|Home | About | Journals | Submit | Contact Us | Français|
Graft-versus-host disease (GVHD) is a major barrier to successful allogeneic hematopoietic stem-cell transplantation (HSCT). The chemokine receptor CCR5 appears to play a role in alloreactivity. We tested whether CCR5 blockade would be safe and limit GVHD in humans.
We tested the in vitro effect of the CCR5 antagonist maraviroc on lymphocyte function and chemotaxis. We then enrolled 38 high-risk patients in a single-group phase 1 and 2 study of reduced-intensity allogeneic HSCT that combined maraviroc with standard GVHD prophylaxis.
Maraviroc inhibited CCR5 internalization and lymphocyte chemotaxis in vitro without impairing T-cell function or formation of hematopoietic-cell colonies. In 35 patients who could be evaluated, the cumulative incidence rate (±SE) of grade II to IV acute GVHD was low at 14.7±6.2% on day 100 and 23.6±7.4% on day 180. Acute liver and gut GVHD were not observed before day 100 and remained uncommon before day 180, resulting in a low cumulative incidence of grade III or IV GVHD on day 180 (5.9±4.1%). The 1-year rate of death that was not preceded by disease relapse was 11.7±5.6% without excessive rates of relapse or infection. Serum from patients receiving maraviroc prevented CCR5 internalization by CCL5 and blocked T-cell chemotaxis in vitro, providing evidence of antichemotactic activity.
In this study, inhibition of lymphocyte trafficking was a specific and potentially effective new strategy to prevent visceral acute GVHD. (Funded by Pfizer and others; ClinicalTrials.gov number, NCT00948753.)
Acute graft-versus-host disease (GVHD) is a major cause of death and complications after allogeneic hematopoietic stem-cell transplantation (HSCT). The condition occurs in 30 to 50% of patients receiving HLA-matched transplants from a related donor and in 50 to 70% of those receiving transplants from an unrelated donor.1 The pathogenesis of GVHD is multifactorial, but ultimately, donor-derived T cells recognize recipient antigens as foreign, resulting in activation, expansion, and cytokine release and leading to destruction of host tissues.2 Current therapies for GVHD target T cells and cytokines, often antagonize T-cell–mediated graft-versus-tumor responses, and delay immune reconstitution.3 Preventing GVHD without intensive immune suppression would represent a major advance for HSCT recipients.
Recruitment of lymphocytes into tissues plays a major role in GVHD.4 This process starts early after transplantation, allowing T cells to interact with antigen-presenting cells of host and donor origin.5-7 The organ-specific fate of lymphocytes is determined by a combination of adhesion molecules and chemokines that interact with specific receptors on the lymphocyte.8
Chemokine (C-C motif) receptor 5 (CCR5) and its natural ligands CCL3, CCL4, and CCL59 have been implicated in the pathogenesis of GVHD and solid-organ rejection.10,11 CCR5 is critical for lymphocyte recruitment to tissues involved in GVHD,12-15 and in certain murine models, the migration of CCR5+CD8+ cells into the liver and gut is markedly reduced by the use of an anti-CCR5 antibody, resulting in protection against GVHD.16,17 Although studies in mice that are genetically deficient for Ccr5 have shown conflicting results,15,17-19 genetic evidence in humans shows that certain CCR5 polymorphisms are protective against GVHD20,21 and correlate with survival in patients undergoing allogeneic HSCT.22
Maraviroc (Selzentry, Pfizer and ViiV Healthcare) is the first drug in the class of CCR5 antagonists. It is a noncompetitive, slowly reversible small-molecule antagonist that prevents signaling by all three ligands23 and is used as part of combination antiretroviral therapy for patients infected with CCR5-tropic human immunodeficiency virus (HIV) (i.e., a subtype that uses only the CCR5 co-receptor to enter cells).24,25 The effects of maraviroc on chemotaxis and alloreactivity have not yet been explored. We hypothesized that CCR5 blockade with maraviroc early after allogeneic HSCT might inhibit lymphocyte trafficking and decrease the incidence of acute GVHD.
