HSV-TK Gene Transfer into Donor Lymphocytes for Control of Allogeneic Graft-Versus-Leukemia

See allHide authors and affiliations

Science  13 Jun 1997:
Vol. 276, Issue 5319, pp. 1719-1724
DOI: 10.1126/science.276.5319.1719


In allogeneic bone marrow transplantation (allo-BMT), donor lymphocytes play a central therapeutic role in both graft-versus-leukemia (GvL) and immune reconstitution. However, their use is limited by the risk of severe graft-versus-host disease (GvHD). Eight patients who relapsed or developed Epstein-Barr virus–induced lymphoma after T cell–depleted BMT were then treated with donor lymphocytes transduced with the herpes simplex virus thymidine kinase (HSV-TK) suicide gene. The transduced lymphocytes survived for up to 12 months, resulting in antitumor activity in five patients. Three patients developed GvHD, which could be effectively controlled by ganciclovir-induced elimination of the transduced cells. These data show that genetic manipulation of donor lymphocytes may increase the efficacy and safety of allo-BMT and expand its application to a larger number of patients.

Allogeneic bone marrow transplantation is the treatment of choice for many hematologic malignancies (1, 2). It is now recognized that the “allogeneic immune advantage,” in addition to the effectiveness of high-dose chemoradiotherapy, is responsible for the curative potential of allo-BMT (1, 2). Although the nature of effector cells has not yet been fully elucidated, transplantation of allogeneic bone marrow produces superior results compared to autologous or syngeneic transplants (3). However, the therapeutic impact of the allogeneic advantage is limited by the risk of a potentially life-threatening complication, GvHD. Severe GvHD can be circumvented by the removal of T lymphocytes from the graft (2). However, T cell depletion increases the incidence of disease relapse, graft rejection, and reactivation of endogenous viral infections (4). Thus, the delayed administration of donor lymphocytes has recently been used for treating leukemic relapse after allo-BMT. Patients affected by recurrence of chronic myelogenous leukemia, acute leukemia, lymphoma, and multiple myeloma after BMT could achieve complete remission after the infusion of donor leukocytes, without requiring cytoreductive chemotherapy or radiotherapy (5). Other complications related to the severe immunosuppressive status of transplanted patients, such as Epstein-Barr virus–induced B lymphoproliferative disorders (EBV-BLPD) (6, 7) or reactivation of CMV infection (8), also benefited. However, severe GvHD represents a frequent and potentially lethal complication of this delayed infusion of donor T cells (9). No specific treatment exists for established GvHD. We investigated the genetic manipulation of donor lymphocytes, which could enable their selective elimination and abrogation of GvHD, thereby making marrow transplantation more efficacious, safer, and available to a larger number of patients.

(HSV-TK) has been successfully transferred into various cell lines to confer ganciclovir sensitivity, and its efficacy has been demonstrated both in vitro and in vivo (10-12). However, an absolute prerequisite for the efficacy of this strategy is that all infused donor lymphocytes carry the “suicide” gene. For this purpose, we devised a simple strategy on the basis of retroviral vector–mediated gene transfer and expression in transduced cells of a cell surface marker not expressed on human lymphocytes, followed by positive immunoselection of the transduced cells (13). This strategy ensures that virtually 100% of the peripheral blood lymphocytes (PBLs) carry the suicide gene (13).

We used a retroviral vector (SFCMM-2) (Fig.1A) (13, 14) for the transfer and expression into human PBLs of two genes: The first encodes a truncated (nonfunctional) form of the human low-affinity receptor for nerve growth factor (ΔLNGFR) (14); the second encodes the HSV-TK–NEO fusion protein (Fig. 1A). ΔLNGFR is located on the cell surface and allows rapid in vitro selection of transduced cells by the use of magnetic immunobeads. In addition, a surface marker allows easy ex vivo detection and characterization of the transduced cells by fluorescence-activated cell sorting (FACS) analysis (13). The safety and efficacy of this vector were extensively tested in vitro and in vivo in small-animal models (15). In this pilot clinical study, the proportion of transduced donor lymphocytes after one round of gene transfer was in the 20 to 50% range. The proportion of transduced donor cells after one round of selection ranged between 95.0 and 99.6% (16).

