Skip to main content

Drug Assays

A New Antibiotic? Yes, Please

New antibiotics against resistant Gram-negative bacteria make me happy, so I’m very glad to see this report from Genentech. They’ve been doing a lot of work on an antibiotic scaffold (arylomycins, Update: on a program that came in whey they bought RQx) which had not thus far really found much practical use, and it looks like they’ve found what could be a valuable drug.

The arylomycins are natural product antibiotics, but they’re not particularly potent, and have found no use in the clinic. That’s a shame, because their target is of great interest: signal peptidase. Those are  key enzymes in an important bacterial signaling pathway, and inhibiting them should lead to all sorts of useful effects. It’s a funny enzyme, though – it uses a catalytic mechanism (serine/lysine dyad) that you don’t see much, and which traditional serine-active-site targeting compounds don’t seem to hit. I should mention that I have personal (unpublished) experience with things that don’t hit signal peptidase I, and I can testify to what a difficult-to-target enzyme class it is.

So any chemical matter that does hit this is worth following up on, but it’s been a hard slog (as it is with all novel antibiotic targets). The Genentech team have put in the work, though, and have extensively modified Arylomycin A while keeping its (essential and funky, for a natural product) biphenyl-peptide-macrolide core, in keeping with clues from the known crystal structures of inhibitor-bound enzyme (see the structures at right). Here’s a writeup at Chemistry World if you don’t have access to the Nature paper, which will give you the main points. The really good news is that the new compounds are potent and efficacious, even against nasty resistant clinical isolates. That’s probably because they’ve never been punched right in the signal peptidase pathway before, and encouragingly, it seems that resistance itself emerges only slowly on repeated exposure.

What comes across as you read the paper, though, is the empirical nature of antibiotic discovery. The peptidase enzyme that G0775 targets is on the inner membrane of the Gram-negative bacteria, so that means you have to get through the outer membrane first. Experiments with defective mutant bacteria indicated that the molecule could actually do that, although you wouldn’t have expected much given its structure. Most antibiotics make it into bacteria through porins, but G0775 apparently manages to bypass that route, and it’s not quite clear how. That alone is definitely worth further study – it’s known that that classic polymixin antibiotics can do this, but G0775 is active (fortunately) against strains with membrane mutations that make them polymyxin-resistant. So there’s something quite interesting going on, and it’s not like we know how to design for it. Also worth noting is that it would seem that G0775 not only makes it in, but manages to not get pumped back out very efficiently either. That’s a common problem with antibiotic lead compounds, and designing away from that is pretty much a black-box process, too.

So the G0775 story represents a great deal of hard work, expertise, and the use of all the tools we have at our disposal in modern drug discovery. But it also has a bit of luck in it. But remember what one of the fathers of bacteriology – Pasteur – had to say about luck in research: “Fortune favors the prepared mind”. And if you’re going to have a successful antibiotic research program, you have to be ready to try a great number of approaches, being as diligent and intelligent as you can, and be ready to run with whatever works. G0775 works, and I hope that its progress in the clinic goes well.

33 comments on “A New Antibiotic? Yes, Please”

  1. Dionysius Rex says:

    I heard this program was already terminated due to preclinical tox findings.

    1. Derek Lowe says:

      I hope not – but if so, then we have to hope for some insights into outer membrane entry, I suppose.

  2. John Wayne says:

    The phrase, “they’ve never been punched right in the signal peptidase pathway before” made me laugh out loud in open office space. This led to an explanation about what I was laughing about to my coworkers, then a quick primer on why cocktails of drugs targeting different biochemical mechanisms are important.

    I can be honest with you guys … I think the personification of biochemistry is hilarious. Wouldn’t it be amusing if the concept of evolution had an internal monologue? Ow! My signal peptidase pathway!

    1. UKPI says:

      Same here…made me think of things a former postdoc used to say (or is that you, GB?)

      1. John Wayne says:

        Nope, I guess scientists have our senses of humor twisted into phenotypically recognizable shapes.

