Tiny bubbles: vesicles that cells spit out are implicated in cancer and AIDS.

When Gustavio Romanio looked through his microscope last year, the University of Michigan cancer researcher was perplexed. Tumor cells that had been treated with chemotherapy agents had bubblelike objects at their edges. These microscopic vesicles were apparently being shed by the cells. What's more, the bubbles contained some of the chemotherapy drugs. Romanio and his colleagues quickly scanned the research literature and discovered that other scientists had seen similar vesicles, dubbed exosomes, coming off many kinds of cells, including cancer cells.

"Although I was trained as a cell biologist, I had not heard of exosomes or vesicle shedding before," recalls Romanio.

Many biologists still haven't. This spring, James Hildreth of Johns Hopkins Medical Institutions in Baltimore spoke at a meeting of AIDS researchers and proposed that HIV travels between cells as cargo in exosomes. "Many of the [people at this meeting] didn't even know what an exosome was," Hildreth recalls.

That may be changing. In the Sept. 16 Proceedings of the National Academy of Sciences, Hildreth and his colleagues detail the provocative connections between exosomes and HIV and call for a new vaccine strategy based on those links. And Romanio's group has just published data suggesting that tumors use exosomes to foil cancer drugs. And other scientists are now using exosomes given off by immune cells to battle cancer and infectious microbes.

"Exosomes are a device for transporting material from cell to cell. It's kind of a biological FedEx," says Jean-Bernard Le Pecq, cofounder of a company that's testing whether exosomes can stimulate immune responses against tumors.

IN THE BLOOD Unless one knows what to look for, it's easy to miss exosomes. Until 20 years ago, the flattened spheres of lipid molecules, ranging from 50 to 200 nanometers in diameter, were dismissed as merely free-floating fragments of a cell's membrane.

In 1983, however, a research group studying immature red blood cells concluded that exosomes have a function. As red blood cells mature, they expel most proteins other than the oxygen-carrying molecule hemoglobin. The investigators determined that exosomes served as cellular garbage bags for the unneeded proteins. "It was interesting, but people were not that excited about [exosomes] at the time," recalls Clotilde Thery of the Curie Institute in Paris.

In 1996, a Dutch group led by Graca Raposo of Utrecht University in the Netherlands reignited the exosome field by reporting that the vesicles are also made by immune cells called B lymphocytes. The researchers showed that exosomes form within larger sacs inside the cell before traveling to the cell surface for release.

Something else very interesting occurs within these sacs: Immune system proteins of the major histocompatibility complex (MHC) join to molecular fragments, called antigens, such as bits of an invading bacterium. In fact, Raposo's team found that exosomes carry these MHC-antigen complexes, which typically stimulate immune action against microbes bearing the antigens.

"At that point, people started to think that exosomes could play a role in immune responses," Thery says.

MHC-antigen complexes are crucial to how a person's immune system defends against infectious microbes. Macrophages, dendritic cells, B lymphocytes, and other immune cells chop up bacterial and viral proteins into fragments. Those cells package the resulting antigens with MHC proteins and present them to T cells, natural killer cells, and other members of the immune system. That way, these mercenaries of the immune system know what they should be hunting down.

In 1998, Raposo, who had moved to the Curie Institute, joined with colleagues in France and Italy to report that dendritic cells also secrete exosomes containing MHC-antigen complexes. More important, the scientists reported that these exosomes could serve as the basis of a new anticancer strategy.

Scientists consider dendritic cells to be the most potent antigen-presenting cells in the body. Some investigators have primed these immune sentinels by loading the cells with antigens in the form of proteins or protein fragments from cancer cells. The dendritic cells are then injected into a patient as a cancer vaccine.

In a twist on that strategy, Raposo and her colleagues loaded laboratory-grown dendritic cells with potential tumor antigens. Instead of injecting the dendritic cells into test mice, the researchers isolated exosomes released by the immune cells and injected the exosomes. The new strategy prompted the animals to mount immune attacks on their tumors that ultimately eradicated the cancer. In mice, the exosome approach has worked against lymphomas, melanomas, and breast cancer, says Thery.

She and her colleagues are now trying to determine how exosomes excite an immune response. In the December 2002 Nature Immunology, they reported that exosomes originally isolated from dendritic cells, when injected into mice, are captured by other dendritic cells. These cells then apparently use the antigens and other molecules within the exosomes to stimulate the activity of T cells, the true warriors of the immune system.