Figure 1

Ex vivo detection of transduced donor lymphocytes. (A) Schematic map of integrated SFCMM-2 proviral genome, indicating the HSVTK internal promoter (T). Solid boxes denote long-terminal repeat sequences. ΔLNGFR, modified form of the low-affinity receptor for nerve growth factor; TN, fusion gene encoding a bifunctional protein carrying both HSV-TK activity and neomycin resistance. Arrows indicate transcription promoters. (B) FACS analysis for expression of ΔLNGFR on the cell surface and appropriate controls (Neg.) from peripheral blood (a and b) and bone marrow (c) samples of patient 1 (a) and 8 (b and c). The proportion of positive cells is indicated. (C) FACS analysis for expression of ΔLNGFR on lymphocytes obtained after G418 selection of PBLs from patient 2 and the appropriate control. The proportion of positive cells reported, 75.3%, is derived from comparison with the negative control and is underestimated because analysis of the sample revealed the presence of a cell population that homogeneously expressed the cell surface marker. On the x axis of (B) and (C), it is labeled as the logarithmic scale of green fluorescence. (D) (a and c) PCR analyses for the presence of the HSV-TK sequence on various samples from patient 8. (a) Lanes 3 to 6, peripheral blood samples obtained at monthly intervals beginning at day 3 from the second infusion (Fig. 3); lane 7, bone marrow sample obtained 23 days from infusion; lanes 1 and 2, negative controls; lanes 8 and 9, positive controls at two different levels of detection of positive cells (10−4 and 10−3, respectively). (c) Lanes 1 and 3, two skin biopsy samples consistent for GvHD; lanes 2 and 5, negative controls; lane 4, positive control. (b) PCR analysis from patient 2. Lanes 3 to 6, peripheral blood samples obtained at monthly intervals beginning at day 4 from the second infusion (Fig. 4); lanes 1 and 2, negative controls; lane 7, positive control.

Twelve patients who experienced severe complication after allo-BMT for a hematologic malignancy or immunodeficiency participated in this study. Here we report on the first eight patients for whom an adequate follow-up is available (Table 1). For the treatment of leukemic relapse, SFCMM-2–transduced donor lymphocytes were infused at increasing cell doses, beginning at 105cells per kilogram of body weight to a total of 4 × 107 cells/kg. For the treatment of EBV-BLPD, transduced donor lymphocytes were infused at an initial dose of 1 × 106 cells/kg. The higher initial dose was necessitated by the rapid progression of this complication and was suggested by the previous clinical study (6). No toxicity or complication that could be related to the gene transfer procedure was observed in this study (17).

Table 1

Clinical characteristics of the eight patients described in the text. Disease status was assessed through examination of marrow aspirates and biopsy, cytogenetic examination, and molecular analysis. EBV-BLPD was determined by serologic data and histologic evaluation as indicated. Exclusion criteria for the infusion of transduced donor lymphocytes included the presence of aGvHD grade II or higher and CMV reactivation requiring ganciclovir treatment. This study was approved by the Institutional Ethical Committee (14,31), and all patients gave informed consent. NHL, non–Hodgkin lymphoma; CML, chronic myeloid leukemia; AML, acute myeloid leukemia; CmML, chronic myelomonocytic leukemia; EBV-BLPD, Epstein-Barr virus–induced B lymphoproliferative disorder; CR, complete remission; PR, partial response; NR, no response; NE, not evaluable.

