Skip to main content


Antibiotics: Not As Easy As They Say

Since we were just talking yesterday about antibiotics and the misconceptions that people have about them, this is an appropriate time to take a look at an article in the popular press about antibiotic discovery. ( counts as “popular press”, for sure). The article focuses on the work of Brian Murphy at UI-Chicago, who’s doing marine natural products research, especially on potential antibiotics from marine/freshwater microorganisms. (Update: Prof. Murphy has shown up in the comments to this post, and he wasn’t very happy with the way the article turned out). That’s certainly a reasonable field of research – soil microorganisms have been large tapped out for this sort of work, and the number of marine candidates is surely huge (although difficult to sample and to culture). The article itself is a mixture of accurate reporting and some basic misconceptions:

“It’s a huge gamble,” he said. “We look for unique environments, and we just have to hope that the evolutionary pressures driven by the challenges of trying to survive in these conditions will yield microorganisms which can produce new drug leads. But we have no idea what we’ll find.”

The costs of such ventures, which have taken him across the globe from Thailand to Iceland, can stretch to tens of thousands of dollars. And with this brings pressure. Any organization willing to stump up such money will demand returns on their investment, but nature doesn’t always play ball.

“Such money”? A few tens of thousands of dollars is a roundoff error too small to see in drug R&D. No one’s glaring away, awaiting a return on their thirty-thousand dollar investment in plane tickets and scuba gear. They’re worried about their tens-and-hundreds-of-millions of dollars, the sort of investment that any real compound has to lead to. The article also quotes from William Fenical at Scripps:

Fenical, the professor at Scripps Institute of Oceanography in San Diego, is one of the original marine biomedicine pioneers. He began investigating the ocean’s potential to provide cures for disease in the late 1960s, long before anyone considered it as a science at all.

That’s not accurate, either – Vice’s reporter is laying it on thickly there. Natural product-based drug discovery has been going on for a long time – it was the first sort of drug discovery there was. Marine natural products have been recognized as active drugs ever since antiquity, just like their terrestrial counterparts.

In 2013, Fenical made one of the most interesting antibiotic discoveries in recent years off the coast of San Diego. A bacteria living in the sediment on the Pacific Ocean floor was producing a compound called anthracimycin. Fenical soon found it was capable of attacking the bacteria MRSA, a hospital superbug which is notoriously difficult to treat.

But in many ways, discovering antibiotics is the easy bit. Finding someone interested in investing in developing them is a far greater challenge. Two years have gone by since Fenical identified anthracimycin and no one has shown any interest in taking it from the research lab to the clinic.

That’s worth a closer look. Anthracimycin is an interesting compound, but the article should have mentioned that one of the things that makes it interesting was that it was one of the few natural-product antibiotics discovered in recent decades (although that doesn’t quite fit with the untapped-bounty angle that the rest of the article takes). Cubicin (daptomycin) was discovered in 1986, and hit the market in 2003, and that’s been it for a while. Anthracimycin has what certainly looks like useful activity against various resistant strains of S. aureus,

“We’ve discovered six antibiotics in the recent past,” Fenical said. “Of those, three to four have serious potential as far as we know, including anthracimycin. But we have no way to develop them. There are no companies in the United States that care. They’re happy to sell existing antibiotics, but they’re not interested in researching and developing new ones.”

That line about how there are no companies in the United States that care isn’t accurate, either. It’s true that many of the larger companies have, over the years, pulled out of antibiotic research. But it’s not for lack of “caring”: it’s for lack of ever finding anything. GlaxoSmithKline had a huge investment in new antibiotics, as did AstraZeneca and others. They must certainly were “interested in researching and developing new ones”, and they put their money down and large teams of researchers on it. GSK came up completely empty, after many years of work, and AZ drilled a lot of dry wells, too (although they’ve spun off their anti-infectives research into a separate company).

The usual line is that companies don’t care because they can’t make any money when they do find a new antibiotic, so why bother? But (as just mentioned) there have been large companies pouring money into it, which they don’t do unless they think that they can make it back. Another case in point is Merck’s buyout of Cubist, the developers of daptomycin. They paid over $8 billion dollars to do that deal, which shows that there is indeed money in the field – if you can find something to sell. That’s the hard part, and that’s why companies have been pulling out.

This article is conflating an apparent lack of interest in developing anthracimycin (so far) with the difficulties in developing any new antibiotics at all. So why isn’t anthracimycin going forward? One reason may be that the compound has no activity against gram-negative bacteria, which is one of the biggest unmet needs in the field. A chlorinated version, which makes it similar to a known soil-organism natural product, chlorotonil A, gives it some gram-negative activity, but not much. Another tricky part is patent coverage and IP. There’s an earlier Japanese patent application (PCT/JP 2011/010586) that appears to have the compound in it, for example. And another problem is availability. A five-liter fermentation of the marine organism gave about 6 milligrams of anthracimycin, which means that getting enough to work with is going to take a significant investment. To turn this into a drug, you’re going to need enough to scoop it around like flour out of a bag.

