1. ¹Department of Gastroenterology and Hematology, Hokkaido University Graduate School of Medicine, Sapporo, Japan
2. ²Department of Hematology and Oncology, Hokkaido University Graduate School of Medicine, Sapporo, Japan

Allogeneic bone marrow transplantation (BMT) is one of the curative treatments for young patients with severe aplastic anemia (SAA).¹ However, graft failure still remains a significant clinical problem and is often treated by a second allogeneic stem cell transplantation.² In this paper, we describe a patient with late graft failure after allogeneic BMT for SAA, who received donor leukocyte infusion (DLI) from the original donor and achieved hematological recovery.

A 21-year-old female with SAA was admitted to our hospital in July 2000 to undergo allogeneic BMT from an HLA-matched and ABO-matched sibling donor. Paroxysmal nocturnal hemoglobinuria was ruled out by negative results both for a Ham test and a sugar water test, and by the fact that almost 80% of the patient's neutrophils were CD55- and CD59-positive, as assessed by flow cytometry. Since she also suffered from chronic hepatitis B and the donor was an HBV carrier, lamivudine was administered to the recipient. The patient was conditioned with cyclophosphamide (50 mg/kg/day on days -5 to -2), antithymocyte globulin (2.5 mg/kg/day on days -6 to -2) and total lymphoid irradiation (7.5 Gy on day -1), and received 4.1 10⁸ nucleated bone marrow cells/kg. The reason for conditioning with radiation, which can induce secondary malignancy, was to prevent graft failure in the patient, who was treated with multiple blood transfusions. GVHD prophylaxis consisted of treatment with cyclosporine A (CsA) and a short course of methotrexate. CsA was later replaced by tacrolimus, because of the complication of CsA-induced encephalopathy. Chimerism analysis was carried out using a capillary electrophoresis system, as previously described.³ After successful BMT without serious hepatic dysfunction, the patient was followed up monthly and showed mixed chimerism without obvious acute or chronic GVHD, allowing discontinuation of tacrolimus treatment 5 months after the BMT (Figure 1). At 7 months after the BMT, pancytopenia with a stable mixed chimerism appeared and gradually progressed. Bone marrow examination revealed severe hypocellularity with no evidence of cytomegalovirus, HHV-6, or human parvovirus B19, and bone marrow scintigraphy using ¹¹¹In showed normal uptake in the central and peripheral bone marrow. The patient also suffered from chronic active hepatitis B (peak AST, ALT and HBV-DNA levels being 371, 788 IU/l and 7.1 LGE/ml, respectively). A second transplantation was therefore considered to be contraindicated because of a high risk of life-threatening hepatitis. Since early cessation of tacrolimus treatment might have enabled the survival of residual host lymphocytes, resulting in induction of durable mixed chimerism, we expected that DLI would be effective for complete elimination of host-derived immune cells, which could be responsible for pancytopenia. The dose of infused leukocytes from the initial donor was gradually increased without any prophylaxis of GVHD, from 1 10⁷/kg CD3 cells to 2 10⁷, 4 10⁷ and 7.5 10⁷/kg cells. Four courses of DLI were carried out without induction of serious hepatic dysfunction or GVHD. Hematopoietic aplasia was not observed because of much donor hematopoiesis before DLI. At 2 months after the last DLI, weekly subcutaneous administration of erythropoietin was started to accelerate erythropoiesis. The pancytopenia gradually improved over the next 4 months with conversion to complete donor-type chimerism (Figure 1). The patient is now well, has required no transfusions, and has shown no clinical signs of GVHD or hepatic dysfunction. Seroconversion of HBeAb-positive immune status was also obtained after three courses of DLI with administration of lamivudine.

Figure 1.

Clinical course of the patient after BMT and DLI. Pancytopenia gradually improved with conversion to complete donor-type chimerism after four courses of DLI. WBC, white blood count; Hb, hemoglobin; Plt, platelet count; G-CSF, granulocyte colony-stimulating factor; PC, platelet concentrate; MAP, mannitol-adenine-phosphate-added concentrated red cells; EPO, erythropoietin; NA, not available.

Full figure and legend (158K)

The fact that hematopoietic recovery was achieved after reconstitution of complete donor-type chimerism means that DLI was effective for the eradication of host-derived immune cells. However, the possibility that cells transfused for DLI contained hematopoietic stem cells that could reconstitute hematopoiesis could not be completely ruled out. The discrepancy between the percentage of donor cells and the pancytopenia as well as the response to DLI remains unclear. Recurrence of pancytopenia in the presence of stable mixed chimerism with the predominance of donor T cells has already been reported by Casado et al, who found late graft failure after durable mixed chimerism in patients conditioned only with cyclophosphamide.⁴ We are not sure if late graft failure is usually preceded by a slow and progressive loss of donor hematopoiesis. One possible factor causing the pancytopenia in our patient is the cessation of tacrolimus treatment 5 months after the BMT, which might have activated residual host lymphocytes even when there was a majority of donor lymphocytes.

As already suggested, the addition of irradiation increases the rate of complete chimerism; however, sensitive techniques such as PCR enable the detection of residual recipient cells even after conditioning involving irradiation.^5,6 Although it is not clear whether graft failure is always preceded by a progressive loss of donor-type hematopoiesis, DLI may become a useful treatment for some patients with mixed chimerism.⁷ Since DLI unexpectedly induces fatal complications such as GVHD and pancytopenia, which occur regardless of the chimerism status, the use of DLI for aplastic anemia patients with graft failure in combination with mixed chimerism may be limited to patients who showed no obvious acute or chronic GVHD. Coadministration with stem cells in G-CSF-mobilized PBSC might be helpful for the prevention of pancytopenia after DLI in patients who have shown little donor hematopoiesis before DLI. In our case, we decided to perform DLI because of the complication with hepatitis B, preventing profound immunosuppression. Treatment criteria that would qualify a patient with mixed chimerism for DLI should be discussed with respect to other therapeutic options. Further studies are required to assess the benefits and risks, particularly with regard to GVHD and long-term sustained hematopoiesis after DLI.