View this table:

Extensive manipulation of the cells that undergo ex vivo gene transfer by retroviral vectors could potentially modify their immune repertoire and their activation status, thus affecting their in vivo survival and function. In all but one of the patients who received transduced donor lymphocytes, genetically modified cells could be repeatedly detected in the circulation, in marrow aspirates, and in tissue biopsies (18). The proportion of transduced donor cells among circulating lymphocytes ranged between the level of detection by polymerase chain reaction (PCR) (10 4) and 13.4% (Fig. 1B). At various times of followup, some patients had only low frequencies of circulating transduced cells, below the level of detection in FACS analysis. In these cases, ex vivo–transduced donor cells could be selected by culture in the presence of G418 (a neomycin analog), and selected cells homogeneously expressed the cell surface marker (Fig. 1C). Additionally, the presence of transduced donor lymphocytes in PBLs, BM samples, and GvHD lesion sites could be confirmed by PCR analysis and followed over time (Fig. 1D). The persistence of transduced cells could still be detected for more than 12 months after the last infusion of transduced donor cells. Persistence of antigen-specific responses was also observed; in the patient treated for an EBV-induced lymphoma, we detected transduced cells that exhibited EBV-specific activity for the first time after administration of donor cells. The frequency of EBV-specific cells was ∼1:1000, as compared to the frequency of 1:1300 that was detected in the donor PBLs before and after ex vivo vector transduction. This level was achieved after administration of a small dose of donor cells (Fig. 2 and Table 1). After administration of ganciclovir to treat GvHD, we could still detect this specific reactivity in the peripheral blood of the patient at the reduced frequency of 1:3250. No linear correlation was observed between the number of infused transduced donor lymphocytes and persistence or detection. Rather, the intensity of antigen response and proliferation of donor cells seemed to affect levels of detection. In particular, patient 1 reached the level of 13.4% of positivity after a total injection of only 106cells/kg to treat EBV-BLPD. Patient 8, who achieved similar levels of positivity (11.9%), did so after receiving an order of magnitude higher infusion of cells to treat chronic myelomonocytic leukemia (CmML) relapse. Other patients who received even higher cell doses did not reach similar levels of positivity. As expected, viral antigens were able to induce a more rapid and intense specific proliferation of transduced cells than leukemic cells (19).

Figure 2

(A) Clinical outcome of donor lymphocyte infusions in a patient affected by EBV-BLPD. Patient 1 underwent a BMT from her HLA-identical and MLC-compatible brother for a high-grade lymphoma in second remission. After successful engraftment, the patient developed EBV-BLPD confirmed by morphological examination of a laterocervical node biopsy and by in situ hybridization for EBV RNA in the nuclei of the neoplastic cells. A myeloaspirate and a bone marrow biopsy showed overt infiltration of the bone marrow by lymphoid paratrabecolar nodules causing marrow failure. The patient received a total dose of 1.5 × 106 cells/kg (6). In the 2 weeks after administration of donor cells (arrow), all clinical symptoms associated with EBV-induced B cell proliferation regressed as shown by body temperature decrease (a). At the time of regression of clinical symptoms, a sharp increase in PBL counts was observed (b). Circulating transduced donor lymphocytes were almost exclusively CD3+CD8+ lymphocytes (>90% of total mononuclear cells from day +10 to day +15), with high proliferation index. Meanwhile, a hematological reconstitution occurred, shown by an increase in the number of circulating platelets (c). During this time, marked donor cells increased progressively in the patient's peripheral blood up to 13.4% of total mononuclear cells (B). (a) Negative control; (b to d) three blood samples obtained at day 3, 10, and 20 after infusion. The proportion of positive cells is indicated.

The in vivo function of genetically modified donor lymphocytes was revealed by antitumor responses, immune reconstitution, and alloreactivity (20). In five out of the eight patients in this series GvL was detected, with three complete responses (Table 1). Three patients developed GvHD. The patients who achieved full remission included patient 1, treated for an aggressive EBV-induced lymphoma, and patient 8, treated for leukemic relapse (Figs. 2 and 3). These results were obtained in the absence of any chemotherapeutic agent. A third patient (patient 7, Table 1) achieved full remission after administration of genetically modified donor lymphocytes and chemotherapy. Patient 3 relapsed twice after two subsequent allogeneic BMTs for CML and could be maintained in CML chronic phase for over 5 years. During this period, all donor cells disappeared from the marrow and the circulation. After the long interval in chronic phase, the CML progressed to accelerated phase and was treated by the infusion of transduced donor lymphocytes (total dose, 4.5 × 106cells/kg). This was followed by conversion to chronic phase and by the reappearance of donor chimerism, as monitored by the progressive conversion of the blood type from B to the donor type A (21). In this patient, the increase of the donor blood type erythrocytes closely followed the appearance of circulating transduced donor lymphocytes.

