Successful treatment of EBV-associated posttransplantation lymphoma after cord blood transplantation using third-party EBV-specific cytotoxic T lymphocytes

Cord blood transplantation (CBT) has been associated with a 2% Epstein-Barr virus–posttransplantation lymphoproliferative diseases (EBV-PTLD) risk,¹  although an incidence of 21% has been documented when antithymocyte globulin (ATG) is added to nonmyeloablative conditioning.²  While EBV-PTLD risk can be reduced by avoidance of ATG or monitoring and preemptive Rituximab,^3,4  once EBV-PTLD has developed, therapy after CBT is limited to rituximab ± chemotherapy. Herein, we describe 2 cases of successful treatment of CB-derived EBV-PTLD with partially human leukocyte antigen (HLA)–matched third-party EBV-specific cytotoxic T lymphocytes (EBV-CTLs).

Patient #1 (Table 1) achieved 100% chimerism with 1 of 2 CB units by day 21. He developed grade 2 acute graft-versus-host disease (GVHD). This resolved with 2 mg/kg of intravenous methylprednisolone, although corticosteroid taper was slowed due to intermittent flares. After routine surveillance for EBV-DNA by quantitative polymerase chain reaction (qPCR) was negative for 6 months, it was discontinued. However, exacerbation of gut symptoms at 8 months prompted endoscopy. This revealed extensive intestinal ulceration with a donor-derived EBV⁺ diffuse large B-cell lymphoma. Simultaneously, blood EBV qPCR was > 8300 copies/mL 18-flourodeoxyglucose positron emission tomography (FDG-PET) demonstrated lymphoma in cervical and mesenteric lymph nodes and throughout the intestine (Figure 1A). Immunosuppression was stopped and rituximab (375 mg/m²/dose × 10) was administered without response. Disease progression with new lung lesions and extensive tonsillar and oropharyngeal ulceration prompted emergent treatment with third-party EBV-CTLs.

Patient #2 similarly developed grade 2 gut GVHD that responded to tacrolimus and prednisone. However, 8 months after transplantation, diarrhea and abdominal pain prompted colonic biopsy that revealed extensive infiltration with donor-derived EBV⁺ diffuse large B-cell lymphoma. FDG-PET revealed widely disseminated disease. Despite Rituxan (375 mg/m²/dose ×4), hydroxyurea, prednisone taper to physiologic levels, and cessation of tacrolimus, EBV-PTLD progressed with left hemiparesis and subsequent coma. Magnetic resonance imaging revealed new space-occupying lesions in the basal ganglia and cerebellum, and EBV-DNA was detected in spinal fluid. She was treated with whole-brain radiotherapy (2450 centi-Gray), dexamethasone, and 4 additional rituximab doses. All neurologic symptoms resolved except left-arm paresis, viremia cleared, and abdominal disease improved on FDG-PET. However, within 2 months FDG-PET scan showed progression in the colon. Hence, the patient was referred for third-party EBV-specific CTL therapy.

We have generated more than 330 fully characterized EBV-CTLs from consenting healthy donors under good manufacturing practice conditions for adoptive therapy of high-risk allograft recipients. EBV-CTLs were generated as previously described.^5,6  Briefly, monocyte/natural killer cell–depleted mononuclear cells were stimulated in vitro at a 20:1 ratio with irradiated autologous EBV-transformed B-cell lines (EBV-BLCLs). After 11 days, T cells were restimulated weekly with irradiated EBV-BLCLs at a 4:1 ratio with interleukin-2. After at least 28 days, T cells were tested forEBV-specificity and their HLA-restriction established by examining their cytotoxic activity in ⁵Cr release assays against a panel of EBV⁺ and EBV⁻ targets, each expressing one of the set of HLA A, B, C, DR and DQ HLA-alleles shared by the T-cell donor as previously described.⁷  T-cell lines lacked alloreactivity and were sterile and endotoxin-free. CTLs with the closest HLA-match to the CB donor (and hence the lymphoma) that were restricted by 1 or more of that donor's HLA-alleles were selected for administration. Memorial Sloan-Kettering Cancer Center Institutional Review Board–approved informed consent was obtained before EBV-CTL administration in accordance with the Declaration of Helsinki. The donors from whom the lines were generated also consented to the use of their cells in a patient other than for who they had donated.

Figure 1A-B shows the EBV-CTL administration time-line, EBV-specific CTL precursor (CTLp) blood levels, and the EBV qPCR and FDG-PET disease course for each patient. Patient #1 received EBV-CTLs (1 × 10⁶T cells/kg/dose ×5) every 21 days ×3, and then at 86 and 108 days. They were homozygous for class I and II HLA-alleles and restricted by HLA-B*0801 and HLA-C*0701 shared by the lymphoma. By day 10 after each infusion, a 2-3 log₁₀ increase in EBV-CTLp frequency was detected that persisted for 7-10 days. By 48 days after the third infusion, and 7 days after infusion 4, EBV-CTLs could not be detected in the blood by PCR analysis. Clinical improvement was documented by resolution of fever, abdominal pain, and melena over the first 18 days after the first infusion. Reduction in PET-avid lesions was documented 14 days after the initial infusion, with marked reductions of all lesions by day 24. Imaging 107 days after CTL initiation revealed near resolution, but colonic lesions remained. Colonic biopsy demonstrated persistent EBV-PTLD. Repeat imaging 176 days after EBV-CTL demonstrated complete remission that is maintained 20 months after EBV-PTLD diagnosis.