Back to the Vice article:

Fenical’s frustrations stem from the fact that developing new antibiotics is in some ways far easier than developing treatments for other diseases such as cancer.

The main way in which it’s easier – and it’s a big one – is that the assays and animal models are far more predictive. But having lots of good means to develop a compound does not help you when you can’t find much chemical matter to develop, or many good targets to develop anything against, and there cancer (and many other therapeutic areas) have raced ahead. The huge majority of existing antibiotics work through a very short list of mechanisms, as opposed to (say) cancer, where there are a wide variety of targets that have been prosecuted at all stages of the disease (with more in the works). When you do big natural-product screens for new antibiotics, then, you mostly find nothing, and what you do find is almost always something that was already known, or works against a target that’s already known. I’d say the biggest interest with anthracimycin is in figuring out how it works – does it hit something that we already hit? Or if it has a new mechanism, then the thing to do would be to work up some targeted screens and see what else you can find.

The article mentions another natural product discovery by Fenical and his group, diazonamide A, and that’s worth looking at in light of just those issues. The compound showed promising anticancer activity, but was extremely difficult to isolate in any quantity from its natural sources. That led to a lot of total synthetic work (there have been several total syntheses thus far, but needless to say, none of them can provide the compound in development-level quantities). Work on the compound’s mechanism of action, though, showed that it has an interesting microtubule effect (probably through ornithine delta-amino transferase), and this mechanism is being studied further as an oncology target. A small company has been formed to try to develop a simpler version of the diazonamide structure itself as a new oncology agent as well, which is something you won’t get from the Vice writeup, either.

Producing antibiotics, on the other hand, is far less environmentally invasive. All scientists need is to collect a few cells of the antibiotic-producing bacterium, which can then be cultivated en masse, enabling the production of the chemical on an industrial scale.

Unfortunately, that’s the sort of thing you will get from the Vice article. And as anthracimycin itself demonstrates, that’s not always how it works. When you find an active natural product, sometimes you can’t even identify the organism that’s making it. And many of those can’t be cultured in the lab. Even when you do, the bacterium or fungus responsible might not make very much of it under lab conditions, or it might not make very much of it under any conditions at all, no matter what. This is not the trivial problem that the above quote suggests.

The article does make some good points, and it does get some things right. But what it leaves out and what it gets wrong make it hard for a lay reader to figure out what’s really going on in antibiotics, and with natural product drug discovery in general.

19 comments on “Antibiotics: Not As Easy As They Say”

  1. Brian Murphy says:

    Excellent commentary Derek. This is [the real] Brian Murphy speaking, and I was quite upset when I saw this article, as the author’s criticisms of industry de-investment of antibiotic discovery were quite naive. I was also upset that I never saw a proof of the article and was not allowed to comment, a mistake that I will not make again. Keep up the good work.

    1. anon says:

      This is the price of using “science” journalists as agents to spread the word about your research.

      1. Ran says:

        This is not a very smart comment. Some journalists are doing a good work in communicating science to the public and reveal to taxpayers what their money is being used for in the academia.

        Journalists also make mistakes, sometimes out of ignorance. Scientists too.

  2. bagger vance says:

    Wow, Harran solved a structure! So that’s where his license to kill comes from.

  3. Anon says:

    “This is [the real] Brian Murphy speaking,..”

    Is it just me, or does that come across as slightly arrogant given that there are hundreds or possibly thousands of Brian Murphys listed on LinkedIn? Must be all fakes…

    1. Ed says:

      No, it doesn’t come across as the slightest bit arrogant. It is a simple way of identifying himself as the individual mentioned in the article.

      1. great unknown says:

        It also emphasizes that the article misappropriates his name to make points he disagrees with, and thus presents [a fake] Brian Murphy.

  4. Z-squared says:

    Is it worth mentioning the recent events at Tetraphase in the context of this commentary?


  5. Soon-to-be-former BMS person MOLS says:

    BMS was among the companies that made a huge investment in microbial genomics back in the 1990s, and learned the hard way that target ID wasn’t the rate limiting step.

  6. ScientistSailor says:

    A few point from a caring and dedicated antibacterial researcher…

    1) the assertion that soil bacterial are a depleted source of natural products has recently been debunked by Kim Lewis at Northeastern. Using new technology to grow the microbial ‘dark matter’ he isolated a new natural product, with a completely new mechanism of action:

    2) Just looking at the abstract of the Chlorotonil paper says it all:

    The poor solubility of 1 in both organic solvents and water again raises the old question of why bacteria create such complex metabolites.

    Granted, I don’t know what the actual solubility is, but if is so low as to make the abstract of the paper, I believe that would rule it out as and antibiotic. It would also be a big strike against it even as a lead. Remember that antibiotics need massive (by industry standards) exposures to be effective. The lowest-dose Gram- drug I know of is levofloxacin, which is given 750mg QD; some of the beta-lactams are given 2g TID!

