Napoleon lice 'n' mites
afarensis notes, however, that they extracted DNA from tooth pulp to amplify B. quintana (a louse-transmitted agent of trench fever), and says,
I will also be the first to admit that I have no idea how the DNA got into the pulp cavity in the first place...
Simple: via the bloodstream. Dental pulp is a living tissue, receiving a blood supply like any other tissue in the body—and any pathogens that are carried in the blood supply can end up there as well. Dental pulp is good for these ancient investigations because since it's surrounded by the tooth material, it's unlikely to be contaminated with other pathogen DNA, and it's more protected from decay and therefore a usable sample is more likely to be extracted. To my knowledge, this was pioneered by Michael Drancourt and Dider Raoult, both of whom are authors on the louse paper; the first publication on the technique was this one on Yersinia pestis DNA, which I'd not a half-hour ago mentioned in the comments to this thread. It's fascinating stuff--Drancourt and Raoult have also used it to examine B. quintana in a 4000 year old tooth, among other projects, which have significantly pushed forward the field of paleomicrobiology.