Cancer Immunotherapy Inches Forward

PRIMING THE IMMUNE SYSTEM:Thom GravesScientists are working on several approaches to coax the immune system into attacking cancer. Here, artificial antigen-presenting cells are used to stimulate the patient's own tumor-specific T cells.Like a modern army, the human immune system possesses an array of sophisticated cellular and molecular detection systems and weaponry. Against most pathogens these forces mount a formidable defense, but not when the disease in question involves the body's own cell

Amy Adams
Jul 18, 2004

Thom Graves

Scientists are working on several approaches to coax the immune system into attacking cancer. Here, artificial antigen-presenting cells are used to stimulate the patient's own tumor-specific T cells.

Like a modern army, the human immune system possesses an array of sophisticated cellular and molecular detection systems and weaponry. Against most pathogens these forces mount a formidable defense, but not when the disease in question involves the body's own cells. Having learned to recognize and ignore familiar proteins early in life, the immune system mostly turns a blind eye to diseases such as cancer.

As researchers look for ways to circumvent cancer's stealth capabilities, they are taking their cues from the immune system, calling their approach "cancer immunotherapy." Their cellular muse, as it were, is the dendritic cell, whose normal job is to present antigens in a way that draws the attention of cytotoxic T...


In developing aAPCs the gold standard of immune activation remains the dendritic cell itself. This cell's normal job is to scavenge foreign bacteria, viruses, and the occasional apoptotic cell, process the protein or DNA contained within, and present the cellular bits on its surface to specific T cells, which then attack cells bearing that antigen.

In immunotherapy protocols, the job of scavenging and loading cancer-specific antigens is performed ex vivo. The thinking goes like this: Feeding cancer-related antigens to dendritic cells will prime the immune system to attack tumor cells, much like putting a dog onto a scent. Molecular "scents" tested in clinical trials have included markers for prostate cancer (prostate-specific antigen, PSA) and skin cancer (MART-1 and pg100), as well as patient-specific molecules for some blood cancers and antigens for rarer diseases such as pancreatic and renal cancer.1 Some of these trials have reached FDA Phase III.

But removing and growing dendritic cells from each patient is prohibitively expensive. Even if cost weren't an issue, chemotherapy and radiation treatments can weaken the immune system, leaving it lethargic and unable to mount an effective attack. And, the dendritic cells may not survive long enough once reinjected to activate an aggressive response in any event.

Yet despite these shortcomings, dendritic cell approaches are far in front of any aAPC competitor, says Jean-Loup Romet-Lemonne, CEO of Paris-based Immuno-Designed Molecules (IDM). "It's just a question of time before the FDA approves a dendritic cell vaccine, and that will be a boost to the whole field," he says.

Among the leading contenders are Seattle-based Dendreon, Igeneon of Austria, and Biomira of Edmonton, Alberta. All three companies have vaccines in Phase III trials: Dendreon's Provenge for prostate cancer, Igeneon's IGN101 for breast and colorectal cancer, and Biomira's Theratrope for breast cancer. IDM recently completed a Phase II trial for Eladem, also directed at prostate cancer. The vaccine eliminated circulating cancer cells in men after radical prostate surgery to remove the primary tumor.2


Though aAPC approaches lag behind dendritic cells, progress is being made. One promising method fuses irradiated tumor and dendritic cells, creating a hybrid product that contains all the dendritic cell's costimulatory molecules and the tumor cell's surface antigens.3 This, says Trevor, is the system's biggest advantage: Such cells present not only known tumor-related antigens such as PSA, telomerase, and survivin to the T cells, but they also can display as-yet unknown but still highly immunogenic proteins that may be present on the tumor.

In the lab, at least, this approach has shown potential, activating T cells against human tumors in vitro and inducing an immune response and shrinking tumors in animals. But Trevor notes several problems. First, because the hybrid cells contain two and often more nuclei, they are bulky and may not efficiently move from their subcutaneous injection site to the lymph nodes. And in fact, though the vaccines safely induced a T-cell response in phase I/II trials, they have not yet successfully shrunk tumors. Another problem: only 10 percent of the dendritic cells in a given sample form fusions, squandering the hard-to-purify dendritic cells, says Trevor.