  3. ScientistSailor says:

    It would not surprise me (nor anyone else in the space) if there was nephrotoxicity limiting the usefulness of a polyamine like this…

    1. David Edwards says:

      I admit to beng puzzled by this statement, but that’s probably because I lack specialist knowledge of renal biochemistry. Even with that lack of knowledge, however, I’m wondering why a compound wth multiple amine groups would be toxic to the kidneys, given that their primary function is to remove urea (a compound with mutiple amine groups) from the bloodstream. Likewise, I’m wondering why there are no reports of adverse renal toxicity issues with metformin, which is the first line treatment for type 2 diabetes, and which is positively littered with amine groups. I suspect there is an answer to this question, but whether it’s a simple one that can be contained in a post here, as opposed to an answer requiring three years of undergraduate study, is again something I’ll have to find out by posting this!

      1. Dionysius Rex says:

        Probably based on clinical experience with polymyxin + analogs that have a very strong tendency for renal tox

  4. AC says:

    I thought the mechanism of inhibition was pretty interesting too. They installed a nitrile covalent warhead to target the serine, but ended up forming an amidine with the lysine instead. I’ve only come across nitrile warheads in DPP-4 inhibitors (e.g. vildagliptin and saxagliptin) although I’m sure there are more examples out there. With DPP-4 the nitrile forms a reversible covalent adduct with the active site serine via a Pinner reaction, but DPP-4 has a more typical Ser-His-Asp catalytic triad. I’m guessing with G0775 and signal peptidase, reversible attack of the serine happens as well giving a carboximidate intermediate, but the active site lysine is close enough that it is able to attack this intermediate to give the more stable amidine.

    1. Derek Lowe says:

      It is indeed, and I think that’s another example of good fortune (and trying a lot of things, since I’ll bet that they tried a lot of different warheads. Most people targeting these nontraditional serine hydrolases end up throwing the kitchen sink at them.

    2. Barry says:

      nitrile warheads were tried against various Caspases, in which they form reversible thioimidates with the active-site cysteine.

  5. Mach4 says:

    Those phenoxyethyl amines I bet are key to its activity. See this time and time again where primary or secondary amines slow antibiotic translocation through negatively charged efflux sphincters in bacteria. Is the free phenol version active? Good work Genentech!

  6. lynn says:

    I was quoted in Science about this paper. I said it was a tour de force – but, I also noted to the reporter that “Due to some structural signals, extended toxicology studies will be necessary”. And Paul Hergenrother, in this same Science commentary, noted other obstacles including tox. So, yeah – really neat study…but adding lots of amines to get into Gram-negatives is somewhat worrying.

    1. Mach4 says:

      I’m curious on the membrane perturbation qualities of such antibiotics, Lynn, as molecular microbiologists test against the 4 main mechanisms in bacteria, DNA, RNA, protein and cell wall synthesis.

      In this business we made up the timely maxim “Hit more than one and your done”. The natural product and free phenols look like non-specific membrane perturbation agents, and all microbiologists should be on the lookout for such bad actors, of which we have plenty and growing everyday.

    2. loupgarous says:

      If that was Jon Cohen’s “Scientists engineer a powerful new weapon against antibiotic-resistant bacteria” posted on September 12th, he quotes Hergenrother as being a bit more optimistic:

      ““It’s a really cool story, but the challenge is going to be for them to push it all the way through, and that’s not an easy thing to do,” says Paul Hergenrother, a chemical biologist at the University of Illinois in Urbana who studies the traits that compounds need to penetrate the outer membranes of gram-negative bacteria.

      Hergenrother stresses that, for approval, new antibiotics have to have relatively little toxicity. “With antibiotics, the tolerance for side effects is very low—it’s not like oncology,” he says. But Hergenrother was impressed that the experiments in tests tubes (sic) and mice required only modest doses of G0775 to substantially reduce bacterial load. “That’s just what you want to see at this stage of development,” he says.”

      I’d really like it if the toxicity concerns for G0775 could be overcome, or perhaps have been overstated.

  7. Dr. Manhattan says:

    This is indeed an interesting achievement with a very interesting mechanism of action & inhibition. As to Lynn’s comment (Hi Lynn!), the role of the amines in penetration is somewhat obscure, as they mentioned that they obtained active compounds against wild type E. coli in the absence of all primary amines, although the compounds were not as potent (8 fold less). In fact, the entire question as to how these modified arylomycins get into the bacterial periplasm (the site of signal peptidase I action) is unclear, even to the authors.

