The science of bone marrow transplantation arose from the realization that irradiation causes lethal bone marrow failure (1), thus generating the idea that intensifying irradiation or use of cytotoxic drugs at myeloablative doses could cure bone marrow diseases. The discovery that the infusion of allogeneic hematopoietic stem cells can “rescue” patients from lethal marrow toxicity and give rise to a new donor-derived immunohematopoietic system resulted in successful treatments for patients with malignant or nonmalignant hematologic disease (2, 3). Because even high doses of irradiation or chemotherapy cannot eradicate all malignant cells in all patients, ever more intensive conditioning regimens dominated efforts in clinical marrow transplantation until the late 1990s.
Copernican revolution in transplantation: dose deintensification
The key observation that led to the emergence of nonmyeloablative allogeneic hematopoietic stem cell transplantation (NST) was the discovery of the immune-mediated allogeneic graft-versus-tumor (GVT) effect (4, 5). The importance of this effect was demonstrated by an increased rate of leukemia relapse in patients receiving transplants in which identical twin donors were used or T cell depletion of donor marrow was performed (5). Additional evidence for the role of donor cells in GVT effects came from findings that recurrent leukemia could be successfully treated solely by infusing additional allogeneic lymphocytes (6). Preclinical dose de-escalation studies in canine marrow transplant models played an important role in showing that doses of total body irradiation as low as 200 cGy in combination with cyclosporine and mycophenolate are sufficient for donor engraftment (7). However, in contrast to the 100% donor hematopoietic engraftment typical of myeloablative transplantation, such nonmyeloablative transplants typically resulted in a state of mixed hematopoietic chimerism. The key element for engraftment in NST is the balance between the host immune system reacting in the direction of rejection (host-versus-graft [HVG]) and the donor immune cells reacting in the direction of the host (graft-versus-host [GVH]). Host and donor T cells play a critical role in this delicate balance (8). In HLA-matched sibling transplants, the primary targets of these HVG and GVH reactions are minor histocompatibility antigens on the lymphohematopoietic tissues, which are also significant targets of the beneficial GVT effect.
In contrast to traditional allogeneic stem cell transplantation, in which the patient receives myeloablative conditioning (chemotherapy and/or irradiation), in NST, if donor cells are not infused or are rejected, the recipient's own bone marrow usually recovers spontaneously. This situation leads to the term “nonmyeloablative.” Such transplants are also called “mini” or reduced-intensity transplants. The intensity of NST conditioning regimens can vary from the relatively intensive and more cytoreductive regimens (e.g., based on busulfan or melphalan) to the relatively mild regimens (e.g., based on 200 cGy of total body irradiation, with or without fludarabine). Mild regimens are associated with no significant neutropenia (average neutrophil nadirs of ∼1,000/μl) and have minimal requirements for platelet or red blood cell transfusions.
According to the International Bone Marrow Transplant Registry, since 1997 the number of NSTs has been rapidly growing. By 2002, a total of 7,532 NSTs were registered, and the number continues to grow (9). The procedure is now readily performed in patients who would be not able to tolerate conventional high-dose treatment regimens, such as those who are older or who have decreased organ function or performance status.
Why should we be doing NST for autoimmune disease?
At least 3 components of allografting can contribute to the control of systemic autoimmune disease: the immunosuppressive conditioning regimen, the use of immunosuppressive drugs for GVHD prevention, and the establishment of a donor immune system that may replace a “faulty” recipient immune system. Animal models of spontaneous and induced autoimmune disease have demonstrated a clear potential for allogeneic marrow transplantation from disease-resistant strains to produce sustained remissions of autoimmunity in a susceptible strain (10). The number of case reports of patients with a hematologic disease and a concurrent autoimmune disease has been constantly growing; among these, lasting remission of autoimmunity following allogeneic transplantation has been documented in many patients (11).