Eliminating the dendritic cell altogether could solve this latter problem. Among many groups trying this approach, London-based Onyvax has created tumor cell vaccines consisting of irradiated tumor cells from different stages of the disease. The vaccine is injected under the skin. The body's dendritic cells then do their normal job, engulfing the foreign injected cells and educating the T cells as to what they've found. The company is currently testing the approach in Phase I/II trials for prostate and colorectal cancers.

Anthony Walker, CEO of Onyvax, says other methods of activating T cells may have merits, but simply injecting tumor cells has the benefit of simplicity. "Our view is that you start with a simple product. If that doesn't work, then you move on to the others," he says. Trevor, meanwhile, continues to improve the hybrid system. Recently she developed a more effective way of fusing the tumor and dendritic cells.3


Tumor cell vaccines and hybrid fusions both rely on the vaccines activating T cells in vivo. But successful implementation requires addressing several practical issues, including how long the tumor cells or hybrids survive in the blood stream, whether they wend their way to the lymph nodes, and how often they must be reinjected to maintain T-cell activation. Some researchers, therefore, go the ex vivo route, extracting T cells, activating them in culture, and then reinjecting the activated lymphocytes back into the patient. As an added benefit this approach provides a way to tweak activation conditions.


Courtesy of Xcyte Therapies

Xcyte Therapies has tested its Xcyte Dynabeads in phase I/II clinical trials of patients with chronic lymphocytic leukemia, multiple myeloma, and prostate cancer. Bearing anti-CD3 and anti-CD28 monocloncal antibodies, these super-paramagnetic particles approximate the size and shape of APCs.

Although details vary, all methods rely on major histocompatibility (MHC)-bound antigenic peptides to expand the appropriate T-cell population. Also required is stimulation through CD3 or CD28, cytokines such as IL-2 or IL-15, and nonspecific adhesion molecules including ICAM-1 and LFA-3 (both of which are involved in CD28 signaling).

Carl June at the Abramson Cancer Center of the University of Pennsylvania developed one such implementation using the erythromyeloid cell line K562.4 These cells endogenously produce the costimulatory molecules LFA-3 and ICAM-1. June and his colleagues engineered the cells also to express 4-1BBL, a ligand for the T-cell receptor 4-1BB (CD137) that enhances the survival of CD8+ T cells, and the Fc gamma receptor (CD32), onto which they loaded anti-CD3 and anti-CD28. Within a week of exposing these cells to a mixed T cell population, flow cytometry showed that the K562 cells had died, leaving behind an essentially pure population of T cells.

This method was the first to stimulate large numbers of expanded T cells and to maintain long-term expansion of the CD8+ population. But these K562 cells activate all T cells in a population, not specific subpopulations. James Riley, June's colleague at the University of Pennsylvania, says antigen-specific T cells can be expanded in two ways using this technique: Either purify the antigen-specific T cells before expanding them, or engineer K562 cells that express human leukocyte antigen (HLA) molecules. By selecting a cell line expressing the appropriate HLA (i.e., MHC class I) type, the researchers could then attach the antigen and expand antigen-specific T cells from a mixed population. Riley says the group has shown that the technique is effective, "Now it's a matter of picking the tumor model to work on," he says. The group has yet to use these cells in a human trial.

Two other cell-based aAPCs build on Drosophila cells or mouse fibroblasts. In the Drosophila system, cells require pulsing with HLA and peptides in order to activate T-cell expansion, in addition to added cytokines or feeder cells for proliferation.5 Although the fly cells die at 37 degrees, which limits how long the T cells can be exposed to the aAPCs, two out of 10 patients in a Phase I trial showed some tumor regression.5

Michel Sadelain at Memorial Sloan-Kettering Cancer Center in New York has developed a mouse fibroblast-based system that effectively amplifies CD8 T+ cells.6 The team engineered the cell line to express the human costimulatory molecules B7.1 (CD80), ICAM-I, and LFA-3, along with an HLA molecule and human β2-microglobulin. The team has created fibroblast lines that express antigens from melanoma, as well as human telomerase reverse transcriptase, which is expressed by many tumors, and a Wilms tumor antigen. According to Sadelain, these cells are easy to handle, much like the K562 cells, and unlike Drosophila cells, they don't require antigen pulsing.