    Also, if you look at the Pseudomonas strain data in the Supplementary section there are four 16 ug/ml and one 32 ug/ml out of 12 strains tested. Unless this compound can be dosed up to substantially high levels (again, PK and toxicity become key issues), Pseudomonas is probably out of reach for this compound. Pseudomonas is a perennial problem when it comes to finding active antibiotics that are efficacious. But, they have very impressive inhibition numbers against multiple MDR enterobacteriaceae, including Klebsiella, a major pathogen. I hope the toxicity issues mentioned do not take this program down, as it is indeed a tour de force in antibiotic R&D!

  8. periodic_table says:

    What about the boronic acid beta-lactamase inhibitors? (e.g. Vaborbactam, approved in 2017). those are novel and hit multiple MDR strains, one of them is approved, possibly others en route, and they do not have the tox issues this one may have. Adding multiple primary amines is well known in antibiotic drug discovery as a means to both enhance potency against gram negatives and tox problems. This is definitely some good work but I am not optimistic about success in preclinical / clinical. Hope to be proven wrong!

    1. John Wayne says:

      I thought that molecule was a beta lactamase inhibitor.

  9. loupgarous says:

    Regarding the toxicity issues, is there still a role for arylomycins as topical antibiotics, even if you can’t take them internally? The standard topical antibiotic cocktail (bacitracin/neomycin/polymyxin) in OTC salves may gradually lose its usefulness as antibiotic resistance proliferates in the wild. There may always be a use for it, but when my wife was bitten by a neighbor’s dog (enough to draw blood), the person treating her prescribed mupirocin salve, which may just have been commendable caution, or a precaution informed by local experience with resistant bacteria.

    1. I’d say your wife was pretty lucky.

      Mupirocin works well against staphs and streps found on the skin, and a few gram-negatives, but not against the anaerobes which make up the bulk of bacterial nasties associated with dog bites.

  10. Andy says:

    This program actually originated at RQx Pharmaceuticals in San Diego. Four of the authors on the paper were from RQx. Genentech bought the program from them a few years ago.

    1. Derek Lowe says:

      Just updated the post to reflect that – thanks!

  11. Mfernflower says:

    I feel the left hand side of the Arylomycins are begging for some lipidomemetic organosilicon chemistry rather than the aromatic mess that is in that Genentech molecule

    I am also a little bit weary of the alkylamine modifications they preformed on the molecule

  12. john adams says:

    Do we know more about the mechanism of nephrotoxicity? Would cilastatin be of help here by chance?

  13. Lustig says:

    What is the reason that the (many) amines are so toxic? It’s mentioned a quite often in the comments here.

    1. Barry says:

      I tend to assume that the nephrotoxity of e.g. polymyxin antibiotic amines is due to the same membrane disruption by which they kill bacteria, but that’s unproven to the best of my knowledge

  14. Barry says:

    kidney is susceptible to the surfactant action
    of the polymyxins through which
    their antibacterial action is mediated.23,48

  15. JC says:

    three aliphatic amines – membrane permeabilizer signature !

  16. Jim Mowre says:

    I like the part about resistance very much. You know within a year of its approval this will be the Frank’s Red Hot of the pharmacy – “I prescribe that **** for everything.” And five years later, the bacteria will ignore it as readily as they do everything else.

    Hopefully we can get the overprescribing issue fixed before we have to research the therapeutic index of KCN. “Well, let’s put it this way…it cured half the people who took it, and the other half? Let’s just say they’re not sick any more.”

    1. Orv says:

      I think if we can successfully prohibit its use in battery-raised farm animals, we can extend its effectiveness a bit longer.

  17. Mister B. says:

    Do you have an estimation of how long and how much this programm cost please ?

    As some very well informed readers have pointed out that this programm could fail due to potential toxicity of the molecule … I would be curious to know how much has been spent !

    Thanks !

    1. Derek Lowe says:

      Those are always very difficult numbers to come by – partly because companies don’t break them down in public, and partly because it can be difficult to figure out the real amounts. People overlap on other projects, assays and equipment get shared, costs get computed at the time by other categories than “by particular project”, and so on.

  18. david kuo says:

    Very cool that things that I worked on over a quarter century ago are just now being developed as potentially viable targets. My boss was working on beta lactam compounds on another project at the time (leukocyte elastase) and on a whim, we decided to test it out on our new “continuous assay” for our engineered E. coli signal peptidase (due to the classic serine protease motif we “saw”) and lo and behold, it actually inhibited the reaction. Really hope that something becomes of all this as pharma decided a long time ago to abandon many of these important targets for monetary reasons.

Comments are closed.