Stem cell transplantation was initially conceived as a way of eliminating leukemia from the bone marrow much as surgery is used to remove solid tumours. However, the marrow needed to be replaced because it was the source of new blood cells, including immune cells, and a person could not survive without their bone marrow. New marrow could be regrown from stem cells obtained from the person (called autologous) before the marrow was wiped out by chemotherapy or radiation, or obtained from a donor (called allogeneic), such as someone with similar genetics. A good match was needed between host and donor to reduce the risk of graft-versus-host disease (GVHD), a potentially fatal immune reaction in which the immune response of the donor graft “rejected” the person receiving the transplant. In the usual case of tissue rejection, the host rejects the donor organ (e.g. a heart). But the situation is reversed in GVHD. Since the bone marrow recipient lacks an immune response, it is the immune response of the tissue graft that rejects the host.
However, as more transplants were performed for CML, a curious phenomenon was seen. If people survived GVHD, they typically did better and were less likely to suffer a relapse of their CML (Weiden and colleagues. N Engl J Med 1981;304:1529-1531). Equally curious was what was seen when efforts were made to lower the risk of GVHD. Since GVHD was caused by the donor immunity, immune components were removed from the donor graft prior to transplantation. What the researchers found that was people were more likely to have a recurrence of CML after receiving a cleaned-up graft (Marmont and colleagues. Blood 1991;78:2120-2130). These findings indicated that the donor graft’s immune response had the potential to seek out diseased cells in the host and destroy them. This was called the graft-versus-leukemia (GVL) effect.
So immune reactions between donor and host were a double-edged sword. GVL had the potential to cure leukemia. GVHD had the potential to be fatal to the host.
As techniques improved in the 1980s and 1990s, one approach was to remove immune cells (called T lymphocytes, a type of white blood cell) from the graft prior to transplantation to minimize the risk of GVHD, then to infuse the donor’s lymphocytes a few months later in a bid to enhance the GVL effect. This technique, called donor lymphocyte transfusion (DTL), was first tried in three people in 1988 and was highly successful (Kolb and colleagues. Blood 1990;76:2462-2465). Two people did develop GVHD, but this reaction resolved over time. All three were maintained in remission for over a decade, demonstrating the long-lasting benefits of the GVL effect (Kolb and colleagues. Blood 2004;103:767-776).
Numerous studies have now reported on how the GVL effect has improved outcomes in people undergoing stem-cell transplantation. In the EBMT-95 survey, GVL was seen in all but one of the groups (Dazzi and colleagues. Blood 2000;96:2712-2716; Porter and colleagues. Biol Blood Marrow Transplant 1999;5:253-261). The sole exception was people who received a transplant from an identical twin – demonstrating that the donor’s immune system needs to be different from that of the host for there to be a GVL effect. Survival at 6 years was about 80%.
GVHD remained a serious complication. In the EBMT-95 study, moderate-to-severe GVHD developed in about 40% of people receiving a transplant (Kolb and colleagues. Blood 1995;86:2041-2050). The rate of GVHD was even higher – about 60% – in the US-97 survey (Collins and colleagues. J Clin Oncol 1997;15:433-444). Generally speaking, more severe GVHD resulted in a better GVL effect, supporting the notion that you can’t obtain the benefits of GVL without the risks of GVHD. However, an important finding in these surveys was that a significant number of people who didn’t develop GVHD were still able to achieve a complete response – raising the possibility that GVHD may not be necessary for a GVL effect.
The trick was (and is) to identify which aspects of immunity contributed to GVL, and which contributed to GVHD. So over the past decade, studies have looked at ways of minimizing GVHD while preserving the GVL effect. As noted above, T lymphocytes are central to the immune response, but there are many different subtypes of T cell. Removing one type (called CD8+) has been shown to reduce the risk of GVHD (Giralt and colleagues. Blood 1995;86:4337-4343). Another method has been to identify T cells that specifically target leukemia cells (rather than healthy host cells), then transplant only those leukemia-specific cells. Perhaps the most “sci-fi” idea is to introduce a “suicide gene” into T cells, which causes the T cell to self-destruct if exposed to a given medication. If GVHD develops, the medication is given so that the T cells causing GVHD are selectively destroyed. An early report in 8 people was promising (Bonini and colleagues. Science 1997;276:1719-1724), but there are many technical hurdles that still need to be overcome.
Immune therapy – enhancing the effectiveness of the body’s immune response to cancers – is rapidly evolving, and we can expect important advances to be made in devising strategies to enable stem-cell grafts to selectively kill leukemia cells without inducing GVHD. The recognition of a graft-versus-leukemia effect has given researchers a goal, and provided new insights on how to make transplantation safer and more effective. Transplant remains the only treatment that offers a possible cure, but it is no longer the best option for most people with CML because of the better safety record of TKIs such as Gleevec. Transplant research has slowed in recent years – there are fewer patients to study because of the success of TKIs – but significant progress has been made. Stem-cell transplants remain an important treatment option for people who don’t respond well to medications, and a source of hope for those with more advanced disease.