We determined the effects of maraviroc on the internalization of CCR5, CCR2, and CCR1 on T-cell subsets and on T-cell chemotaxis in normal donors and donors who were homozygous for the CCR5 Δ32 allele (CCR5-Δ32) and thus were CCR5-deficient. We also assessed the effects of maraviroc on functional T-cell assays and hematopoietic colony formation. Detailed methods are described in the Supplementary Appendix, available with the full text of this article at NEJM.org.
We performed a phase 1 and 2 single-group clinical trial to study the role of maraviroc when it was added to conventional GVHD prophylaxis after reduced-intensity conditioned HSCT for patients with hematologic cancers. The phase 1 portion was designed to confirm the safe dose level that was established in patients with HIV and to confirm consistent target drug levels in vivo. This dose was then used in the phase 2 portion of the study. Study participants included adult candidates for HSCT who met protocol eligibility criteria. We calculated the HSCT comorbidity index for each patient, as described previously.26
From June 2009 through March 2011, we enrolled 38 patients. All eligible patients at our institution were offered the chance to participate. Enrollment was held during the analysis of safety and pharmacokinetic data in the phase 1 portion. All patients received a uniform conditioning regimen of intravenous fludarabine (120 mg per square meter of body-surface area) and busulfan (6.4 mg per kilogram of body weight), followed by the infusion of granulocyte colony-stimulating factor–mobilized peripheral-blood stem cells from either a related or an unrelated donor. A single-antigen HLA mismatch was allowed. We determined the amount of CD34+ and CD3+ cells in the graft using standard procedures.27
All patients received standard GVHD prophylaxis with oral tacrolimus (0.06 mg per kilogram per day) in two divided doses starting 2 days before transplantation and intravenous methotrexate (15 mg per square meter on day 1 and 10 mg per square meter on days 3, 6, and 11). The dose of tacrolimus was adjusted to a target trough level of 5 to 15 ng per milliliter. All patients received antimicrobial prophylaxis, which typically included voriconazole and acyclovir starting 2 days before transplantation and trimethoprim–sulfamethoxazole starting at the time of engraftment.
Maraviroc was given orally twice daily starting 2 days before transplantation until day 30. Administration was suspended in cases of toxic effects that investigators considered to be related to the study drug or in cases of severe mucositis. Planned infusions of donor lymphocytes were not allowed. Donor lymphocyte infusion or chemotherapy could be administered to treat relapse or decreased chimerism levels at the discretion of the treating physician.
Primary end points were safety and the cumulative incidence of grade II to IV acute GVHD by day 100. Secondary end points included rates of grade III or IV acute GVHD by day 100 and day 180, organ-specific acute GVHD, moderate-to-severe chronic GVHD, disease relapse, death that was not preceded by disease relapse (non–relapse-related death), and overall survival. We used the Consensus Conference Criteria28 and the National Institutes of Health Consensus Development Project Criteria29 for the grading of acute and chronic GVHD, respectively. We monitored hematopoietic engraftment using molecular chimerism studies (see the Methods section in the Supplementary Appendix). We used the National Cancer Institute Common Terminology Criteria for Adverse Events, version 3.0, to analyze safety.
To determine the inhibitory capacity of patients' serum on CCR5 internalization, we used a modified internalization assay (see the Supplementary Appendix).30 In order to show the ex vivo effect on chemotaxis, we tested the capacity of patient serum to inhibit chemotaxis (see the Supplementary Appendix).
The study was approved by the institutional review board at the University of Pennsylvania, and patients provided written informed consent. Independent oversight was provided by a medical monitor and the clinical trials review committee at the Abramson Cancer Center at the University of Pennsylvania. The study was conducted in adherence to the protocol, available at NEJM.org. Authors' contributions are detailed in the Supplementary Appendix. Pfizer provided the experimental drug, partial research support, and assistance in the interpretation of pharmacokinetic data without a data confidentiality agreement. Pfizer played no additional role in the analysis of the data or in writing or approval of the manuscript. All the authors vouch for the accuracy and completeness of the data and the fidelity of the study to the protocol.
