Yesterday, four people emailed me, asking about Brian Palmer's Slate article about antibiotic resistance. Since I'll probably get more such emails (and thank you for sending them), I'll offer my thoughts below:
1) Palmer's basic point about antibiotic development not being the answer is right. All a new drug does is kick the can down the road, since resistance will evolve to the new drug. Having said that, we currently do need new drugs, so we shouldn't stop developing them.
2) Palmer is not correct about plasmid curing as being a solution to antibiotic resistance. If we come up with a compound that causes bacteria to lose plasmids (plasmid 'curing')--mini-chromosomes that often contain antibiotic resistance genes--bacteria will evolve resistance to that compound. In addition, plasmids are important for the survival of ordinary commensal bacteria (and, in developed countries, most infections are due to opportunistic commensals, not obligate pathogens).
3) The missing 'non-drug' component is infection control. A rise in the percentage of resistant infections also typically means a rise in the absolute number of infections. Resistant infections, in other words, are extra bonus infections. This means that a rise in resistant infections indicates a breakdown in infection control, including simple things such as adequately cleaning surfaces.
4) Palmer is right, in that gene transfer certainly matters, but clonal spread is the ultimate problem. Palmer correctly notes that the 'mutational model' of antibiotic resistance is insufficient:
Antibiotic resistance is often described as a simple evolutionary response to environmental pressures--when a bacterial colony is exposed to drugs, the cells that develop defenses will survive and multiply. If it were this simple, bacteria would rarely survive an antibiotic attack. The few cells that mutated to defeat the drug would be killed off by your immune system before they could flourish. But bacteria can transfer genes to one another. When one cell "solves" a drug, it can package up the genetic recipe and transfer it to other bacteria. An entire colony of bacteria can develop antibiotic resistance with a single lucky mutation. And your body, which graciously hosts about 2 quadrillion bacterial cells--20 times your total number of human cells--is one enormous genetic swap meet. Most of your resident bacteria are either helpful or harmless. But some of them have been in our guts long enough to have seen our full menu of antibiotics. So even the so-called "normal flora" can archive antibiotic resistance and either go rogue themselves or spread it to more virulent invaders.
Obviously, bacteria acquire resistance genes through horizontal gene transfer. But antibiotic resistance becomes a problem through epidemic (or clonal) spread. That is, a handful of resistant genotypes spread widely and rapidly, as is the case with 'piggy MRSA.' We know resistance transfers within a person, but what scares me is having the same resistance gene in the same genetic background (i.e., genome) showing up in multiple hospitals. This means it's good at spreading from person to person.
While I'm on this topic, we also should distinguish between the evolution of resistance--how a previously sensitive organism becomes resistant--and the ecology of resistant organisms--how (and when and if) they spread and multiply. Obviously, these two things feedback on each other, but gene transfer in bacteria should be thought of as an analog of mutation, not as a way of increasing resistance per se.
5) There are some small problems with the description of KPC. I'm glad that Palmer discusses KPC, which are a very serious emerging threat. But Klebsiella pneumoniae-associated carbapenamase genes only confer resistance to all penicillin derived antibiotics ('only'--shudder). The problem is that, as Palmer notes, KPC-gene containing plasmids also have genes that confer resistance to other classes of antibiotics (and to make things even worse, these plasmids are often found in bacteria that can also have chromosomally-encoded resistance genes).
I'm still glad someone is calling bullshit on new drugs as the only option though.
Plasmid curing might just as easily result in chromosomal carriage of the resistance genes. That could be worse because the metabolic cost of keeping a few genes on the chromosome is less than keeping a few copies of a plasmid floating around.
I have some thoughts on new drugs that I'd inflict on everyone but I'll withhold for now as I'm being delinquent in finishing up some lab work.
We might have been reading different articles. From my reading, he was in fact arguing that new antibiotic development *was* the answer, and that reduction/optimization of antibiotic use was not. His main point seemed to be that we silly biologists and chemists hadn't realized that trying to halt genetic transmission was something we should be studying. Pretty crap article, in my opinion.
A couple of things
1. Anyone who thinks plasmid curing is a potential therapy hasn't ever tried to cure a plasmid :). More seriously, you are right that even if someone finds a medically relevant "curative", it won't have much value as a therapeutic.
2. The role of HGT seems to me to be more complex than outlined here. The evolution of antibiotic resistance outside the host shouldn't be overlooked. Chemical warfare in the soil and at the plant root surface has been going on for a lot longer than humans have been around, and this has driven (at least in part) the assembly of multiple resistance elements on mobile elements. This has exacerbated the problem in treating disease, of course. You are quite right to note that clonal spread is more relevant to the epidemiology of resistant disease, but HGT is an essential part of the path to a multiply resistant organism in the first place.
I remember some vague references in old Science News and/or Scientific Americans to bacteriophages and how there was work on phages before penicillin came into wide use. Would training an army of phages help?
And is inoculating cattle against 'evil' E. coli strains going to really help keep our food safe (am I conflating too much here?).
I see a lot of promise to strategies that disrupt quorum sensing.
We may sometimes use other businesses to perform certain services for us, such as maintaining the Site and our mailing lists, processing orders and delivering.