Skip to Content

New Antibiotic Action

I’m always happy to hear about new agents to treat gram-negative bacterial infections, especially after a stint working in antibiotic drug discovery myself. Seeing just how hard it is to kill these things did not improve my peace of mind about the problem of bacterial resistance, that’s for sure. So this paper caught my eye, although it’s not exactly a new drug – just a new antibiotic.

Glatiramer acetate (GA) has been a multiple sclerosis drug for many years now – enough years that it’s the subject of a long-running series of brutal patent fights between Teva (who sells it as Copaxone) with competitors, which could be the subject of another blog post all by itself. The drug itself is not a single substance, but rather a mixture of polypeptides containing lysine, alanine, glutamic acid, and tyrosine (constituents of myelin basic protein). It has not-very-well-defined effects on the immune system in MS patients, and seems to reduce the frequency of relapses while not really modifying the underlying progression of the disease itself.

Now it’s been found (by a multicenter team from Denmark, the UK, and the US) that GA is actually a pretty good gram-negative antibacterial agent, with fast bactericidal activity against many species. It’s especially effective against Pseudomonas aeruginosa, and even kills organisms that have formed biofilms. The mechanism looks to be similar to some other antimicrobial peptides, although GA is still active when administered in the context of human plasma (which is the downfall of many other AMPs). That mechanism is almost certainly through membrane disruption, as is the case with other cationic, amphiphilic peptides. The advantage here is that this is a drug whose human safety has already been well-established, because that’s another thing that has tripped up other attempts to use naturally-occurring peptides as systemic drugs.

Some other good news in this area comes from the CARB-X partnership which involves a number of academic and nonprofit research groups. So far, after about a year, they have a good-sized portfolio going:

Of the 17 drug discovery projects CARB-X supported in its first year, 12 focus on ‘traditional’ antibiotics — small-molecule, bacteria-killing drugs.

One-third of these projects are working on drugs with novel chemotypes and new mechanisms of action, offering high development risk but also high reward.
For example, two small molecules inhibit LpxC, the enzyme that catalyses the synthesis of lipid A, a key component of the outer monolayer of Gram-negative bacteria. Another third are advancing new chemotypes for established mechanisms of action, such as new topoisomerase and gyrase inhibitors. The last third act via established chemotypes and on established mechanisms of action, with the lowest scientific and developmental risk but also offering the lowest potential antimicrobial resistance payout.

The new chemotypes against traditional targets are particularly interesting – as that article points out, it’s been a long time since any such molecule hit the market. There are other efforts underway (such as the CO-ADD group) to mine chemical space for insights into penetrating gram-negative bacterial membranes and for new chemotypes in general. Screening in the manner is really a no-brainer; there are so many collections that have never been looked at that we would be completely remiss not to go through them. We’re going to need all the help we can get.

17 comments on “New Antibiotic Action”

  1. Sebastian says:

    Might the antibacterial properties of GA contribute to its action in MS? I’m not too familiar with the drug, but is it possible killing gram-negative bacteria shifts the immune response from Th1 to a Th2 and Treg profile, reducing the autoimmune degradation of myelin?

  2. lynn says:

    Sort of interesting – but GA is dosed at 20-40 mg/day in humans and the effective in vitro dose for P. aeruginosa [the most sensitive bug tried] is 25 – 50 micrograms/ml. I couldn’t find PK data for GA, but antibiotics with MICs in that range are dosed at grams/day. [I’m open to corrections here!] This is a general problem with repurposing of non-antibiotic human health drugs – required concentrations are way out of line.
    Efforts like CARB-X and CO-ADD do give hope and broad screening [as well as studies of characteristics needed for penetration of small molecules into Gram-negatives] is worthwhile, but finding compounds that kill bacteria is a long way from arriving at drug candidates.

    1. Derek Lowe says:

      Here’s what the authors had to say on the question of dose – they’re wondering about CF applications:

      “With the already high dosages of GA employed clinically, it seems likely that critical levels of compound to kill microbes can be reached without compromising safety. For example, this could be relevant in treatment by inhalation of lung infections with P. aeruginosa. e total dosage of 20 mg used routinely in MS therapy would replenish 400 ml of lung uid to the maximum 50 μg/ml concentration used in our studies. e volume of the epithelial lining uid in the lower respira- tory tract, o en a site of P. aeruginosa infection in cystic brosis (CF) patients, has been estimated at 15–70 ml42; therefore, infection in this location would only require a dose of 3.5 mg GA, and hence permit a considerable overload of drug delivered to other sites of infection and compensate for losses to zones outside infection foci.”

    2. MBP says:

      It’s all ADME. Neglecting distribution and clearance, 30 mg/day in an adult human (~5 L blood volume) by IV dosing would give a Cmax of 6 ug/mL. Distribution effects could lower that number, while clearance could affect it in either direction in the context of a multiple dosing regimen based on the relationship between the dosing interval and the elimination half-life. So it’s really impossible to say without the PK data on the drug, but a Cmax of 6 ug/mL is in the ballpark of the effective blood concentration you mentioned.