We performed a cumulative incidence analysis to estimate the cumulative risk of acute GVHD, chronic GVHD, relapse, and death not preceded by relapse. We used the Kaplan–Meier method to estimate the probability of survival. Death and donor lymphocyte infusions were considered to be competing risks for GVHD and death not preceded by relapse. Cumulative incidence and survival analyses were conducted in R (R Project for Statistical Computing at www.r-project.org). We used Student's t-test or analysis of variance to perform group comparisons of results on flow cytometry and chemotaxis ratios using Stata software, version 11.1.
Maraviroc inhibited CCL5- or CCL3-driven internalization of CCR5 at concentrations as low as 10 nM (Fig. 1A and 1B), an effect that was seen on effector CD8+ and CD4+ T cells but not on naive (CCR5-negative) CD8+ and CD4+ T cells. In contrast, maraviroc did not block CCL5- or CCL2-driven internalization of the structurally related receptors, CCR1 and CCR2 (Fig. 1C).
We tested the effect of maraviroc on T-cell chemotaxis by measuring the migration of lymphocytes from normal donors and CCR5-Δ32 homozygous donors (Fig. 1D). Although Δ32 cells were unresponsive to CCL5- and CCL3-induced chemo-taxis and were unaffected by maraviroc, the drug caused dose-dependent inhibition of chemotaxis of normal donor lymphocytes.
To determine whether CCR5 blockade affected T-cell function, we performed a series of functional assays. Pharmacologic concentrations of maraviroc (0.5 to 5.0 μM) had no effect on T-cell proliferation and cytokine secretion in response to stimulation by cognate viral peptide (cytomegalovirus). Moreover, specific T-cell cytotoxicity against cytomegalovirus peptide-loaded T2 cells was not impaired (Fig. 1A, 1B, and 1C in the Supplementary Appendix). To investigate the effect of maraviroc on hematopoiesis, fresh peripheral-blood mononuclear cells from normal donors were plated in methylcellulose in the presence of appropriate cytokines; formation of myeloid and erythroid colonies was not significantly affected by the presence of maraviroc (Fig. 1D in the Supplementary Appendix).
We added maraviroc to standard GVHD prophylaxis in 38 patients undergoing reduced-intensity conditioned HSCT to assess safety and the effect on the incidence and severity of acute GVHD. A majority of the patients had high-risk features in terms of age, donor–recipient HLA compatibility, and underlying diseases (Table 1). The median follow-up was 19.6 months (range, 13.8 to 34.8).
We obtained detailed pharmacokinetic profiles for the first 13 patients at two dose levels (300 mg twice daily in 6 patients and 150 mg twice daily in 7 patients) (Fig. 2 and the Methods section in the Supplementary Appendix). Among patients receiving 150 mg twice daily, 3 of 7 patients did not reach a target level of 100 ng per milliliter, whereas among those receiving 300 mg twice daily, all 6 patients reached the target level. In view of the pharmacokinetic data and the lack of significant drug-related toxicity, the 300-mg twice-daily dose was chosen as the phase 2 dose.
Maraviroc had few documented toxic effects. Administration of the drug was briefly suspended in seven patients because of grade 3 abnormalities on liver-function testing (in two patients) or grade 3 or 4 mucositis (in five). Liver-function abnormalities did not recur when the drug was restarted. The adverse-event profile was similar to the expected toxicity observed in patients undergoing reduced-intensity conditioned HSCT without maraviroc (Table 1 in the Supplementary Appendix).
Engraftment in recipients of maraviroc was prompt. One patient died from sepsis on day 12 and thus could not be evaluated. In patients who could be evaluated, the median time to neutrophil engraftment (absolute neutrophil count, >500 per cubic millimeter) was 15 days (range, 10 to 27), and the median time to achieve a platelet count of more than 20,000 per cubic millimeter without transfusions was 19 days (range, 9 to 84). One patient continued to require platelet transfusions because of an early relapse of lymphoma. The median percentages of chimerism in whole-blood and T cells at day 100 were 96% (range, 0 to 100) and 81% (range, 0 to 100), respectively (Fig. 3 in the Supplementary Appendix). Patterns of donor chimerism were similar to those in our past experience with HSCT. Three patients had secondary graft losses on days 56, 149, and 331: two cases that were coincident with progressive myelofibrosis and one in a patient with chronic lymphocytic leukemia (CLL) who had autologous recovery of normal hematopoiesis without evidence of CLL.