    3) The real problem here is the lack of lead chemical matter against Gram-negative organisms. MRSA is not as difficult to treat or as scary as people make out. Linezolid (soon to be generic) and dapto work well against this bug. We’ve been using LNZ for 15 years now, and don’t see wide-spread resistance.

    On the other hand, the most recently approved Gram-negaitve antibiotic, Avycaz, already has resistance mechanisms out there. A bit too late for Halloween, but this is one scary story, don’t read right before going to bed:

    Furthermore, there are now strains of Klebsiella pneumoniae that are resistant to ALL available antibiotics, So the assertion that

    “Before long the world may be faced with a situation last seen in the pre-penicillin era when even the most minor infections, such as those resulting from a child’s grazed knee, could prove life-threatening, and every operation was fraught with danger.”
    is wrong, we are already there!

    I guarantee you that if someone has an interesting lead against a G- species, especially CRE, P. aeruginosa or A. baumannii, we will be interested, and will care enough to take a look.

    4) Regarding the TTPH, their recent result supports Derek’s analysis about the predictability of the pre-clinical models in the ABx space. Many of us saw that result coming years ago, the exposures they obtained upon PO dosing weren’t even at the MIC of the bugs they were trying to kill…

    5) Another falsehood:
    This remains the case today, but there are now no national programs aimed at tackling drug-resistant bacteria.
    BARDA and NIH have significant antibacterial efforts

    etc, etc…

  7. Ellis says:

    Finding new antimicrobials is relatively easy, although the majority act through already exploited mechanisms of action and so resistance may already be prevalent. The problem is that very few infections are multiple antibiotic resistant (for now at least) and thus the patient base is too small to make a reliable pay back on investment. Until we are happy to pay $50,000 for a pill that will save you from certain microbe based death that will not change. And maybe some of the drugs put aside because of side effects will be re-considered if they will cure an otherwise lethal infection.

    1. lynn says:

      Finding chemicals that kill bacteria, especially Gram positives, IS easy. But finding such chemicals that are non-toxic and have suitable pharmacologic properties is hard. Finding anti-Gram negative agents is even harder. Yes, it can be problematic to develop antibacterials, but there are few novel candidates with suitable properties and low toxicity that are not subject to rapid [overnight] resistance development.

  8. Grim Reaper says:

    It is sad that the author, had he had more neurons to fire, might have gotten the story right. Maybe he went to the same School of Journalism as Brian Williams or Dan Rather. The bulk of science stories, particularly about pharma, unreliable.

  9. Andy II says:

    I agree with many of the comments above as I have been in antibiotic discovery/development/licensing business myself. There are still number of clinical developments going on as well summarized at a website (The PEW Charitable Trust): Still a number of study for MRSA and many for gram negative infections. Reflection of the current medical needs, which translate good marketability. Finding new compounds with “noticeable” antibacterial activities against “typical” strains is just the first step. Their are merely potential lead compounds, as most of the readers of this site are very aware of, after the first screening. If there are no “tangible” differentiation factors/benefits expected in the future TPP (target product profile) compared to those of the current SOC (standard of care) or those in the pipeline, pharma companies will just pass the opportunities. If pharma is interested in new treatment option for infections caused by those difficult treat such as KPC, Pseudomonas, Acinetobacter, the registration trials are tough to complete under the current guideline (# of cases with these specific bacteria, # of patients to prove safety and efficacy) though FDA is becoming receptive (?) to sponsor companies:

  10. Dave says:

    Ah, yes, Levofloxacin. Look at the list of side effects it can have. That’s one wicked drug. “spontaneous tendon ruptures and irreversible peripheral neuropathy”?!? “Tendon damage may manifest months after therapy had been completed.”?!? Wow!


  11. ScientistSailor says:

    You don’t need to worry about the side effects of levo, resistance rates among Gram-bacteria are >20% now, so it is rarely used.
    Instead, if you are lucky, you can have an IV carbapenem, or if you are less lucky, colisitin. Or you can die of your infection…

  12. ScientistSailor says:

    Hah! Just in time, a lay-press article that gets it right!

  13. gippgig says:

    …and even more bad news:
    Emergence of plasmid-mediated colistin resistance mechanism MCR-1 in animals and human beings in China: a microbiological and molecular biological study
    The Lancet Infectious Diseases
    doi: 10.1016/S1473-3099(15)00424-7
    The situation is out of control and there is every grounds for panic.

  14. Dave says:

    Ha! AZ’s effort was staffed by that crew from ArQule, hardly a serious go at it.
    They all conned investors into thinking they were the leaders of “automating ” organic chemistry when their VP of chem was so computer illiterate he didn’t even have one on his desk, if you sent him an email his secretary had to print it and hand it to him. They couldn’t even make a shaker table that functioned correctly. Now they have populated Novartis, another phoney big pharma research group.
    The main reason so many chemists are out of work these days is because of such fraud.
    Investors please take note, look at where “esteemed” scientists come from using LinkedIn.

Comments are closed.