  3. Dr. Manhattan says:

    GA is a random mixture of random peptides of four amino acids ( lysine, glutamate, alanine, and tyrosine) so the basis of antimicrobial activity is uncertain at best. Maybe some specific peptide sequence has the antimicrobial efficacy and could be further tested and refined. As Lynn points out (Hi Lynne!) an in vitro activity of 25-50 ugs/ml is far from spectacular, and depending on PK, could mean large doses are necessary. This is really important since we are talking about peptides in this case, which are notorious for rapid clearance and other challenging ADME properties. There are also a reports of few serious side effects (CV, liver) at the much lower current dosage. So whether this can be effectively repurposed for infection is very much an open question.

    Yes, Carb-X and CO-ADD are two different focused efforts to improve availability of novel antimicrobials. The Carb-X list of programs funded is available online ( in their annual report, as is information for groups wishing to look into funding by the organization.

    1. Chrispy says:

      I wonder if Copaxone could be treated like a crude natural product and the antibiotic component purified. This might go a long way towards eliminating side effects and helping to determine the underlying mechanism.

      1. Isidore says:

        Four amino acids would yield 256 different tetrapeptides, assuming that each amino acid can be used more than once, fewer if not. Add 64 tripeptides, 16 dipeptides and 4 single amino acids and it is certainly quite possible and fast to test each one of them for antimicrobial activity.

        1. Barry says:

          “Glatiramer acetate is a random polymer (average molecular mass 6.4 kD)”

          unlikely we’re talking tetrapeptides. More like 64mers. and 4 to to the 64th (3.4 x 10 to the 38th) is an inconveniently large number of peptides to synthesize and screen

          1. Barry says:

            But the number of candidates to synthesize, test is vastly lower if we acknowledge that it’s probably the Lysines doing all the work:


          2. Chrispy says:

            The combinations don’t present a problem — we have historically isolated trace ingredients from a fungal broth with nothing more to go in than that it killed bacteria. Do a couple rounds of HPLC on Copaxone to find the active peak(s), then determine the composition. Knowing that it came from only four amino acids makes the problem much simpler.

  4. MikeRobe says:

    Its a known fact and patented that polyamines, polylysines and other nitrogenous compounds including heterocycles can affect the growth of gram negative bacteria. Take a look at polymyxin itself and its not really earth shattering news that complex mixtures of such compounds can cause growth inhibition including in those bacteria harboring efflux pumps of the Mex family and others. But that’s what is left, complicated molecules that need chemistry finesse and evolution, the biology is straightforward.

  5. Vaudaux says:

    As so often noted in this blog, the dose makes the poison.

    This study uses the peptide LL-37 as comparator, not an antibiotic.

    Hundreds of small cationic antimicrobial peptides have been described previously. Generally they have fairly weak activity compared to clinically used antibiotics, and are too toxic to allow high enough doses to be efficacious. None has yet reached phase 2 evaluation as a systemic drug.

    Re administration by inhalation rather than systemically: There are several antibiotics already used in this way for cystic fibrosis. Another one might not add much value. Considering other uses of inhaled antibiotics, there have been a couple of failures recently (ciprofloxacin and amikacin, both from Nektar and Bayer).

    1. slcimmuno says:

      “This study uses the peptide LL-37 as comparator, not an antibiotic.

      Hundreds of small cationic antimicrobial peptides have been described previously. Generally they have fairly weak activity compared to clinically used antibiotics, and are too toxic to allow high enough doses to be efficacious. None has yet reached phase 2 evaluation as a systemic drug.”

      Not exactly true — see Brilacidin, a synthetic mimic of Host Defense Proteins, successful P2b in ABSSSI w Dapto the comparator … designed by William DeGrado and others … currently being tested in OM and IBD too.

      See cites

  6. RBW says:

    With regards to the peptide diversity comments, why stop at tetrapeptides? In the early days of combichem, people were screening peptide libraries for antimicrobial activity. Like Houghton’s split and mix hexapeptide library. I think he fished out hexaphenylalanine as a lead.
    Glatiramer is a great success story of how not to make peptides. Normally it’s a stepwise process with lots of protection and deprotection. But for glatiramer it seems the sequence doesn’t matter. Neither does the length as long as it’s a polymer. So it’s a blockbuster made by uncontrolled peptide synthesis.

  7. Polynices says:

    So my crazy theory is that MS is actually caused by occult gram negative infection and GA helps MS because it actually treats this infection, not because it does whatever to T-cells. I don’t believe this but I bet someone will suggest it in earnest. Occult infection has been blamed for just about everything, including MS.

    1. Matthew K says:

      Seems like the right place to link David Wheldon’s site – these kinds of theories are usually borderline crank but I have a certain sympathy for his reasoning and deductions:

  8. MFernflower says:


    Can I have my NDA now?

    Interesting property of COP-1 but antimicrobial peptides fail 99% of the time to reach the market (exception being daptomycin) – I am cautiously awaiting more research into the bactericidal properties of this drug

Leave a Reply

Your email address will not be published. Required fields are marked *

Time limit is exhausted. Please reload CAPTCHA.