Three patients who received low-dose maraviroc did not reach a predetermined pharmacokinetic target and were excluded from the efficacy analysis. Among 35 patients who could be evaluated, the cumulative incidence rate (±SE) of GVHD at day 100 was 14.7±6.2% for grade II to IV acute disease (the primary end point) and 2.9±2.9% for grade III or IV disease (Table 2 and Fig. 2A). In the first 100 days, there were no cases of acute GVHD involving the liver or gut (Table 2 and Fig. 2C). At day 180, the cumulative incidence of grade II to IV acute GVHD was 23.6±7.4%, and organ involvement remained largely confined to the skin, with low rates of GVHD involving the liver (2.9±2.9%) or gut (8.8±5.0%). At day 180, the cumulative incidence of grade III or IV GVHD was only 5.9±4.1%, largely attributable to low incidence rates of gut and liver GVHD, which were absent before day 100 and remained infrequent up to day 180 (Table 2 and Fig. 2C). In the 11 evaluable patients who received an HLA-matched sibling graft, there were no cases of acute GVHD before day 100 and no grade III or IV GVHD by day 180.
The cumulative incidence of moderate-to-severe chronic GVHD was 23.6±7.5% at 1 year (Table 2 and Fig. 2B). In 10 patients with moderate-to-severe chronic GVHD, organ involvement included skin (in 7 patients), mouth (in 5), eyes (in 4), lungs (in 4), liver (in 3), and gut (in 2). Ten donor lymphocyte infusions were administered to 9 patients to treat dropping chimerism levels (in 3 patients) or relapse (in 7 patients) between days 161 and 433. Immunosuppressive therapy or additional maraviroc were not administered to these patients. After infusions of donor lymphocytes, grade IV acute GVHD developed in 2 patients, and severe chronic GVHD developed in 1 patient.
The cumulative incidence of death not preceded by relapse at day 180 was 2.9±2.9%. At 1 year, the cumulative rate of death not preceded by relapse was 11.7±5.6%. Infectious complications in the first 100 days included several cases of bacterial infections at a rate that was expected after reduced-intensity conditioned HSCT (Table 2 in the Supplementary Appendix). There were no viral or fungal infections other than asymptomatic reactivation of cytomegalovirus.
At 1 year, the cumulative incidence of relapse in recipients of maraviroc was 55.9±8.8% which was not higher than expected considering the disease characteristics of the patients and the reduced-intensity conditioning regimen, which is associated with a high risk of relapse. A subset analysis of patients with acute myeloid leukemia (AML) or the myelodysplastic syndromes and lymphoid cancers did not reveal any significant differences in relapse rates (data not shown). The estimated 2-year survival rate was 47.1±8.6%.
To ensure that the antichemotactic effect of maraviroc was preserved in vivo, we tested the capacity of serum from treated patients to inhibit CCR5 internalization and chemotaxis. Serum samples from multiple time points during steady state on day 12 abrogated CCL5-induced internalization of CCR5 on normal donor T cells (Fig. 3A and 3C). As a control, we tested each patient's serum at 60 days (30 days after the last dose of maraviroc) and observed full recovery of CCL5-induced CCR5 internalization at that time point.
Next, we tested the capacity of serum from treated patients obtained on day 0 to inhibit in vitro lymphocyte chemotaxis. By incubating normal donor T cells in patients' serum and then allowing the cells to migrate in a Boyden chamber, we found that CCL5-induced chemotaxis was significantly impaired by serum obtained on day 0 as compared with day 60 (Fig. 3B and 3D).
In this study, we found that maraviroc was biologically active as an inhibitor of lymphocyte chemotaxis both in vitro and in vivo. The addition of maraviroc to standard GVHD prophylaxis resulted in a low incidence of GVHD in high-risk patients with hematologic cancers after allogeneic HSCT. Although cases of skin GVHD were still observed at expected rates, the addition of maraviroc was associated with an absence of liver and gut GVHD through day 100, leading to a low incidence of severe (grade III or IV) GVHD. CCR5 blockade did not disrupt hematopoietic engraftment or lead to a higher-than-expected incidence of relapse or infectious complications.
