Science Editor

From your mouth to your intestines, large and small, stretches an unbroken, seamless tube of tightly knit epithelial cells. Alongside this food canal extends a parallel air passage, running from nostrils to bronchii and lungs also lined with epithelial cells. These undulating surfaces are home to unnumbered kerzillions of infectious bacteria, all lying in wait for the odd scrape or cut that will gain them entry below the surface, with a license to multiply and infect.

When white cells of the immune system that police the mucosal cells encounter a wound in the surface, they signal the epithelia to fabricate their own protective molecular shield. Those alarm bells turn on genes that activate a class of compounds called bactericidal/permeability-increasing protein BPI.

Cell biologists at the Harvard University-affiliated Brigham and Women’s Hospital in Boston discovered this emergency signaling pathway, as they report in today’s Proceedings of the National Academy of Sciences (PNAS), dated March 12, 2002. Their paper bears the title: “Lipid mediator-induced expression of bactericidal/permeability-increasing protein (BPI) in human mucosal epithelia.” Its senior author is cell biologist Sean Colgan, an associate professor of experimental therapeutics at Harvard.

“The significance, I think,” Colgan told BioWorld Today, “is that it provides a new strategy to potentially combat infectious diseases at mucosal surfaces. The mucosa of the epithelial cells are the first line of defense for any infectious disease. An important aspect is that you would circumvent other anti-microbial problems, such as antibiotic resistance to some bacteria. The BPI protein,” he pointed out, “is a naturally occurring compound, so no chemicals here.

“In the mid 1980s,” Colgan recalled, “BPI was a hot topic. What led us to it was that we had utilized a transcriptional profiling approach. This is a fairly new technique, a microarray analysis whereby one can screen the expression of thousands of human genes in one fell swoop. That was our first clue, and we pursued it because it was somewhat aberrant to what had been shown before that BPI was only expressed in neutrophils.

“The upshot of the microarray scan was to define anti-inflammatory genes induced by lipoxin. Our hypothesis was that a number of anti-inflammatory pathways were induced by the lipoxins. Those are the lipid molecules that we used to screen anti-inflammatory genes in epithelial cells. These are compounds that are made during inflammation, and they’re lipid molecules derived from the cell membrane of different cell types.”

Bad’ And Good’ Molecules In Play

“One of my co-authors here, Charlie Serhan, originally discovered the lipoxins,” Colgan said. “The unique idea about lipoxins is that they are really the first anti-inflammatory lipid molecule defined in nature. He and I set out to define the profile of an epithelial cell in response to exposure to the lipoxins what genes are turned on in an inflammatory setting in the epithelium.

“A salient point in the PNAS paper,” Colgan continued, “is that ATL the aspirin-triggered lipoxin compounds that we’re working with induce a whole new family of molecules on the mucosal surface of epithelial cells. An interesting aspect of the BPI gene,” he went on, “is its close proximity to another molecule, the LBP binding protein. It falls in the family of lipopolysaccharide-binding proteins. I think most scientists would consider LBP to be a bad’ protein and BPI to be a good’ protein in this regard. LBP is important in responses to endotoxemia, whereas BPI battles against endotoxemia.

“The crystal structure of the protein is known,” Colgan observed. “It’s a boomerang-shaped molecule, which basically has two ends. One end is important in killing bacteria and in binding lipopolysaccharide endotoxin. The other end is important in the molecule binding to the surface of the immune system’s leukocytes. That function was the original traditional concept. Previous to these studies it was thought that BPI’s only existence was in soldier cells of the immune system that is, neutrophils battling infections, mostly. But our studies suggest that BPI was predominantly expressed on the surface of the cell, so that any bacteria that comes in close contact with the epithelium could be killed.”

Colgan described his group’s bacteria-killing experiment:

“We were using an intestinal epithelial cell,” he recounted. “For that reason we thought it important that we use a bacterium important to the intestine. Salmonella typhimurium is a common source of food poisoning. So we selected a strain of Salmonella that we knew would be sensitive to BPI killing. We used that as a gauge for whether BPI was functional, and whether we could induce function with the aspirin-triggered lipoxin molecule. Essentially what we did was allow bacteria to adhere to the epithelium, and addressed whether killing of Salmonella was inhibited with an antibody directed against BPI. We found that epithelial cells of the intestine do express functional BPI,” he went on, “which could be inhibited with antibody. And that killing of bacteria was increased by 60 percent with exposure to the aspirin-triggered lipoxin.”

Colgan explained why BPI kills only Gram-negative bacteria: “First and foremost only Gram-negative bacteria make LPS. And the outer membrane of a Gram-negative bacterium is different, and those differences appear to explain the mechanism by which BPI kills those bacteria, but not Gram-positive ones.”

The sheath by which BPI coats an epithelial structure with its antibacterial layer is not readily discernible. “To recognize its physical presence,” Colgan said, “we looked mainly by confocal microscopy. That’s a CAT scan of cells one layer thick. So we could scan through the cell and determine its localization. And using an antibody against BPI we were able to show that it was expressed predominantly on the cell surface.”

Colgan observed that Harvard filed patent applications last December with him and Serhan as co-inventors. “Certainly licensing would be welcome,” he concluded.

Stephen Carrol, vice president of preclinical research at Xoma Ltd., of Berkeley, Calif., supplied Colgan with recombinant BPI and anti-BPI antibodies. “We have constructed a recombinant form of that protein,” Carrol told BioWorld Today, “and we have taken it into a number of clinical indications that Xoma has licensed to Baxter. They are currently conducting a Phase II clinical trial in Crohn’s disease. What’s interesting from a scientific perspective,” Carrol added, “is that Colgan’s study suggests that they can find BPI in endothelial cells which was unexpected.”