Orlando and colleagues (DOI link) provide 123 base-pairs of sequence from a tooth dating to 100,000 years ago from Scladina cave, Belgium. The study's lede is that this sequence is more divergent from other Neandertal sequences than those other sequences, so far, have been from each other.
Neandertal mtDNA sequences are not uniform -- there are polymorphisms among the known Neandertal control region sequences. So far, the news has been that Neandertals are about as diverse in their mtDNA polymorphisms as modern humans. That's interesting, because modern humans are notably limited in mtDNA variability. A similar lack of variation among Neandertals tends to indicate that the same thing that happened to us may also have happened to them.
As for what that "thing" might be, my number one hypothesis would be natural selection. It makes a great deal of sense for mtDNA to have undergone one or more adaptive substitutions during human evolution. Many of the effects of current mtDNA polymorphisms in humans impact precisely those physiological features of humans that have evolved most markedly during the past few hundred thousand years -- including longevity, energetics, and the brain. Certainly the latter two, and possibly longevity, also changed over time within European Neandertals. So, I would propose the hypothesis that two (or more) selective sweeps occurred in the ancestors of modern humans and in Neandertals.
Orlando and colleagues present possible evidence for one of these selective sweeps -- the Neandertal one. The Scladina sequence adds four more polymorphisms across the 123 bp of its sequence compared to other Neandertals. All the other Neandertal sequences across this same 123 bp show only a single polymorphism. The paper proposes that the additional diversity displayed in the 100,000-year-old specimen is evidence for a reduction in Neandertal mtDNA diversity over time.
The interpretation? Genetic drift:
While the diversity of the more recent Neandertals is similar to that of modern humans worldwide, the sequence from Scladina reveals that more divergent Neandertal haplotypes existed before 42,000 years ago. This could suggest that Neandertals experienced genetic drift as demographic bottlenecks eliminated the phylogenetically more recent (i.e. less expanded) haplytpes from populations. Consequently, the most likely conserved Neandertal haplotypes could also be the phylogenetically most ancient (i.e. the most closely related to the common ancestor of modern humans and Neandertals). This could explain the shift towards modern human pairwise distributions observed between 100,000 and 40,000 years ago. Whatever this shift should be related to cohabitation, climatic changes, or any subdivision of populations, the Scladina sequence has revealed that the genetic diversity of Neandertals has been underestimated (Orlando et al. 2006:R401).
I would interpret this as evidence of selection -- a selective sweep in Neandertal mtDNA after 100,000 years ago. Clearly, more sequences would be necessary to assess the pattern of reduction more closely.
There is a reason to be cautious about this new sequence, though. Out of the four polymorphisms by which it differs from other Neandertals, two are transitions from C to T, and one from T to C. A large number of such transitions, especially C to T, result from DNA damage in ancient specimens. The paper notes that other ancient human specimens don't typically have damage at these specific locations, but that doesn't really answer whether this especially old specimen may have had such damage. And the fact that the changes make the sequence less like modern humans when compared to Neandertals is a strike against them, because DNA changes should be more likely to have occurred on the 60,000-year-longer branches leading to more recent fossil Neandertals.
So, there may be some damage there. Conservatively, I would say that the case for mtDNA diversity among Neandertals uncertain. Bring on the autosomes!
Orlando L, Darlu P, Toussaint M, Bonjean D, Otte M, Hänni C. 2006. Revisiting Neandertal diversity with a 100,000 year old mtDNA sequence. Curr Biol 16:R400-R402. DOI link