The outcomes of this study are especially favorable considering the study population, which included older patients (68% over the age of 60 years) and a high proportion of matched unrelated donors (50%) and HLA-mismatched donors (16%); the anticipated incidence of acute GVHD in similar patients is typically more than 50%.31-33 The patients in our study also had major coexisting illnesses, with an intermediate or high co-morbidity index26 in almost half the patients. At our institution, by day 180 among patients receiving tacrolimus and methotrexate for prophylaxis after reduced-intensity conditioned HSCT, the rates of acute GVHD are 38.5% for grade II to IV disease and 21.9% for grade III or IV disease. During the same time period in our study, the use of a combination of maraviroc, tacrolimus, and methotrexate resulted in cumulative incidence rates of 23.6% for grade II to IV disease and 5.9% for grade III or IV disease. Results of this study also compare favorably with other studies of reduced-intensity conditioned HSCT that used similar conditioning regimens.34-41 For example, in a recent report from the Center for International Blood and Marrow Transplant Research involving 1174 patients who underwent reduced-intensity conditioned HSCT with a peripheral-blood stem-cell graft, by day 100 moderate-to-severe acute GVHD occurred in 35.0% of patients receiving matched related-donor transplants and 57.0% of patients receiving unrelated-donor transplants.41 In a similar study, which focused on patients with AML or the myelodysplastic syndromes, grade II to IV acute GVHD occurred in 46.0% of patients undergoing reduced-intensity conditioned HSCT.42 The inclusion of acute GVHD only through day 100 in these large studies probably led to an underestimation of the true incidence of acute GVHD, which is frequently delayed after reduced-intensity conditioned HSCT.37 Of course, the value of maraviroc in lowering the rate of acute GVHD will need to be assessed in a prospective, randomized trial.
The prevention of primarily visceral GVHD is consistent with murine models in which a blocking CCR5 antibody ameliorated GVHD of the liver and gut.16,17 From a clinical standpoint, liver and gut involvement is reported in more than 50% of patients with GVHD43 and frequently determines severity, which itself predicts response to therapy and survival.44 Our own experience with reduced-intensity conditioned HSCT involving similar patients receiving the same conditioning regimen and tacrolimus plus methotrexate as GVHD prophylaxis has shown that among patients with acute GVHD, 44.0% had liver involvement and 39.0% had gut involvement (unpublished data). Although it is unclear why CCR5 blockade would limit visceral GVHD specifically, murine models suggest that tissue distribution may be related to the intensity of responses in type 1, type 2, and type 17 helper T cells.45-47 It is possible that organspecific protection is related to the preferential expression of CCR5 on type 1 helper T cells.48 Other chemokine-signaling mechanisms may be involved in skin infiltration of lymphocytes.
The administration of maraviroc only until day 30 after transplantation resulted in low rates of GVHD up to 6 months after transplantation. Data from murine models suggest that interactions between T cells and antigen-presenting cells that set the stage for GVHD occur early after HSCT.6 It is therefore possible that early and brief CCR5 blockade is sufficient to reduce subsequent GVHD. Conversely, the emerging cases of acute and chronic GVHD after the first 6 months suggest that a longer treatment course of maraviroc may have a beneficial effect.
In summary, a 33-day course of maraviroc, an inhibitor of T-cell chemotaxis, appeared to reduce the incidence of acute visceral GVHD without major adverse events.
Supported by Pfizer; a Specialized Center of Research grant (to Drs. Porter and Vonderheide) and a Special Fellow in Clinical Research Award (to Dr. Reshef) from the Leukemia and Lym-phoma Society; grants (P30-CA16520, to Drs. Vonderheide, Stadtmauer, and Heitjan and Ms. Mick; K24-CA117879, to Dr. Porter; and U01-HL069286, to Drs. Porter and Stadtmauer) from the National Institutes of Health; a Pilot Project Grant from the Abramson Cancer Center (to Drs. Reshef, Vonderheide, and Porter); and an American Society of Hematology Scholar Award and an American Society of Clinical Oncology Young Investigator Award (both to Dr. Reshef). Dr. Reshef is a fellow of the Institute for Translational Medicine and Therapeutics, University of Pennsylvania.