The current interest in the use of NST as therapy for autoimmunity is essentially an extension of the earlier results of transplantation for severe aplastic anemia, which has an 80–90% success rate using a nonmyeloablative regimen with cyclophosphamide and antithymocyte globulin (12). The therapeutic capacity of allogeneic transplantation as applied to autoimmunity may lie in the putative graft-versus-autoimmunity (GVA) reaction that is analogous to the GVT effect (11, 13). It is important to emphasize that after successful allogeneic hematopoietic stem cell transplantation, in contrast to solid organ transplantation, most commonly there is no need for indefinite administration of immunosuppressive drugs, and achievement of transplant tolerance is a typical outcome. In sum, NST may offer the possibility of sustained remission or even cure for severe autoimmune diseases.
Why are we not doing it?
The main reason for avoiding allogeneic hematopoietic stem cell transplantation in the treatment of patients with autoimmune disease is the concern about transplant-related mortality and acute and chronic GVHD. With the current NST regimens, the rate of transplant-related mortality is estimated at 5–15%, although further improvements in safety are likely. The key event in the pathogenesis of acute GVHD is the contact between antigen-presenting cells bearing host antigens and donor T cells. The intensity of this reaction is amplified by inflammatory cytokine release precipitated by host tissue damage caused by the intensive conditioning regimens (14). However, tissue damage after NST is minimal. It has been shown that after NST the secretion of tumor necrosis factor α is much decreased compared with that after standard conditioning, and this correlates with a lower incidence of and less severe acute GVHD (15). Some studies suggest a decreased incidence of acute and chronic GVHD after NST (16), but the incidence in most series is still ∼50%. Chronic GVHD is of particular concern due to its potential impact on survival and quality of life, and better prevention strategies are needed. In most cases, however, chronic GVHD ultimately resolves, and patients are able to discontinue therapy with immunosuppressive drugs, although the treatment may take several years.
The higher-than-expected incidence of GVHD after NST, which has been observed primarily in patients with malignancy, may relate to the older age of the recipients studied and to the frequent use of donor leukocyte infusions for the treatment of progressive disease. Although the number of infections appears to be reduced in the early posttransplantation period after NST, the incidence of infectious complications after NST still remains a significant problem. Graft rejection, which represents a rare but lethal complication after myeloablative transplantation, is slightly more common with the current NST regimens but, when it does occur, is typically nonlethal due to recovery of endogenous host hematopoiesis. Rarely, irreversible bone marrow aplasia has been described in NST recipients in whom grafts from matched unrelated donors were rejected. Finally, although most case reports of NST in the setting of autoimmunity have described remission, instances of recurrent autoimmunity have occurred despite full donor engraftment. It is possible that genetic factors may be a limitation when the donor is a close relative or has a predisposition to an autoimmune disease, but the degree of risk of transferring autoimmunity with allogeneic transplantation is not clear (17).
Is mixed chimerism desirable?
The term chimerism in the setting of hematopoietic stem cell transplantation refers to the engraftment or presence of lymphohematopoietic cells of donor origin, typically detected by using techniques of DNA analysis (18). Mixed chimerism, in contrast to full donor chimerism, indicates the presence of both donor and recipient lymphohematopoietic cells. Because mixed versus total donor chimerism can now be reliably generated with specific NST regimens and avoided with other NST regimens, investigators must address an important question: is the state of mixed chimerism after transplantation desirable?
There are at least 3 theoretical advantages of mixed chimerism over full chimerism as an approach to transplantation tolerance (19), as follows: 1) mixed chimerism can be achieved with nonmyeloablative regimens that are less toxic, 2) mixed chimeras have better immune function, and 3) thymic deletion of host-reactive cells occurs to a greater degree in mixed chimeras due to the intrathymic presence of both donor and host antigen-presenting cells. Whether the last of these advantages is relevant in older patients with little thymic function is not clear.
Posttransplant mixed chimerism is not new in allogeneic transplantation (20, 21). Two key observations were made in older studies. First, mixed chimerism is associated with a lower incidence of acute GVHD (but not a lower incidence of chronic GVHD). Second, mixed chimerism is associated with higher rates of graft rejection and malignancy relapse. It is unknown whether an analogous correlation between mixed chimerism and relapse exists in autoimmune disease. Data from animal models of diabetes or lupus suggest that disease remissions can be achieved with strategies that yield mixed chimerism (22, 23).
