|Home | About | Journals | Submit | Contact Us | Français|
The myeloproliferative neoplasms (MPNs) are a group of clonal disorders that arise from a pluripotent hematopoietic stem cell and are characterized by excess cellular proliferation. These disorders tend to be chronic in nature and can terminate over time into a bone marrow failure syndrome characterized by marrow fibrosis or transform into a leukemic phase. MPNs are predominantly diseases of the elderly and this is one reason why until very recently the standard treatment was supportive care. The only curative modality for these disorders is allogeneic hematopoietic cell transplantation. The introduction of reduced-intensity conditioning regimens now allows this life-saving therapy to be offered to elderly patients who were previously considered ineligible for high-dose conditioning owing to age or comorbidity. In this review, we will summarize the current strategies and future directions regarding the use of reduced-intensity conditioning regimens in the treatment of MPNs.
The myeloproliferative neoplasms (MPNs) consist of a group of clonal disorders that are thought to arise from a malignant hematopoietic stem cell. These disorders are rare, with an annual incidence of six to ten out of 100,000 people, and generally affect the elderly with a peak incidence in the fifth to seventh decade of life . In the initial stages of the disease, there is abnormal proliferation of the myeloid compartment, and bone marrow examination reveals hypercellularity. In general, these disorders are considered to be chronic in nature, as patients can live with these neoplasms for years to decades. However, they all have the common ability to terminate in a marrow failure syndrome characterized by marrow fibrosis, or transform into acute leukemia .
The most recent WHO classification of MPN is shown in Table 1 . The four classic and most commonly diagnosed MPNs are chronic myelogenous leukemia (CML), polycythemia vera (PV), essential thrombocythemia (ET) and primary myelofibrosis (PMF), and they will be discussed later in this article. In addition, very rare disorders, such as chronic eosinophilic leukemia, chronic neutrophilic leukemia and mastocystosis are also MPNs. There are very few data regarding the role of transplantation in these rarer MPNs, and they will not be considered in this review.
The majority of the MPNs have mutations/rearrangements involving receptor or cytoplasmic tyrosine kinases that are involved in the pathogenesis of the neoplasm (Table 1) . There has been great interest in targeting these abnormalities in the hope of curing the disease. The best example of this is CML. Prior to the last decade, the standard of care for patients with CML who had HLA-matched donors was to proceed to hematopoietic stem cell transplantation (HCT) as soon as possible, as this was the only known long-term curative therapy . With the introduction of effective tyrosine kinase inhibitors, such as imatinib, dasatinib and nilotinib, in the last decade, the majority of CML patients do not proceed to HCT early after diagnosis. However, tyrosine kinase inhibitors are not curative, and HCT still plays an important role in the treatment of this disease, as will be discussed below .
Polycythemia vera, ET and PMF are related disorders that often have considerable clinical overlap . In 2005, four groups simultaneously and independently identified an activating mutation (V617F) in the JAK-2 kinase in the majority of patients with PV, as well as in a significant percentage of patients with ET and PMF [7–10]. Although the JAK-2 mutation is important in the pathogenesis of these disorders, other factors also contribute to the varying clinical spectrum of MPN . Although patients can be managed conservatively for many years, the majority of patients will develop marrow failure or leukemia . Although JAK-2 inhibitors have shown activity in vitro  and current clinical trials show some promise [14–20], HCT remains the only curative therapy .
Unfortunately, the majority of patients with MPN are elderly and cannot withstand the toxicities associated with conventional high-dose conditioning regimens. In the last decade, reduced-intensity conditioning (RIC) regimens have been developed in an effort to extend this curative therapy to the elderly . Readers are also referred to a number of excellent reviews that have been published recently on this topic [22–29]. Observations in the late 1970s to early 1980s led to the rationale for the development of RIC regimens . Initially, it was thought that high-dose chemotherapy or radiation was an essential component of eradicating the underlying disease. However, it was later noted that the development of graft-versus-host disease (GVHD) after HCT was strongly associated with the lowest relapse rates [31–33]. The observation that recipients of T-cell-depleted grafts had a higher risk of relapse, and studies in which patients who relapsed post-HCT and achieved remission after donor lymphocyte infusion (DLI), both confirmed that donor T cells play an important role in the graft-versus-tumor (GVT) effect [34–37]. Furthermore, studies in which conditioning regimens were intensified in the hope of decreasing relapse only yielded more toxicity and did not impact survival [38,39]. These findings have led to a major paradigm shift in the field of HCT towards RIC regimens, which rely heavily on maximizing GVT to eradicate disease while minimizing toxicity .