Some scientists have moved away from cells entirely, using beads as their antigen-presenting vehicles instead. Seattle-based Xcyte Therapies, whose Xcellerate technology activates T cells with paramagnetic beads bearing antibodies to both CD3 and CD28, announced results from Phase I/II trials at the American Society of Clinical Oncology annual meeting this past June. "Antitumor effects and sustained increases in lymphocyte counts were observed in all of the clinical trials," Xcyte CEO Ron Berenson said in a press release announcing the results. The company has since initiated a Phase II trial for patients with multiple myeloma and non-Hodgkin lymphoma.


© 2004 Nature Publishing Group

Researchers have developed several varieties of aAPCs, including both antigen-specific (a, d) and non-specific (b, c) varieties. (Reprinted with permission from Nat Biotech, 22:403–10, 2004)

Jonathan Schneck at Johns Hopkins School of Medicine in Baltimore also works with beads.7 According to Schneck, this approach has many advantages. First, by creating beads with the six most common MHC types, the group can produce off-the-shelf antigen-presenting reagents that cover 95 percent of the population. "They are easy to make, simple to manufacture, have a long shelf life, and come at an order of magnitude lower cost," Schneck says. And they appear to work.

Schneck's system involves beads coated with an HLA-immunoglobulin fusion protein and the costimulatory anti-CD28 antibody. Loading the HLA with melanoma-specific peptides generated clinically relevant levels of antigen-specific CD8+ T lymphocytes in vitro.

This bead is the first version of a versatile system that could eventually include costimulatory molecules similar to June's K526 cells, says Schneck. "There may be many second or third generations that might be better, but the leap from cell-based to bead-based is so substantial that we don't want to lose track of the goal of getting this into a clinical trial," he says. By adding different costimulatory molecules they may be able to better overcome tolerance or activate different components of the immune system, he says. Other groups, for instance, have used anti-CD3-coated beads to activate CD4+ rather than CD8+ lymphocytes. Plus, the bead already comes in a clinical grade and can be produced according to good manufacturing practice (GMP) standards.

But beads are not the only artificial substrates available; some scientists work with liposomes and exosomes instead. The former are synthetic spheres created by mixing proteins with lipids in a test tube; the latter are biologically derived, retaining the antigens, MHC, and costimulatory molecules of the original cell. In constrast to beads, the particles (both roughly 80 nm) retain a fluid lipid layer that creates a more natural interface with the T cell.

Unilayer liposomes pulsed with antigen have activated mouse antigen-specific CD4+ cells in vitro,8 while exosomes have been used to deliver antigen to dendritic cells or to activate T cells in animal studies.9 Neither system has been explored in much detail, though Menlo Park, Calif.-based Anosys has tested exosome therapies in two phase I trials.


Any truly successful immunotherapy treatment involves turning the immune system against the body's own tissues. "You don't want to wake up the immune system for some antigen that you don't want the immune system to react to. That's why we didn't use the brain as a model," says Romet-Lemonne.

Many immunotherapy targets include tissues that aren't critical, such as the prostate, or that can tolerate a mild autoimmune reaction, such as the skin. That's in contrast to the colon, where an autoimmune reaction may produce a disease as devastating as the cancer itself. At this time autoimmunity remains a desirable problem, says Trevor. "The day you've seen autoimmunity, that's the day you've seen success," she says. For now, autoimmunity remains rare.

Even if immunotherapy does reach its theoretical potential, it is unlikely to rival traditional chemotherapy, radiation, or surgery. Walker points out that the immune cells are better at preventing new cancer cells from surviving rather than eradicating an entire tumor. "You want to start with minimal disease and keep it there," he says. He expects immunotherapy in some form will be most effective after a cancer has been removed by conventional methods. This approach has many advantages, not the least being that, since many cancers reappear far from where they originated, the immune system can be on guard throughout the body.

Which aAPC is most likely to rival dendritic cells depends on a number of factors, including how effectively they expand antigen-specific T cells, how well those T cells eliminate cancer cells, and cost. So far, no technique has a clear edge. "At this point I think it's silly to throw any strategy away," Pennsylvania's Riley says. Instead, people in the field should focus on the best way of reaching the common goal. "In the end everybody is happy if we cure cancer."

Amy Adams is a freelance writer in Mountain View, Calif.

Article Extras

Related Articles

Feature | Vaccines: Victims of Their Own Success?

Research Vision | Panning for T-cell Gold

Profession | Ongoing Battle over Transgenic Mice