Bacteriocins have been known since the 1920's - originally called Colicines - and a nobel prize winner gave a shout-out to their seemingly miraculous killing ability during his laureate lecture (Luria, 1969). However, they, like probiotics and phage therapy, have been too difficult to make great success stories like the silver bullet classic small molecule antibiotics.

The chief problem with bacteriocins is that they are produced by strains relatively closely related to the target. Classical antibiotics are produced by molds or actinomycetes and can target common molecules among large groups of bacteria. However, if you are closely related to your target, you have to be much more selective - and that makes mutation away from sensitivity easier. Resistance and immunity already generally exist in nature, and humans mass producing these bacteriocins will be just the sort of pressure needed to push these factors into the populations.

Phage have a related problem. Resistance to phage is already common, and phage can't be given in controlled dosages. This means that everybody near a person under phage therapy is also getting doses of the phage. This expands the population under selection for resistance, making resistance more readily available next time around.

Probiotics are also hard to modulate and control, and because they must be closely related to the pathogens, to hold a related niche, they can take up pathogen genes and become pathogens themselves. Further, they may cause some of the same problems as the pathogen, if they are put in an unwelcome place that should normally be sterile, like the bladder.

One interesting idea is to combine probiotics and bacteriocins. Use the bacteriocins to upset the applecart of the rock-paper-scissors game going on in the world (Kerr et al, Nature 2004/Kirkup and Riley, Nature 2004) and then add resistant probiotic bugs to quickly fill in the gap left behind.

There is lots to discuss about all the ways of dealing with antibiotic resistance. I hope people start thinking about this...

1st off excellent post and suggestion! However, your proposal would be quite a complex feat. But how would the body know how tell a mediocre invading bacteria which is really helping the body out from the really bad invading bacteria which would destroy a person? It seems as if the immune system would launch a 2 front war. And wouldn't it be more likely for the mediocre bacteria to be eradicated 1st then the deadly bacteria by the immune system? If could somehow relay this information to the immune system then problem solved, but the consequencies of this kinda work blow my mind. If only we could have a weak or dead bacteria being introduced into the body kinda as a training for our immune system, and see how well that works. Then hopefully the information and tactics gathered from that training will payoff when the real thing comes to town. Unfortunately with such superbacterias, a dead one or a weak one is whole different ball game then a fully armed and ready beast. Regardless alot more research will have to go into this, but I just would like to see how well our immune systems can handle this threat without drug aide. After all other organisms have been doing it the natrual way for millions of years.

I'm glad you liked the post.

Here's my thought - a mediocre bacteria might have a set of adhesins and some nutrient scavenging molecules that allow it to thrive, without toxins and proteases which cause it to really dig into the nearby host cells. It would be able to occupy the empty niche, fill the space, but not really exploit the host the way a full on pathogen would. It would certainly have a harder time, in absence any assistance, because it can't access, say, all the free iron caused by internal bleeding. However, it might be just enough to keep pathogens at bay when a small infectious dose comes along.

Anyway, the large intestine is at a state of continuous mild inflammation anyway, from birth on. The immune system is always modulating the bacteria there, and this would be no different.

But yes, it takes a certain boldness to do this... and these are things that can't be done with probiotic lactobacilli. You need to use something more closely related to the actual pathogens.

Although Ben points out the narrow killing breadth of bacteriocins as a problem, in fact, it is just their narrow-mindedness that makes them a super alternative to classical antibiotics. One will only use a particular bacteriocin when required, rather than constantly and in large doses as is currently the case for the currently prefered antibiotics. Thus, the target of selection for resistance stays small and is negligible at any moment. Compare the target of selection for resistance to amoxicillin in our gut (perhaps 1E12 bugs - essentiall all bacteria in our gut) versus the target for selection for resistance to a bacteriocin like colicin (perhaps 1E5, only members of E. coli). THat is the beauty of using bacteriocins, they are used infrequently and the selection target size for resistance is much, much lower. In fact, in animal trials we have yet to see a single case of resistance!

Glad to hear it is working.

I wonder why resistance was not observed.

I presume, if no cases of resistance were observed, that the infections were carefully screened or created sensitive. There cannot have been a single case of resistance already dominating a test pathogen population. This strikes me as an unusual case given Gordon's demonstration of the prevalence of resistance in natural populations. I suppose it raises the question of whether a GI population and a UT population are structured at all similiarly; and furthermore, what their natural relationship to the host might be. After all, the GI tract is effectively in a state of moderate inflammation but is tolerant to the native biota. The urinary system may be much less tolerant, and in this case the classic assumption about the immune system cleaning up remnant resistant cells after antibiotic administration may be appropriate as a model for why rare cases of resistance do not emerge. Or there may be some other reason I'm not considering.

Certainly, if there were no pre-existing resistant microcolonies, new resistant mutants might arise at 1E-7; in an animal UTI model that might result in 1 or a few mutant cells per animal. This seems less likely to develop into a clinical infection. Further, the obvious mutations might, if the -cin is cleverly chosen, target features critical to UTI development, making the effective mutation rate for UTIs even lower, since most would be relatively crude - stops and deletions in key infection determinants. One wonders, if this is the case, how many individuals would need to be treated before effective resistance emerged, how widespread the therapy would have to be to start tipping the scale such that new cases were sensitive with a lower and lower probability. This seems to demand an analysis beyond single case studies, perhaps epidemiological models of entire populations. It's great that such models can be envisioned prior to drug deployment, at least.

Naturally I'm not saying anything you don't know.
It must be quite fascinating data, from those animal models.