Thery speculates that the first dendritic cells to encounter an infectious microbe produce exosomes bearing microbial antigens as messages to other dendritic cells, sparking a chain reaction that amplifies and perhaps speeds the overall immune response.

CANCER FIGHTER OR HELPER? Whatever their mechanism of immune stimulation, exosomes have already been drafted by researchers for the fight against human cancer. A company called Anosys, based in Menlo Park, Calif., and cofounded by Le Pecq, has sponsored safety trials of exosomes derived from dendritic cells. So far, a few patients with either small-cell lung cancer or melanoma have received them intravenously.

"We don't see any toxic side effects" says Le Pecq. "We have seen slight tumor regressions, but nothing spectacular:'

Still, the results from these preliminary studies are encouraging enough that Anosys-funded researchers in Prance and the United States will

next year launch a larger trial of exosome treatment for people with severe lung cancer. The firm is also exploring whether exosomes loaded with microbial protein fragments might offer a new way of vaccinating people against infectious diseases.

As researchers try to determine whether exosomes can be used as weapons against tumors, that effort begs the question of why tumor cells themselves typically secrete the vesicles. Thery's colleagues, led by Laurence Zitvogel of the Gustave Roussy Institute in Villejuif, France, first documented that phenomenon in 2001, and Romanio's observations last year seemed to confirm it. Zitvogel's team has speculated that the tumor makes exosomes to somehow blind the immune system to the growth of the cancer cells.

Romanio's research offers another potential dark side to exosomes. His group has been studying how tumor cells respond to anticancer drugs. Many tumors can withstand the actions of such medications, and investigators have identified several molecules that pump drugs out of resistant cells or transport them to regions of the cell that corral and disarm the chemicals.

In the Aug. 1 Cancer Research, Romanio and his colleague report that the chemotherapy agent doxorubicin and similar compounds accumulate in vesicles shed by cancer cells. Immediately after doxorubicin is added to cells, it travels to a cell's nucleus, but within 3 hours, it's seen in exosomelike vesicles at the cell's edges. "We don't really know if they are exosomes," cautions Romanio.

Supporting the hypothesis that these vesicles offer a way for tumors to rid themselves of toxic drugs, the researchers also document that genes involved in vesicle formation are more active than normal in tumor cells with increased drug resistance. Indeed, the amount of vesicle shedding by various cancer cells correlates with each type's degree of drug resistance.

Romanio speculates that certain anticancer drugs, such as those attracted to the lipids that make up exosome membranes, are most susceptible to removal by the vesicles. Conceivably, interfering with the expulsion of the drugs by the vesicles could make chemotherapies more potent. "We are exploring strategies to constipate the cancer cell to death," says Romanio.

THE TROJAN EXOSOME Beyond their potential connection to cancer, what has thrown the spotlight on exosomes is a recent paper authored by Hildreth, his Johns Hopkins Medical Institutions colleague Stephen J. Gould, and Gould's student Amy M. Booth. Part of the attention stems from the paper's snazzy title-"The Trojan exosome hypothesis"--and its implication that the AIDS virus can sneak into cells via exosomes, much as Greek soldiers invaded Troy by hiding in a wooden horse.

The hypothesis came about after Gould, a cell biologist, and dreth, who studies HIV, began discussing the fact that each HIV is adorned with a variety of human-cell proteins. "It's now very clear that the preponderance of proteins in the membrane of the virus is host proteins, says Hildreth.

Most biologists simply attributed that phenomenon to the virus particle budding out of an infected person's cell and becoming coated with part of that cell's protein-studded membrane. But if that's the case, it's hard to explain why some abundant cell-surface proteins are all but absent on HIV, while relatively rare cell-surface proteins are present at high concentrations. Moreover, an array of proteins from inside human cell is also present in the

Over the past few years, several research groups have shown that HIV can assemble itself inside cells in the same sacs where exosomes From "It became obvious to us that retroviruses generally, and HIV in particular, evolved to become cargo of this pathway," says Hildreth.

HIV usually infects cells when proteins in its envelope attach to specific proteins on an immune cell and initiate fusion between the virus and cell Yet there have been controversial reports that the AIDS virus doesn't need its envelope proteins to infect cell and that the virus can infect nonimmune cells lacking the molecular handholds that the envelope proteins grip.