Figure 3

Clinical outcome in a patient affected by chronic myelomonocytic leukemia (CmML) relapse treated by the infusion of donor lymphocytes. Patient 8 underwent BMT from her HLA-identical MLC-compatible brother but relapsed 3 months after transplant, as revealed by cytogenetic analysis, followed by a decrease in peripheral blood counts and marked morphologic myelodisplastic bone marrow. The patient received a total of 2 × 107cells/kg donor lymphocytes. Marked cells could be easily detected by FACS analysis (Fig. 1B) and PCR (Fig. 1D) in the peripheral blood (reaching a peak at 11.9% of circulating mononuclear cells) and in bone marrow (reaching a peak at 5.3%). Infusions (arrows) were followed by a hematological reconstitution, as shown by the increase in the number of circulating (leukocytes and platelets. Bone marrow cytogenetic analysis revealed a progressive decrease in the number of host hematopoietic cells (shaded area), and 100% donor engraftment (unshaded area) was observed after four infusions.

Thus, the procedures associated with ex vivo gene transfer of the marking and therapeutic genes by the SFCMM-2 vector preserved the in vivo potential of allogeneic PBLs and their antigen-specific reactivity. Additionally, this observation was further confirmed in a small series of patients in which genetically modified and unmodified donor PBLs were used. In a patient with moderate skin GvHD, it was possible to determine by immunostaining and PCR that both cell populations positive and negative for the cell surface marker ΔLNGFR were contributing to the lymphocyte infiltrate present in the lesion (22).

Two patients in this series developed acute GvHD and were treated with ganciclovir (23) (Table 2). Patient 2 relapsed 6 months after allo-BMT and was treated by increasing doses of transduced donor cells. Marked cells could be detected in the peripheral blood by FACS analysis and PCR. After the third infusion of transduced donor lymphocytes (total of 1 × 106 cells/kg), values of the patient's bilirubin and liver function enzymes rapidly increased (Fig. 4). A liver biopsy revealed acute GvHD, and the patient was treated with ganciclovir (10 mg/kg per day). All transduced donor lymphocytes and all GvHD clinical and biochemical signs then disappeared (Fig. 4 and Table 2), in the absence of any local or systemic toxicity. Elimination of transduced donor cells preceded normalization of liver enzymes with similarly sharp kinetics (Fig. 4 and Table 2). Indeed, after 1 day of treatment, genetically modified cells decreased below the level of PCR detection (10 4) and remained below this level until a second infusion of transduced cells was administered 21 days later. Because no immunosuppressive drug was administered to this patient, the complete abrogation of the severe liver GvHD could be attributed exclusively to the expression of the TK transgene. Four weeks after the infusion of donor lymphocytes, patient 1 progressively developed signs of acute GvHD, confirmed by a skin biopsy. The intravenous administration of two doses of ganciclovir (10 mg/kg) resulted in progressive rapid reduction from 13.4% to below PCR detection (Table 2). The treatment was followed by the disappearance of clinical signs of skin GvHD. About 3 weeks after infusions of HSV-TK–transduced cells, patient 8 presented chronic GvHD involving skin, oral mucosa, and lungs. Immunohistochemical analysis and PCR revealed the presence of genetically modified cells in the biopsy. This patient was then treated with ganciclovir (10 mg/kg per day) for 7 days, resulting in improvement of clinical conditions and a decrease of GvHD signs. The correlated reduction in the proportion of circulating transduced donor cells to 2.8% was achieved in the first 24 hours (Table 2). Complete elimination of genetically modified donor lymphocytes could not be obtained despite extended treatment.

Figure 4

Values of bilirubin and liver function enzymes in patient 2, who developed acute GvHD after the infusion of transduced donor lymphocytes. Long arrows, infusions of transduced donor cells; short arrow, timing of the liver biopsy performed to detect acute GvHD (biopsy). The gray area indicates the period of the four infusions of ganciclovir (10 mg/kg per day).

