Even after the bacteria that cause meningitis are finally dead, they still won't leave the blood-brain barrier alone.

That's the take-home message from a paper published by scientists from the Charité, the medical school of Berlin's Humboldt University, in the May 2005 issue of Journal of Clinical Investigation.

"We know a lot about what endothelial cells do in infection, but it has never been demonstrated in such a straightforward way that bacteria kill endothelial cells and vascular cells, and that they do it in two ways," said senior author Joerg Weber.

Endothelial cells are the interior lining of the body; they are an important part of the blood-brain barrier, and they also separate blood from tissue and the inside from the outside world in several other organs. Weber and his colleagues exposed cultured endothelial cells to either pneumococci, the bacteria that cause meningitis, or pneumococcal cell wall. "When [the bacteria] grow, they lyse, and pneumococci grow very quickly," Weber explained. "And when you treat pneumococcal infection, they basically disintegrate."

Either scenario leads to the appearance of free-floating debris of pneumococcal cell wall. That debris, it turns out, can kill endothelial cells, just like living pneumococci and the toxins they secrete, but with different time courses and mechanisms.

Exposure to living bacteria killed 80 percent of endothelial cells within 12 hours, and 100 percent within 18 hours. In contrast, bacterial cell walls took 72 hours to induce maximal cell death.

The differences in time course could be the reason for the differences in mechanism. While pneumococcal cell walls induced classical apoptosis, as assessed by both the activation of caspases, which are apoptotic enzymes, and physical characteristics of the endothelial cell nuclei after death, the living bacteria induced an apoptosis-like death via mitochondrial damage caused by the bacterial toxins pneumolysin and hydrogen peroxide.

Exposure to living bacteria caused an increase in intracellular calcium, while exposure to bacterial cell wall did not. Usually, calcium influx is one mechanism that triggers caspase activation. Weber said that did not occur there because "very early on, there is significant injury to mitochondria that induces death via a separate mechanism. My guess is that at least in these first hours, this toxic process is so rapid that the killing mechanism comes before caspase activation."

In endothelial cells exposed to bacterial cell walls, the opposite dissociation occurred: There was no influx of calcium, but caspase-mediated apoptosis nevertheless occurred. In that case, the caspases were activated through a type of Toll-like receptor, TLR-2.

Toll-like receptors are major immune system sensors. Weber and his colleagues investigated their role, focusing on TLR-2 and TLR-4, which play a role in the inflammatory response to both living pneumococci and pneumococcal cell wall debris.

Using a model cell line transfected to express either TLR-2 or TLR-4 receptors, they again found differences between the effects of the living bacteria and the pneumococcal cell wall. While the living bacteria were able to kill those cells independently of whether those cells were expressing either TLR-2 or TLR-4, cell wall debris exerted its effects by activating the TLR-2 receptor.

The results presented by the Charité scientists could point to new treatment strategies for diseases besides meningitis. Meningitis also is a good model system of sepsis, which, with a fatality rate of around one-third of cases and more than 30 unsuccessful randomized clinical trials over the past few decades, could certainly use better treatment options.

"Currently, in sepsis and meningitis, we only target the bacteria," Weber said, adding that his research suggests that it is necessary to treat the cellular consequences of having bacterial debris around after treatment, as well as killing the bacteria themselves. "With knowledge of the mechanisms, it will be much easier to create protective and, in the case of meningitis, neuroprotective strategies.

"In five to 10 years, there will be [clinical] strategies based on interfering with apoptosis. This is where we are trying to go now."