The Trojan-exosome hypothesis could offer an explanation for these phenomena, since the vesicles have their own methods of attaching to cells and dumping their cargo inside. "If we model HIV as an exosome, it means that this virus possesses an ancient, preexisting method to fuse to [cell] membranes," says Hildreth.

The Trojan-exosome hypothesis has drawn a mixed reaction from AIDS researchers. If exosomes were significant players in HIV infection, then cells everywhere in the human body would be infected, and that's not the case, says Robert Gallo of the Institute of Human Virology in Baltimore. Also, newly developed drugs that block HIV's envelope proteins from binding or fusing with cells wouldn't slow infections as dramatically as they seem to, he adds.

"I can't accept [exosomes] as being of real significance," concludes Gallo.

Gould counters that he and Hildreth don't claim that exosomes account for most of HIV's virulence in patients with AIDS, but they do believe the vesicles provide a way for the virus to hide out in the body.

"If your definition of clinical relevance is whether this alternative, backdoor pathway has any relevance to the virus' ability to persist in the face of potent antibody and [immune cell] responses, then it is relevant," Gould argues.

A RETROVIRUS IS BORN Wesley I. Sundquist and his group at the University of Utah in Salt Lake City have provided much of the evidence showing that the release of HIV from infected cells depends on the same intracellular pathway used to assemble exosomes. Although he remains skeptical about the importance articles in the spread of the AIDS virus, Sunduquist says he's intrigue by another proposition that Gould and Hildreth the floated: Exosomes may have spawned retroviruses, such as HIV.

Scientists have previously suggested that these viruses evolved from genetic sequences known as retrotransposons, which show up in most animals' cells. Indeed, the sequences make up about 10 percent of the human genome. The only apparent function of retrotransposons, which are known as selfish DNA, is to make copies of themselves. They do this by compelling the cell to make an RNA template of their DNA and then using that RNA to make another strand of DNA that is stitched back into the cell's DNA.

Similarly, HIV and other retroviruses have genes made of RNA. The retroviruses infiltrate cells by making a DNA copy of their viral RNA, and the DNA copy gets integrated into a host cell's DNA. Consequently, scientists have proposed that retroviruses evolved from retrotransposons. Hildreth says that a mutation that redirected a retrotransposon's RNA into the exosome-assembly pathway could have offered a ready-made way to package that RNA, get it out of a cell, and deliver it into another cell. Voila a retrovirus.

"By this one simple [mutation], we've solved the problem of how retrotransposons became retroviruses," says Hildreth. Over time, retroviruses could have evolved envelope proteins that would have enabled them to more efficiently infect cells and target specific cell types, the researchers add.

"I could kick myself for not thinking of the idea that viruses likely evolved from exosomes," says Sundquist. "It is a clever and attractive idea."

The most dramatic conclusion reached by Gould, Booth, and Hildreth concerns HIV vaccines. The researchers argue that if HIV spreads, in part, via exosomes, then current experimental vaccines may not offer enough protection because they raise an immune response only against viral-envelope protein. These scientists call for vaccine developers to look into immunizations based on a phenomenon called alloimmunity.

The quick-acting alloimmune response occurs when a person rejects a transplanted organ. It happens because the donated organ's surface molecules, particularly the MHC proteins, don't match the recipient's MHC proteins closely enough.

Hildreth and his colleagues at Johns Hopkins argue that this alloimmune response actually represents the body's original way of defending against retroviruses. The researchers theorize that because HIV uses the exosome pathway to break out of a host cell, newly formed viruses should bear MHC proteins from the people they're infecting at the time. If the viruses then move in to a person with dissimilar MHCs, this second person should quickly mount an alloimmune response to the virus. Several lines of evidence confirm that HIV transfers less efficiently between MHC-mismatched people than between others, Hildreth points out.

Consequently, the researchers endorse a previously discussed but largely dismissed vaccine strategy: immunizing populations with MHC-based vaccines that would boost the alloimmune response. "We should start to see a dramatic drop in the rate of [HIV] transmission," says Hildreth. "This may be a very simple way to slow the virus down:'

Gallo questions whether HIV researchers will be discussing the Trojan-exosome hypothesis a year from now, but Gould and Hildreth argue that their proposals will have a lasting impact. "There's a tremendous amount to be learned about the biology of exosomes,' says Hildreth. "We're hoping this will shine a bright light on the exosome field. The more we learn about exosomes, the more we will learn about HIV."