Patient #2 initially received 2 courses of 3 weekly infusions of EBV-CTLs (1 × 10⁶ T cells/kg/dose ×6). These were matched to the lymphoma at HLA-A*0301 and B*3501 but differed for alleles of HLA-DRB1*13 (CB donor/lymphoma DRB1*1302 and EBV-CTL DRB1*1301). The EBV-CTLs included cytotoxic CD8⁺ T cells restricted by HLA-B*3501 and CD4⁺ T cells restricted by HLA-DRB1*1301. This resulted in partial remission after course 1 (day +38), and complete remission after course 2 (day +106). However, because of persistent left-arm paresis, 3 additional infusions were given at the same dose. Disease response was correlated with a 4-fold EBV-CTLp increase after the first, and a 10-fold increase after the second EBV-CTL course. The persistence of the CTLp frequencies likely reflects the shorter intervals between infusions. We also analyzed the number of tetramer-binding HLA-B*3501–restricted EBV-specific T cells before and after CTL administration. CD8⁺ EBV-CTLs restricted by HLA-B3501 were a dominant population in the CTLs. They were not detected before treatment, but dramatically increased by 16 days after the first CTL infusion. The contribution of HLA-DRB1*13–restricted T cells is unclear. We did not have tetramers to measure HLA-DRB1*1301–restricted T cells after infusion. Furthermore, their activity against targets expressing the HLA-DRB1*1302 of the lymphoma could not be ascertained. Subsequently, the patient improved with complete remission maintained at 15 months after EBV-PTLD diagnosis. In both patients infusions were without toxicity; neither patient had GVHD flare despite cessation of GVHD therapy and CTL administration.

Rituximab monotherapy in EBV-PTLD has yielded a response rate of only 44%,⁸  and can be limited by lack of CD20 expression.⁹  However, chemotherapy (± rituximab) after allograft can be associated with significant toxicity and mortality.⁸  An alternative therapy is the administration of donor-derived EBV-CTLs.⁹  However, the requirement for a seropositive donor and the time needed for CTL production has limited application of this approach. Moreover, until recently generation of donor-derived virus-specific CTLs from naive CB T cells was not thought possible. Hanley et al¹⁰  have now confirmed 2 earlier studies^11,12  indicating that cytomegalovirus and EBV-specific T cells can be generated from CB. However, the EBV-specific T cells generated from naive CB-derived T cells are specific for epitopes not targeted by adult memory T cells. While these T cells can lyse autologous EBV-BLCL in vitro, it remains to be determined whether they have the requisite avidity and resistance to apoptosis to be effective in vivo. Furthermore, production time remains lengthy (30 days for EBV-BLCL sensitization plus 21 days for CTL production).

In contrast to transplant donor–derived T cells, cryopreserved third-party EBV-specific T cells are immediately available and can be selected on the basis of their partial HLA-restriction and their level of HLA-match. Haque et al were the first to demonstrate that partially HLA-matched, third-party, EBV-specific T cells can induce remissions of EBV-PTLD in organ allograft recipients and 2 marrow allograft recipients.^13,14  However, they did not establish the clonality of the EBV-PTLD. In our patients, the lymphomas were monoclonal and had failed immunosuppression taper and Rituximab. Historically, such lymphomas are usually lethal. In our patients, sequential third-party T-cell infusions induced transient increments in circulating EBV-specific T cells, clearance of EBV-DNA, and clinical and radiologic remissions. In contrast to the extended survival that has been reported for donor-derived EBV-specific T cells,¹⁵  the third-party T cells did not durably engraft. Thus, multiple infusions were required. Nevertheless, the disease regressions achieved have been sustained, likely reflecting subsequent establishment of effective transplant donor-derived EBV-specific T cells. Our results provide evidence that “off-the-shelf” third-party CTLs selected on the basis of partial HLA-matching and HLA-restriction can be effective treatment of monoclonal EBV-PTLD after CBT. While transplant donor–derived T cells may constitute the only option for adoptive cell therapy for a proportion of CBT recipients who inherit very rare HLA genotypes, a review of the HLA typing of marrow and CB allograft recipients transplanted at our center also suggests that our diverse bank of fully characterized EBV-CTLs derived from consenting donors and manufactured under good manufacturing practice conditions could provide effective treatment for a large proportion of EBV-PTLD patients. This treatment approach could also be applied to other infections, and, potentially, to the eradication of residual or recurrent malignancy.

This work was supported in part by the P01 CA23766 and CA59350 from the National Institutes of Health, The Tow Foundation, The Aubrey Fund, The Smead Foundation, and The Laura Rosenberg Foundation.

National Institutes of Health

Contribution: J.N.B., G.K., and R.J.O. performed the research, analyzed and interpreted the data, and wrote the manuscript; E.D. generated and characterized the EBV-CTLs and the immunologic responses and assisted in writing the manuscript; C.S. assisted in writing the manuscript; J.J.J., M.A.P., and S.E.P. assisted in clinical management; and M.D. assisted in the analysis of imaging studies and in writing the manuscript.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Juliet N. Barker, MBBS (Hons), Department of Medicine, Memorial Sloan-Kettering Cancer Center, 1275 York Ave, New York, NY 10065; e-mail: barkerj@mskcc.org.