Stem cell transplantation with reduced-intensity conditioning for hemophagocytic lymphohistiocytosis

Primary hemophagocytic lymphohistiocytosis (HLH) is a disease of childhood that presents with fever, hepatosplenomegaly, and pancytopenia.^1,2  It is characterized by a selective deficiency of cellular cytotoxicity with hyperactivation of T cells and macrophages.^3-5  A number of genetic abnormalities, including perforin and hMUNC13-4 defects in 20% to 40% of patients and mutations in SH2D1A/SAP have been described,^6-11  suggesting that a deficiency of triggering apoptosis may be responsible for the pathologic features.

Patients may achieve remission with immunomodulation consisting of epipodophyllin and steroids with or without cyclosporin A.^10,12  However, the only curative option is allogeneic stem cell transplantation (SCT).^13,14  Patients often have significant pretransplant morbidity and require intensive cardiorespiratory support before SCT. This toxicity results in a high transplant-related mortality (TRM) with conventional intensity conditioning, mostly from noninfectious pulmonary toxicity and veno-occlusive disease (VOD). In the HLH-94 study the TRM was 30%,¹⁵  with a 3-year overall survival rate of 71% following matched related donor transplantation, 70% for matched unrelated, 54% for mismatched unrelated donor, and 50% for those with a haploidentical donor.

We hypothesized that SCT with reduced-intensity conditioning (RIC) might be sufficient to restore normal immune regulation and hence cure HLH, while decreasing TRM. In support of this, the HLH-94 registry reports that 8 of 43 evaluable patients receiving transplants have mixed chimerism, and all remain disease-free at a median of 3 years after bone marrow transplantation (BMT), demonstrating that mixed chimerism is sufficient for long-term cure. Our experience in children with primary immunodeficiencies suggests that RIC may decrease the TRM from SCT.¹⁶  We therefore investigated if this approach could be used in children with HLH.

Twelve consecutive patients with primary HLH at Great Ormond Street Hospital underwent SCT from an unrelated (fully matched unrelated donor [MUD], n = 5, one antigen-mismatched unrelated donor [MMUD], n = 3) or phenotypically matched family (n = 1) or haploidentical (n = 3) donor using RIC between June 1999 and November 2004. Three patients had underlying primary immunodeficiencies and one child had Down syndrome. Although patients were prospectively enrolled, there were no formal entry criteria. Patients were chosen based on the clinician's judgment that TRM was likely to be unacceptably high with conventional conditioning because of significant coexisting morbidity. Patient characteristics, including pre-SCT organ dysfunction are described in Table 1. Three patients presented with multiorgan failure requiring ventilation, 3 had central nervous system (CNS) involvement, 2 had hypertrophic cardiomyopathy, 4 had liver disease, and one had renal infiltration with HLH. Four of 9 patients assessed had a detectable perforin mutation. One further patient had no detectable mutation, but absent perforin expression and an additional patient had absent SAP expression, with no detectable mutation. Details of perforin mutations are included in Table 1.

All patients had previously received the HLH-94 protocol. Eight of 12 were in complete remission (CR) and 4 in partial remission (PR) before SCT. The median age at transplantation was 14 months (range, 4-151 months). SCT was performed at a median of 6.5 months from diagnosis (range, 3-14 months).

Approval was obtained from the Great Ormond Street Hospital Ethics Committee for the use of the RIC regimen. Informed consent was provided according to the Declaration of Helsinki from all parents or next of kin.

Eleven patients were conditioned with fludarabine 30 mg/m² for 5 days (days -7 to -3) and melphalan 140 mg/m² (day -2) in the patients receiving matched family/unrelated donor (MFD/MUD) transplants or 125 mg/m² (day -1) in the patients receiving haploidentical grafts. Patients receiving MFD/MUD transplants received serotherapy with alemtuzumab 0.2 mg/kg for 5 days (day -8 to -4) and those receiving haploidentical transplants received antithymocyte globulin (ATG) 5 mg/kg for 5 days (day -5 to -1) together with busulphan 4 mg/kg for 2 days (days -9 to -8) as additional myelosuppression. One patient, who was ventilated at the time of transplantation, received fludarabine 30 mg/m² for 3 days (day -4 to -2) and 2 Gy total body irradiation (TBI) in a single fraction. Grafts were T-replete marrow (for MFD/MUD) or peripheral blood stem cells (PBSCs for MMUD) and granulocyte colony-stimulating factor (G-CSF)-mobilized CD34 selected using the CliniMACs system (Miltenyi Biotec, Bisley, United Kingdom) for the haploidentical donors.