In this issue of Arthritis & Rheumatism (24), Burt and colleagues report a case suggesting that remission can be obtained in rheumatoid arthritis (RA) in the context of mixed chimerism. With the current followup of this patient (12 months), it is too early to assess whether this state of “remission” remains durable. The conditioning regimen described in this case report can be considered “lymphoablative” and has not previously been used in RA; as such, it is difficult to conclude with certainty the role of allogeneic cells and mixed chimerism in the reported outcome. It is also important to note that this patient received mycophenolate for the first 9 months posttransplantation; this immunosuppressive drug may have contributed to the observed disease remission.
In the case described by Burt et al, multiple therapeutic approaches were used to prevent GVHD, including in vitro T cell depletion of the allograft, in vivo T cell depletion (anti-CD52; CAMPATH-1H), and administration of 2-agent chemoprophylaxis with cyclosporine and mycophenolate. The use of multiple immunosuppressive agents in this case may complicate interpretation of the results of NST. This approach is relatively aggressive in the setting of matched sibling transplantation but may have been warranted in an attempt to reduce GVHD and maximize autoimmune therapy. However, such benefits may need to be considered in balance with any potential negative effect, such as delayed immune reconstitution. From this case report, we can conclude that NST from a matched sibling donor can be performed safely, and that the patient is so far doing well posttransplantation. The mechanism(s) accounting for this result, however, is currently not clear. Nonetheless, the demonstration of a sharp drop in the titer of rheumatoid factor concurrent with increasing donor T cell engraftment is strongly suggestive of an immune-mediated GVA effect.
Should investigators aim for mixed chimerism in the initial NST protocols for autoimmune disease? The answer is probably yes, not because we know the relative advantages of mixed versus full chimerism in treating human autoimmune disease, but because, with the current transplant regimens, it seems to be a safer strategy than pursuing more intensive NST protocols that result in rapid full donor engraftment. It is important to recognize that, in humans, a stable state of mixed chimerism is rarely achievable, and the most common outcome of mixed chimerism is a gradual conversion over a period of months to full donor engraftment. If the strategy of mixed chimerism is to be pursued, it will be essential to determine the minimal degree of donor engraftment required for control of autoimmunity. As has been reported for other nonmalignant diseases (18), it is possible that the percent of donor chimerism required for disease control in autoimmunity may vary depending on the disease being evaluated. Furthermore, chimerism at the site of disease, as well as both T cell and B cell subset chimerism, may also be important variables. Finally, similar to the NST experience in hematologic malignancies, the ease of engraftment after NST for autoimmune disease may depend on the type and amount of immunosuppressive chemotherapy patients received prior to undergoing transplantation.
How to proceed: new partnerships
There will likely be a place for NST in the treatment of severe autoimmune disease. Without doubt, the primary emphasis of future studies should relate to defining transplantation regimens with reduced morbidity and mortality while maintaining sufficient donor chimerism for the desired anti-autoimmunity effect. Due to a higher risk of transplant-related mortality, use of donors other than HLA-matched siblings should be avoided in initial studies. One may question whether RA represents the most appropriate disease for further study, or whether diseases with higher mortality such as systemic sclerosis or systemic lupus erythematosus should be targeted. Unfortunately, systemic sclerosis and lupus have been associated with the highest transplant-related mortality risks in trials of autologous transplantation (up to 12% mortality) (25). RA is a disease in which the toxicity profile of autologous transplantation is favorable, with no deaths reported in 67 patients after treatment with high-dose cyclophosphamide–based regimens (26). RA patients are less likely to have serious vital organ dysfunction, which may account for their favorable risk profile posttransplantation.
Because investigators who typically perform transplantation may not have expertise in autoimmune disease, future efforts in NST will optimally employ equal engagement of rheumatologists very early in protocol conception. National and international coordination would be helpful, with the goal of developing guidelines regarding the selection criteria and clinical and scientific end points. Formation of successful teams of disease experts and transplant experts is essential as we approach what may be the emergence of a new clinical discipline (27).