Antibiotic resistance

Antibiotic use selects for resistant bacteria, that then can predominate in the population. As a result, bacterial infections become more difficult to treat. Important resistant bacterial pathogens we work on include Meticillin Reistant Staphylococcus aureus, Carbapenem Resistant Enterobacteriaceae, and the pneumococcus. As resistance can make certain infections almost impossible to treat, the question of how it emerges and spreads is of vital importance.

As noted elsewhere, we actually have an effective vaccine against some pneumococcal serotypes. Because these serotypes are disproportionately associated with resistance, this provides an opportunity to see how and if resistance re-emerges following widespread vaccination. In the first generation of pneumococcal conjugate vaccines, initially the prevalence of resistance decreased as the vaccine-type strains were excluded from the population, but then returned.

In some cases this was due to former vaccine targets that were resistant recombining with a non-vaccine type strain and hence escaping the vaccine. We showed that one such, ST 320, became the dominant strain in Massachusetts (and important nationwide).

After the 13-valent vaccine was introduced in 2010, it has been expected that resistance will decline. The possibility of resistance re-emerging again is cause for concern. We are working with the CDC to sequence resistant cases of IPD collected as the vaccine was introduced and define and characterize the major lineages. We are especially interested in the role of recombination in possibly generating new fit clones. We are also addressing the hypothesis that resistance emerges in certain regions, likely those using more antibiotics than others, and then spreads. Modeling work undertaken in the lab suggests that variation in the burden of carriage can impact how resistance survives a vaccine campaign – results that may be important in sub-Saharan Africa and other places with a high burden of pneumococcal carriage.

The focus on recombination reflects the potential for horizontal transfer of resistance determinants. Presumably any lineages that take up DNA from the environment more readily than others will have more opportunity to sample and integrate resistance genes. We know that pneumococci vary greatly in their relative rate of recombination – defined as the ratio of the amount of change introduced by recombination to that introduced by point mutation. In the Massachusetts dataset, the lineage with the highest relative recombination rate by far was ST 320, which as noted above was a vaccine escape variant harboring multiple resistance determinants.

This could be argued to be a case of circular logic. Surely in some cases things that are resistant must have more evidence of recombination, because by definition they must have undergone recombination in order to become resistant. This is not the case however, because the evidence of recombination is at multiple points across the genome – not only resistance loci or the capsular biosynthetic locus. This is consistent with our prior analysis of the MLST database using BAPS, that identified strains with more evidence of recombination at the core housekeeping genes used in MLST. These are not linked to resistance (one that is linked, ddl was excluded). Strains with more evidence of recombination at the MLST loci are significantly more likely to be resistant to β-lactams, macrolides, chloramphenicol and tetracycline.

This finding is all the more striking because these resistance mechanisms are transferred by different means: the β-lactams by altered penicillin binding proteins that can be acquired through natural competence and homologous recombination, while macrolide resistance is found on mobile elements capable of transferring themselves without assistance from the cell’s own DNA uptake machinery (often along with tetracycline). The rate of recombination may not be constant over the history of a strain. We know that loss of capsule, which happens naturally if rarely through stop codons and similar lesions in the relevant genes, can be followed by an increased rate of DNA uptake, and we have proposed that this may lead to temporarily enhanced rates of recombination in those lineages that is lost when the capsule is restored. There is clear evidence for more than one pattern of recombination from population genomic studies of the pneumococcus.

Other work and interests of the group include the role of agriculture in the origin and dissemination of resistance, which is almost certainly important but difficult to pin down precisely due to a lack of adequate data.