Although the omission of high-dose myeloablative therapy is associated with significantly decreased nonrelapse mortality (NRM), in some cases this comes at the cost of higher relapse rates. This is especially true when the tumor is associated with rapid proliferation or the burden of disease is very high at the time of HCT . Therefore, in an effort aimed at minimizing NRM as much as possible without increasing relapse rates, a spectrum of RIC regimens have been introduced, with varying intensities ranging from very low, to moderate, to very high.
There is broad agreement that RIC regimens are characterized by a few general principles: these regimens cause reversible myelosuppression when given without stem cell support, they have limited nonhematopoietic organ toxicity, and they will result in a state of mixed chimerism at the time of the first evaluation post-transplant . Based on results in a preclinical canine model, investigators at our center have further decreased conditioning intensity and developed a non-myeloablative conditioning regimen that consists of fludarabine, 2-Gy total-body irradiation (TBI) and post-grafting cyclosporine, and mycophenolate mofetil (MMF) [43,44]. In contrast to conventional RIC regimens that include some amount of debulking chemotherapy, the aim of non-myeloablative conditioning is to provide profound immunosuppression to prevent graft rejection, and this type of transplant relies entirely on GVT to eradicate the underlying malignancy .
The Center for International Blood and Marrow Transplantation (CIBMTR) definition for retrospective analysis defines a RIC as any regimen that includes the following: a TBI dose of 5 Gy or less if given as a single fraction or 8 Gy if given in a fractionated manner, a busulfan dose of 9 mg/kg or less, a melphalan dose 140 mg/m2 or less, a thiotepa dose no greater than 10 mg/kg, and the BEAM regimen (carmustine, etoposide, cytarabine and melphalan) . A recent survey of transplant professionals indicated that the majority of respondents agreed with the general principles and most of the CIBMTR definitions; however, only 32% considered the BEAM regimen to be a RIC regimen. The conclusion of this study was that a consensus needs to be achieved on what constitutes a myeloablative regimen so that RIC regimens could be better defined. Current efforts are underway by the CIBMTR and European Bone Marrow Transplant Registry (EBMT) to address these issues .
As discussed earlier, in the pre-tyrosine kinase era, it was recommended that CML patients proceed to HCT as soon as possible . However, the introduction of effective tyrosine kinase inhibitors has led to a dramatic reduction in the number of transplants performed for CML. Thus, currently, patients with CML only proceed to transplant if they have failed or are intolerant of these drugs . However, tyrosine kinase therapy needs to be taken for the patients entire life, is very expensive and is not curative. Furthermore, patients who present in accelerated phase or blast crisis very rarely achieve a complete cytogenetic response with imatinib . Treatment with second-generation tyrosine kinase inhibitors, such as dastinib, can increase the rate of complete cytogenetic response to 20–30%, but the median duration of this response is very short, usually a few months [46,47]. Thus, HCT still plays a large role in the treatment of CML in these patients, especially in advanced disease [46,48].
In comparison with acute leukemia, there are only a few studies of RIC regimens in CML, but this disease is quite sensitive to the GVT effect, as demonstrated by the success of DLI ; therefore, RIC might be particularly effective in this disease. Results of currently published RIC studies are summarized in Table 2 [49–54]. It should be noted that the majority of these studies were performed in the pre-tyrosine kinase inhibitor era.
These studies all suggest that RIC regimens are effective in the treatment of early-stage CML with low NRM. However, results of RIC in advanced CML are not as encouraging. This is particularly highlighted by a study from the MD Anderson Cancer Center (MDACC; TX, USA), which reported long-term results of patients with CML treated with a variety of fludarabine/melphalan-based RIC regimens . The majority of patients had advanced disease: chronic phase (CP)1 (20%), CP2 (26%), accelerated phase (AP; 45%) and blast crisis (BC; 8%). Early toxicity was seen with NRM of 33, 39 and 48% at 100 days, 2 and 5 years, respectively. The 5-year overall survival (OS) and relapse-free survival (RFS) were 33 and 20%. Data from the EBMT and the Seattle Consortium are also similar in that patients with advanced disease have poor outcomes. These results highlight the need to optimize RIC with addition of novel agents for more advanced disease.