We thank Marcelo Pasquini of the Center for International Blood and Marrow Transplant Research and Hernan Valdez and Manoli Vourvahis of Pfizer for their helpful discussions; Oren Litvin of Columbia University for help with preparation of the figures; Grace Jeschke for help with colony-formation assays; Amy Marshall and Kristin Coffan for regulatory support; and Pavel Vassilev for data management.
Presented in part at the 2010 American Society of Hematology Annual Meeting, Orlando, FL, December 4–7, 2010; the 2011 American Society of Hematology Annual Meeting, San Diego, CA, December 10–13, 2011; and the 2011 Blood and Marrow Transplantation Tandem Meeting, Honolulu, February 17–21, 2011
Dr. Frey reports receiving lecture fees from Pfizer; Dr. Stadt-mauer, consulting fees from Pfizer; Dr. Vonderheide, grant support from Pfizer; and Dr. Porter, consulting fees from Bristol-Myers Squibb and lecture fees from Celgene and Millennium. No other potential conflict of interest relevant to this article was reported.
Disclosure forms provided by the authors are available with the full text of this article at NEJM.org
Ran Reshef, The Abramson Cancer Center and the Division of Hematology and Oncology, the Abramson Family Cancer Research Institute, Perelman School of Medicine, University of Pennsylvania, Philadelphia.
Selina M. Luger, The Abramson Cancer Center and the Division of Hematology and Oncology, Perelman School of Medicine, University of Pennsylvania, Philadelphia.
Elizabeth O. Hexner, The Abramson Cancer Center and the Division of Hematology and Oncology, Perelman School of Medicine, University of Pennsylvania, Philadelphia.
Alison W. Loren, The Abramson Cancer Center and the Division of Hematology and Oncology, Perelman School of Medicine, University of Pennsylvania, Philadelphia.
Noelle V. Frey, The Abramson Cancer Center and the Division of Hematology and Oncology, Perelman School of Medicine, University of Pennsylvania, Philadelphia.
Sunita D. Nasta, The Abramson Cancer Center and the Division of Hematology and Oncology, Perelman School of Medicine, University of Pennsylvania, Philadelphia.
Steven C. Goldstein, The Abramson Cancer Center and the Division of Hematology and Oncology, Perelman School of Medicine, University of Pennsylvania, Philadelphia.
Edward A. Stadtmauer, The Abramson Cancer Center and the Division of Hematology and Oncology, Perelman School of Medicine, University of Pennsylvania, Philadelphia.
Jacqueline Smith, The Abramson Cancer Center and the Division of Hematology and Oncology, Perelman School of Medicine, University of Pennsylvania, Philadelphia.
Sarah Bailey, The Abramson Cancer Center and the Division of Hematology and Oncology, Perelman School of Medicine, University of Pennsylvania, Philadelphia.
Rosemarie Mick, The Department of Biostatistics and Epidemiology, Perelman School of Medicine, University of Pennsylvania, Philadelphia.
Daniel F. Heitjan, The Department of Biostatistics and Epidemiology, Perelman School of Medicine, University of Pennsylvania, Philadelphia.
Stephen G. Emerson, The Abramson Cancer Center and the Division of Hematology and Oncology, Perelman School of Medicine, University of Pennsylvania, Philadelphia.
James A. Hoxie, The Abramson Cancer Center and the Division of Hematology and Oncology, the Center for AIDS Research, Perelman School of Medicine, University of Pennsylvania, Philadelphia.
Robert H. Vonderheide, The Abramson Cancer Center and the Division of Hematology and Oncology, the Abramson Family Cancer Research Institute, Perelman School of Medicine, University of Pennsylvania, Philadelphia.
David Porter, The Abramson Cancer Center and the Division of Hematology and Oncology, the Abramson Family Cancer Research Institute, Perelman School of Medicine, University of Pennsylvania, Philadelphia.