Table 2

Effect of ganciclovir treatment on elimination of genetically modified donor lymphocytes and on GvHD.

View this table:

Although fluctuations in the numbers of circulating transduced lymphocytes were observed at various times of followup, reduction of circulating transduced donor lymphocytes comparable to that described in these three patients was never observed in any of the patients in the study in the absence of ganciclovir treatment. This is best compared to the lack of any spontaneous remissions of a grade II or higher of GvHD in the 258 patients treated with donor lymphocytes in the EBMT (European Bone Marrow Transplantation) international collaborative study (24).

The two treated patients who had achieved full remission before ganciclovir administration remained in full remission after elimination of donor lymphocytes by ganciclovir. Thus, the transfer of the HSV-TK gene by the SFCMM-2 vector may provide a tool for in vivo modulation of alloreactivity and effective and specific treatment of acute GvHD in the absence of any immunosuppressive drug.

As anticipated, no local or systemic toxicity was observed that could be related to the gene transfer procedure or to ganciclovir administration. No significant modifications of the total counts of leukocytes, PBLs, natural killer cells, or other T cell subsets could be detected in patients who had received ganciclovir treatment for GvHD. This confirms that lymphocytes do not produce the in vivo bystander effect observed in other cell types (10,12). The only unwanted effect observed in this study was a specific immune response to HSV-TK–NEO in one patient who had received genetically modified cells late after transplant, after return to immunocompetence (25). The immune response was expressed only to the HSV-TK–NEO fusion protein. No immunity to the cell surface marker ΔLNGFR was detected in any patient (25).

This study confirms the therapeutic potential of donor T cells in the context of allo-BMT (5, 6). Because it is not yet possible to differentiate between GvL and GvHD effector cells and to predict which patients will develop the more severe grade of GvHD after an unmodified allo-BMT (26), immunosuppressive prophylaxis remains an absolute requirement. This immunosuppression and the more intense regimens associated with GvHD that arise despite prophylaxis significantly limit the benefit of allo-BMT, eliminating or reducing the allogeneic advantage (27). No specific treatment yet exists for established GvHD, and immunosuppressive regimens carry significant side effects and limited efficacy (28). In this context, the transfer of a suicide gene for selective and specific elimination of effector cells of GvHD provides a new tool to combine the benefits of allo-PBLs with the possibility of eliminating GvHD without toxic effects. However, the extensive manipulation of the cells that undergo ex vivo gene transfer could potentially modify the clinical outcome produced by transduced donor PBLs as compared to unmodified counterparts, resulting in abrogation or reduction of the allogeneic advantage, or exacerbation of GvHD. Therefore, the long-term survival and immunocompetence of transduced PBLs observed in this study is crucial not only for this model but also for all other applications involving ex vivo gene transfer into polyclonal lymphocytes (29). Because the mechanisms underlying GvHD, GvL, and antiviral responses are likely to be different and to involve different effector cells, the preservation of an intact immune repertoire as well as all functional properties in transduced PBLs is of crucial relevance. We propose that optimal effects achieved were attributable to the rapid positive selection system for transduced cells allowed by the cell surface marker (13). Ganciclovir-mediated elimination of HSV-TK–transduced cells was efficacious in the presence of acute GvHD. In only one case of chronic GvHD was ganciclovir treatment not fully effective. Transduced cells isolated ex vivo from this patient showed unmodified sensitivity to ganciclovir, suggesting that in vivo resistance could indeed be due to the activation state of the genetically modified lymphocytes. This is supported by the well-known cell replication–dependent activity of HSV-TK (30). If confirmed in extended clinical studies, these results will expand the number of candidate patients benefiting from T lymphocyte–depleted allo-BMT, by allowing the use of less-compatible marrow donors.

  • * To whom correspondence should be addressed at Istituto Scientifico San Raffaele, Via Olgettina, 60, 20132 Milan, Italy. E-mail: bordigc{at}


View Abstract

Stay Connected to Science

Navigate This Article