The median number of bone marrow mononuclear cells infused was 4.1 × 10⁸/kg (range, 1.7-7.5 × 10⁸/kg) in unrelated donors and the median of CD34⁺ selected cells infused was 15.5 × 10⁶/kg (range, 13.8-32.0 × 10⁶/kg) in the haploidentical transplants. Graft-versus-host disease (GVHD) and antimicrobial prophylaxis were as described.¹⁶ 

Engraftment was assayed on the peripheral blood using variable numbers of tandem repeats (VNTRs) analysis with multiple informative alleles. Lineage-specific chimerism analysis was performed after selection of CD3/CD15/CD56/CD19⁺ cells using the magnetically activated cell sorting (MACS) system (Miltenyi Biotech, Surrey, United Kingdom).

At a median of 14 days (range, 9-26 days), all patients had neutrophil engraftment and at a median of 16 days they had an unsupported platelet count more than 20 × 10⁹/L (range, 4-34 × 10⁹/L). All patients had 100% donor cells at engraftment. Three of the 9 survivors subsequently developed mixed chimerism. Lineage-specific chimerism was analyzed in mixed chimeras and is shown in Table 1. Chimerism was stable after 6 months following SCT. No patient has rejected the graft or relapsed.

Three patients had severe mucositis; none developed VOD. Two patients developed pneumonitis. Six patients had cytomegalovirus (CMV) reactivation and 3 reactivated Epstein-Barr virus (EBV). Nine of 12 (75%) children are alive and in CR a median of 26 months from transplantation (range, 5-69 months). One patient each had TRM from CMV pneumonitis, multifactorial pneumonitis following T-cell sequestration and CMV disease on the background of previous pulmonary HLH, and hepatic rupture after transjugular liver biopsy.

Four patients developed acute GVHD; 3 had grade 2 (skin and gut) and one had grade 4 (skin). Two patients have developed limited skin chronic GVHD and one has low-risk extensive chronic GVHD of the skin and gut.

Eight patients are alive with more than 6 months of immunology follow-up. All have achieved normal T-cell function, as assessed by the phytohemagglutinin (PHA) stimulation index, a median of 7.5 months (range, 4-13 months) from SCT. Seven have normal age-adjusted CD3 counts a median of 8 months (range, 5-16 months) from SCT, and 8 have normal CD4 counts (median, 8 months; range, 6-16 months) from SCT. Two patients have normal B cells 8 and 13 months from SCT. Six have normal IgM levels. Patient 1 with X-linked lymphoproliferative (XLP) disease had decreased natural killer (NK) T cells (CD3⁺CD56⁺) before transplantation, which increased to normal levels after transplantation (2.9 × 10⁵/L to 2.08 × 10⁶/L), and remains in remission.

Nine of 12 patients are alive and in CR with Lansky scores of 90% to 100% a median of 30 months from SCT (Table 2).

Our results with RIC SCT in patients with HLH and significant pre-SCT morbidity are encouraging. Although too small for statistical analysis, the overall survival data compare favorably with historical data, particularly for patients receiving mismatched transplants. In our cohort, 4 of 5 patients receiving cells from a MUD and 5 of 6 from an HLA-mismatched donor survive in CR, compared with corresponding figures of 70% and 52% in the HLH-94 study.¹⁵  Viral reactivations were frequent, as has previously been reported with the immunosuppressive RIC,¹⁶  but only one patient died of directly infectious causes.

Three surviving patients remain in long-term remission despite having mixed chimerism. All had high-level donor T-cell engraftment. Patient no. 10 is especially intriguing, in that he remains disease-free 6 years after SCT despite having only donor T-cell and not NK cell engraftment. These data suggest the presence of donor T cells is sufficient to restore normal immunoregulation and prevent hypercytokinemia and macrophage activation and that myeloablation is not critical for cure of HLH. We have not performed cellular cytotoxicity assays after SCT because the majority of our patients were 100% donor engrafted. In future studies, it would be of interest to assess T-cell and NK cell cytotoxicity in patients with mixed chimerism.

Although it is predicted that the incidence of end-organ damage, infertility, and growth retardation may be lower than with conventional intensity conditioning, demonstrating this will require longer-term follow-up. One particular concern is the potential for etoposide-associated therapy-related acute myeloid leukemia (t-AML) in patients with residual recipient myelopoiesis. However, data from patients not receiving transplants with EBV-associated HLH¹⁷  have shown that this risk is very low and we believe it is likely to be even lower in patients undergoing SCT, who receive a much lower cumulative dose of etoposide.

Given these results, we would recommend RIC SCT in patients with severe organ toxicity before transplantation. Larger studies are indicated for standard-risk patients undergoing unrelated donor SCT. We have previously noted a higher incidence of rejection in patients with immunodeficiency receiving transplants from a matched sibling donor using the alemtuzumab/fludarabine/melphalan protocol, reflecting the reduced donor-versus-host allo-reactivity. Given the risk of rejection (which may predispose to relapse¹⁸ ) in such patients and in those given a transplant from a haploidentical donor, we would recommend conventional intensity conditioning in these patients at present, unless precluded by significant pre-SCT morbidity.

We would like to thank Jan-Inge Henter for sharing outcome of transplant data from the HLH-94 study before its publication in the British Journal of Haematology and for his helpful comments on the manuscript.