It is also well known that minimal residual disease post-conventional HCT using quantitative PCR for bcr–abl transcript level can accurately predict relapse. In the myeloablative setting, detection of bcr–abl transcripts 6–12 months post-HCT is an independent predictor of relapse and would warrant further intervention, such as DLI or tyrosine kinase therapy. Although molecular remission can be achieved within 1 month using even non-myeloablative conditioning, in general, it is delayed with reduced-intensity approaches [55,56]. This is usually because of the fact that many patients develop mixed chimerism after RIC. In the majority of cases, conversion to full donor T-cell chimerism is associated with clearance of minimal residual disease [49,51,56]. The kinetics of this conversion to a full donor chimera is slower when using RIC versus conventional myeloablative conditioning and usually depends on the intensity of conditioning. Thus, patients may have detectable bcr–abl transcripts up to 1 year post-transplant but eventually achieve a complete molecular remission [29,51]. As discussed earlier, the stage of disease at transplant also affects the outcomes post-RIC, as patients with advanced disease often have early detection of minimal residual and rapid relapse after transplant. Therefore, molecular remission can be induced using RIC and non-myeloablative approaches, and is useful in predicting relapse. However, the kinetics are likely to depend on the stage of disease prior to transplant and the intensity of the conditioning regimen.
The majority of patients with PV and ET can be managed conservatively for many years and are not routinely recommended to proceed to transplant until they develop myelofibrosis or leukemia. Patients with myelofibrosis can also have a chronic course and the timing of recommending a transplant can be a difficult decision. A variety of scoring systems have been developed to help determine prognosis in patients with myelofibrosis, including the Dupriez (Lille), Mayo, and International Working Group (IWG) scoring systems for myelofibrosis (Table 3–Table 5) [57–59]. Patients with intermediate- or high-risk disease are generally considered to have poor long-term prognosis and should be considered for HCT . The majority of conventional high-dose transplant series for myelofibrosis report a high rate of NRM up to 40%, which increases in older patients especially. Specifically, there have been increased rates of veno-occlusive disease of the liver owing to pre-existing hepatosplenomegaly . A study from our center reported results on 104 patients transplanted for a diagnosis of primary or secondary myelofibrosis . Although the majority of patients received myeloablative regimens, most commonly BuCY, nine patients received non-myeloablative conditioning with fludarabine and 2-Gy TBI. This cohort comprised of elderly patients up to the age of 70 years who had other comorbid conditions. The overall survival in this group was 56%, which did not differ statistically from patients who received a conventional transplant. These results were encouraging and suggested that a GVT effect does exist in MPN.
Thus, there has been considerable interest in RIC in this disease. The Myeloproliferative Disease Consortium is currently conducting a prospective Phase II trial evaluating RIC in myelofibrosis, and the results of this study should add to our understanding.
Results of currently published RIC studies are summarized in Table 6 [62–67]. These studies suggest that RIC is effective in the treatment of myelofibrosis with low NRM. There have been concerns regarding engraftment using RIC given that most patients have considerable marrow fibrosis; however, published data suggest that engraftment is not a problem with RIC or even non-myeloablative conditioning . Similar to conventional HCT, rapid regression of marrow fibrosis is also seen after RIC . There has also been some concern that splenomegaly, especially when it is profound, can increase the risk of graft failure or can lead to delayed engraftment [69,70]. However, recent studies suggest that even with profound splenomegaly, engraftment was not a problem and that spleen size decreased over time and paralleled the reduction in marrow fibrosis . Although these results are encouraging, they still highlight the need for improvement in patients with advanced disease.
Although results of RIC regimens are promising in MPN, results in advanced disease are still poor, similar to high-dose regimens. A number of issues are currently being studied to optimize RIC for better results. These are reviewed below.
It has long been known that the use of oral busulfan leads to varying levels of plasma concentrations owing to differences in bioavailability and pharmacokinetics. Patients who have low concentrations have a higher risk for relapse/graft rejection and those with higher levels have more toxicity. A study from our center identified that targeting of oral busulfan to a level of 900 ng/ml resulted in better results . Similarly, the use of intravenous busulfan has been reported to reduce hepatotoxicity . In addition, alternative agents with better toxicity profiles are being incorporated into conditioning regimens. The alkylating drug treosulfan has significant cytotoxicity against malignant myeloid cells while having limited non-ematologic toxicities. Treosulfan is also more advantageous compared with oral busulfan as it has a more predictable pharmacologic profile [74,75]. Current studies are underway in the use of this drug in conditioning regimens in a variety of malignancies [76,77]. Cyclophosphamide is also being eliminated in more RIC regimens as it has been demonstrated that metabolites of this drug may increase morbidity and mortality . Fludarabine, a potent immunosuppressive agent that has a favorable toxicity profile, is being incorporated into the majority of RIC regimens as it favors engraftment of donor T cells [79–81].
