personalized” therapy for cancer will be far more difficult than even its biggest boosters had thought. This might make a nice counterpoint, to prevent my post from being too much of a downer. :-)

  3. mdstudent says:

    @ cervantes

    “When it does arrive, it’s seldom what it was cracked up to be.”

    Hence the cautious and skeptic attitude taken by the researchers which Dr. Novella is highlighting.

    Fair enough, the title of their article is a little naive but for all we know they just finished watching the new Hobbit trailer.

  4. mdstudent says:

    @ Dr Gorski

    “That will be the key question as targeting CD47 is moved from mice to humans: Will it result in widespread hemolysis in humans and will human bone marrows be able to turn up the production of various blood cells lost due to treatment? In other words, what will the toxicity be? It might well be more toxic in humans.”

    I was thinking the exact same thing. It’ll definitely be interesting to see how human trials turn out and ultimately compare to current standards of care.

    Cell biology is fascinating.

  5. David Gorski says:

    The comments after the news story are rather depressing, too. Various quacks and trolls have appeared, including Burzynski supporters, the “cannabis cures cancer” troll (or someone like him), “cancer industry kills” trolls, and the like.

  6. I agree – the anti-angiogenesis story is a perfect analogy. These breakthroughs that all sound like they are going to cure cancer, at best turn into a reasonable adjunct and incremental advance. That is what I predict for this treatment – if the toxicity isn’t a deal-breaker.

    The reporting is typical, kind of a hybrid science/lay reporting. The headline is provocative and hype. The beginning the article focuses on the hype, but then at the end we have a more realistic assessment. This is better than a mainstream news outlet would do, but not as dry as a technical paper.

    Still – it’s a good way to point out the scientific, vs the quack, approach.

  7. ConspicuousCarl says:

    Indulge me a bit here. Without knowing much in detail, I don’t even get the logic of this proposed therapy.

    If CD47 is expressed by cancer cells in order to disguise themselves as healthy cells, how is it helpful to notice this, and why would anti-CD47 drugs be considered anything but toxic to the patient’s normal body? It sounds as if someone noticed that terrorists are infiltrating our military bases by wearing US uniforms, therefore we should build robots which kill anyone in a US uniform? There seems to be a missing hypothesis on how CD47-killing therapy is supposed to only kill cancer cells.

    I notice that they said the negative effect on healthy blood cells was temporary in mice. I would guess that there could be a bunch of reasons why something designed to kill CD47-endowed human cells simply doesn’t work in mice.

  8. mdstudent says:

    @ conspicuouscarl,

    “Moreover, Weissman’s lab found that cancer cells always had higher levels of CD47 than did healthy cells. How much CD47 a tumor made could predict the survival odds of a patient.” -From the article

    Because tumor cells express more CD47 than normal cells they are expected to be more likely to get caught and destroyed. In theory.

    Imagine the terrorists, in their enthusiasm to fit in, put on too much uniform and the robots were programmed to kill the most uniformed soldiers first, ideally the terrorists.

  9. jli says:

    Here’s the problem I see. CD47 isn’t just expressed by cancer cells, but by a lot of normal cells too. True, it’s expressed at higher levels on cancer cells, but it’s still expressed on normal cells.

    This is where I see a potential problem in using anti-CD47 in humans. It is kind of analogue to the use of anti-EGFR. They work wonders in some non-mutated KRAS/BRAF cancers, but they are certainly not without toxic side effects attributable to disturbance of the normal function of the EGFR-pathway.

    We don’t know if the tolerable dose will be high enough to be damaging to cancer cells (if it works in patients). One could speculate that over expression of CD47 in cancer cells could be the very reason that toxic doses wouldn’t be enough to be effective.

    What they found is that those mice that were treated generally experienced shrinkage of their tumor …..

    This often understandably works as a persuasive argument. I couldn’t find the reference in a quick scan on Pubmed, but it reminds me of this paper: http://www.nature.com/bjc/journal/v84/n10/pdf/6691796a.pdf where the researchers looked at the results of clinical trials on substances that reduced tumor bulk by at least 60%. Even this large tumor shrinkage in mice weren’t very good predictors of effect in clinical trials.

  10. Roadstergal says:

    If I had a dollar for everything that worked great in mice and didn’t pan out in humans, I’d be a 1%er. I could cure mice of all kinds of xenograft tumors and autoimmune diseases with agents that work a treat in mice, but are either too unsafe or too ineffective for human use.

    Yes, it’s a cool idea, but it’s utterly preliminary at the moment. Sure, it might turn out to be safe and efficacious in humans – and if it does, then we can start crowing about its therapeutic potential, not before.

  11. gippgig says:

    One thing to try is to link anti-CD47 antibodies to a toxin that is only active as a multimer (for example, as I recall, some pore formers). The higher density of CD47 on tumor cells would greatly increase the chance of forming the active toxin multimer resulting in selectivity for the cancer cells.

  12. nybgrus says:

    @conspiciouscarl:

    mdstudent gave a pretty good analogy. To extend it a bit, instead of thinking of the terrorists wearing US military uniforms, think of them trying to infiltrate as 5 star generals with lots of medals.

    Since the cancer cells have a higher level of CD47 they will bind a higher proportion of the anti-CD47 Ab. Also, due to their cancerous nature, they will be more antigenic than regular cells. So the theory is that they will be preferentially targeted by the immune system and then hit harder.

    So the robots go after the most highly decorated soldiers and then blast them with extreme prejudice. Those that are less decorated only get a flesh wound or two as a test.

    And that is why the toxicity is important – will the differential toxicity in human clinical trials be enough to make it a useful treatment? I’d guess not. But combined with other therapies in smaller doses it may prove a useful adjunct.

    Gippgig’s comment is an interesting one, though I think that it would suffer from much the similar issues whether the multimeric toxin is there or not. Either regular cells will suffer the same fate, or the multimer will have to be large enough to only select for the highest expressors of CD47, thus rendering it useful only for a certain subset of cancers (which is still not such a bad thing).

  13. BillyJoe says:

    My wife recoils in disgust at such “side-effects”.
    In fact I feel a little nauseous myself.

  14. Scott says:

    I really hope the banhammer hits Franky soon. If posting piles of irrelevant comments to tout a commercial website isn’t spam, I don’t know what is.

  15. Scott – good pickup. Definitely spam, now banned and comments removed. Thanks.

  16. David Gorski says:

    D’oh! How’d I miss that?

  17. cervantes says:

    You’ll need a subscription, but if you have one, this is highly relevant to this post: Drug development: Raise standards for preclinical cancer research. C. Glenn Begley & Lee M. Ellis

  18. ConspicuousCarl says:

    mdstudent and nybgrus:

    Thanks, I missed that line. It will be interesting to see if some in vitro testing of cancerous human vs healthy human cells shows a difference, but until that happens I am now wondering if there is any prior reason to suppose that those differences in total CD47 amounts is more likely to explain the mouse results than just minor protein sequence or folding differences between the mouse and human forms, which I thought were always at the top of the list for why a drug might not work the same in different animals.

  19. Earthman says:

    “Using anti-CD47 in addition to chemotherapy, for example, could be counterproductive if the stress from chemotherapy causes normal cells to produce more CD47 than usual.”

    From my experience of chemotherapy, the last thing I would want is to have anaemia on top of all the other nasty side effects.

Comments are closed.