Targeted therapies are also being added to conditioning in the hope that these agents will decrease relapse rates without increasing NRM. Investigators at the FHCRC recently reported results with a RIC regimen of 2-Gy TBI, fludarabine and I131 radiolabeled anti-CD45 antibody in elderly patients with high-risk leukemia or myelodysplasia . The use of this radiolabeled antibody allows for delivery of very high-intensity therapy to the target organ, while sparing toxicity in nonhematopoietic tissues. This regimen was used to transplant 58 patients (median age: 63 years) with advanced MDS and AML from HLA-matched related (n = 22) and unrelated (n = 36) donors. A total of 20 patients (35%) had refractory disease that either was primary refractory (n = 8) or was in relapse after prior remission (n = 12) at the time of transplant. According to Southwest Oncology Group criteria, 55% of the patients had high-risk cytogenetic abnormalities. The maximum tolerated dose of the antibody was estimated to be 24 Gy delivered to the liver, which allowed us to deliver an average of 36 Gy to the marrow space and 102 Gy to the spleen. After a median follow-up of 2.6 years, the 1-year probabilities of OS, RFS and NRM were 41, 40 and 22% in this very high-risk population of elderly patients which is very encouraging. Similarly, investigators at MDACC have incorporated gemtuzumab ozogamicin, a monoclonal antibody against the CD33 antigen conjugated to calichemicin, into a RIC regimen of fludarabine and melphalan, and early results show some promise . Further preclinical studies are also being conducted to identify the optimal radioisotope/toxins to use in conjugation with antibodies [84–87]. More studies will be needed to determine if these therapies will be effective in MPN.
Finally, the addition of tyrosine kinase inhibitor therapy prior to and after transplant is also being studied to determine if relapse rates can be reduced. In CML, the use of tyrosine kinase inhibitors is clearly effective in controlling disease in patients who have relapsed following allogeneic transplant [88–90]. After RIC, the maximal GVT effect does not occur immediately, and rapidly proliferative diseases may relapse early . Thus, there has been interest in the use of adjunctive therapies, such as tyrosine kinase inhibitors, to modify the kinetics of the disease post-transplant until the GVT effect can occur. Furthermore, a study from our center reported that prophylactic administration of imatinib post-myeloablative transplant for bcr–abl-positive diseases was safe and effective at reducing relapse . One publication from the UK reported results on 22 patients transplanted with a RIC regimen on fludarabine–busulfan–alemtuzumab followed by 11 months of post-transplant imatinib . The rate of acute GVHD was 5%; there was no chronic GVHD seen with NRM of 4% at 1 year. Of the 21 patients who completed 11 months of imatinib, 15 patients relapsed and required DLI. With a median follow-up of 3 years, OS was 87%. There is also some evidence that imatinib is effective in the treatment of chronic GVHD and may explain the very low rates of GVHD seen in this study . Further studies are needed to determine the optimal tyrosine kinase inhibitor and the length of time needed for treatment, but this strategy does appear quite promising. Furthermore, novel drugs, such as aurora kinase inhibitors, which are effective in imatinib-refractory disease, will also need to be studied . Similarly, several JAK-2 inhibitors are now being introduced into clinical trials and do have significant activity in JAK-2-positive MPN [13–20,95]. Studies will be needed to determine if the use of JAK-2 inhibitors post-transplant will also be effective at decreasing relapse rates.
As discussed earlier, the first evidence of a GVT effect in humans was shown in the late 1970s . In this study, patients with acute and chronic GVHD after high-dose conditioning had lower rates of relapse when compared with patients who had no GVHD. Furthermore, patients who only developed chronic GVHD had the highest rates of survival after HCT . A recent retrospective analysis from our center looking at 322 patients who under-went non-myeloablative HCT, also identified chronic GVHD as being highly associated with decreased risk of relapse . In this analysis, acute GVHD was associated with higher NRM but did not influence relapse rates. Similar analyses from other groups looking at RIC transplants in a variety of hematologic malignancies also suggest that only the occurrence of chronic GVHD is associated with better disease-free survival [50,96–98]. Consequently, attempts to minimize severe, acute GVHD without affecting chronic GVHD may decrease NRM even further, without impacting survival when using RIC.
A common method used to decrease acute GVHD is the use of antibodies targeted against donor T cells . The most commonly used agents are anti-thymocyte globulin or alemtuzumab, which targets the CD52 antigen found on T lymphocytes and natural killer cells . Although these antibodies are quite effective at decreasing the incidence of GVHD, the rates of relapse/disease progression are also higher [50,100–102]. Furthermore, the rates of certain infectious complications can be significantly higher, particularly CMV infection with the use of alemtuzumab .
Another method proposed by the Stanford group is the use of total lymphoid irradiation of 8 Gy combined with ATG . This regimen was translated from murine studies where the conditioning prevented mice from developing GVHD by increasing the proportion of regulatory natural killer T cells . This regimen was used to treat 111 patients with hematologic and lymphoid malignancies using HLA-matched related and unrelated donors. The rates of acute GVHD were 2% for related and 10% for unrelated donors. The incidence of chronic GVHD was 27% and NRM at 1 year was 4%. The 36-month probabilities of OS and RFS were 60 and 40%, respectively. With relatively short follow-up, the results suggest that this regimen is associated with very low levels of acute GVHD while apparently preserving GVT . Further studies will be needed to determine if this regimen would be effective in MPN.
The current drugs that are most commonly used for GVHD prophylaxis are the calcineurin inhibitors, methotrexate and MMF . Although these drugs are effective, GVHD continues to be a problem, and the use of newer pharmacologic agents that limit GVHD without compromising GVT has generated a great deal of interest. In a recent retrospective analysis, the use of rapamycin as GVHD prophylaxis after RIC was associated with improved survival in patients with lymphoma . In addition to direct anti-tumor effects, rapamycin may also affect Tregs, and this drug is currently being studied further . The proteasome inhibitor bortezomib and the histone deacetylase inhibitor SAHA have both been studied in murine models of transplantation and have been shown to reduce GVHD without compromising GVT [108,109]. However, these studies need to be interpreted with caution as similar effects of keratinocyte growth factor in murine models did not translate to human clinical trials [110,111]. In a recent Phase I study, the combination of bortezomib, tacrolimus and methotrexate was evaluated as GVHD prophylaxis after HLA-mismatched HCT and the rate of acute GVHD was very low at 13% . These data are quite promising, but further studies with these novel agents in human trials will be needed to show their efficacy.
Finally, the infusion of mesynchymal stem cells (MSCs) is also being studied in modulating GVHD. These are fibroblastoid cells derived from the marrow and have potent immunomodulatory properties in animal models of transplantation [113,114]. The use of MSCs in humans was first reported in 2004 in a case report where the infusion of haploidentical MSC resulted in resolution of steroid-refractory GVHD . At present, clinical trials are underway to study the use of third-party MSCs in the treatment of GVHD and their effects on GVT.
Given that the majority of patients with MPNs are elderly, comorbid medical conditions are often a major cause of NRM. Sorror et al. have developed a useful tool, the Hematopoietic Cell Transplantation-Specific Comorbidity Index that risk-stratifies patients into several groups to help identify patients that would have increased NRM . Patients with higher comorbidity scores have higher rates of NRM when conditioned with increasing intensity .
Over the last decade, the introduction of RIC has made significant strides toward reducing the NRM associated with transplantation. However, relapse rates are still high when using RIC, especially when the stage of disease is advanced. Over the next 5 years, we anticipate that relapse rates will decline, especially as the role of tyrosine kinase inhibitors in the peri- and post-transplant period is better defined. Currently, imatinib is the most studied drug in CML. However, the second-generation tyrosine kinase inhibitors nilotinib and dasatnib are also being used widely, and more drugs are in development. Further studies will be needed to define the ideal inhibitor, dose and duration of therapy. Similarly, the JAK2 inhibitors are just entering the clinic, and further study is required regarding their potential role in maintenance therapy after transplantation. We also anticipate the incorporation of novel chemotherapeutic agents, such as treosulfan, and targeted therapies, such as monoclonal antibodies, into conditioning regimens that aim to maximize efficacy while minimizing toxicity.
The authors would like to thank Bonnie Larson, Helen Crawford and Sue Carbonneau for assistance with the preparation and editing of the manuscript.
Financial & competing interests disclosure
This work was supported in part by PHS grants DK082783, HL36444, and CA18029 from the NIH. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.
No writing assistance was utilized in the production of this manuscript.
Aravind Ramakrishnan, Associate in Clinical Research, Fred Hutchinson Cancer Research Center, Acting Instructor, University of Washington School of Medicine, 1100 Fairview Avenue N, D1-100, PO Box 19024, Seattle, WA 98109-1024, USA, Tel.: +1 206 667 2908, Fax: +1 206 667 6124, Email: gro.crchf@irkamara.
Brenda M Sandmaier, Member, Clinical Research Division, Fred Hutchinson Cancer Research Center, Professor, University of Washington School of Medicine, 1100 Fairview Avenue N, D1-100, PO Box 19024, Seattle, WA 98109-1024, USA, Tel.: +1 206 667 4961, Fax: +1 206 667 6124 ; Email: gro.crchf@iamdnasb.
Papers of special note have been highlighted as:
• of interest
•• of considerable interest