Transfer of either allogeneic or genetically modified T cells as a therapy for malignancies can be accompanied by T cell-mediated tissue destruction. The introduction of an efficient “safety switch” can potentially be used to control the survival of adoptively transferred cell populations and as such reduce the risk of severe graft-vs-host disease. In this study, we have tested the value of an inducible caspase 9-based safety switch to halt an ongoing immune attack in a murine model for cell therapy-induced type I diabetes. The data obtained in this model indicate that self-reactive T cells expressing this conditional safety switch show unimpaired lymphopenia- and vaccine-induced proliferation and effector function in vivo, but can be specifically and rapidly eliminated upon triggering. These data provide strong support for the evaluation of this conditional safety switch in clinical trials of adoptive cell therapy.

The capacity of allogeneic T cells to recognize and destroy tumor cells has been demonstrated in chronic myeloid leukemia patients receiving donor lymphocyte infusions (DLIs)4 following an HLA-matched allogeneic hematopoietic stem cell transplantation (1). This graft-vs-leukemia (GvL) effect is based on the recognition of polymorphic peptide sequences that are presented by HLA molecules on host cells but not donor cells and that are hence considered foreign by the infused T cell population. However, presentation of minor histocompatibility Ags is obviously not restricted to leukemic cells or hematopoietic cell types, and T cell recognition of minor histocompatibility Ags on other host tissues can result in the development of graft-vs-host disease (GvHD). The morbidity and mortality associated with GvHD remains a major limitation of allogeneic hematopoietic stem cell transplantation and the introduction of a “safety-switch” that could be used to eliminate potentially autoaggressive donor T lymphocytes in case of GvHD is therefore desirable.

In addition to treatment protocols involving allogeneic T cell transfer, an increasing effort is made to produce autologous tumor-reactive T cells for T cell therapy. In case of melanoma, infusion of ex vivo expanded tumor-infiltrating lymphocytes has shown considerable promise (2, 3). Furthermore, by introduction of TCR or chimeric receptors (CR) that target tumor-associated self-Ags, a high-avidity tumor-reactive T cell repertoire may also be created for other tumor types (4, 5, 6). These different types of autologous T cell products also have the potential to cause autoimmunity, due to either (on-target) recognition of the intended self-Ag on other tissues, or due to the (off-target) recognition of other self Ags (7, 8). Consequently, the pharmacological control of the in vivo survival of infused autologous T cell products may in certain cases also be desirable.

The most extensively studied safety-switch to date is the HSV I-derived thymidine kinase (HSV-TK) gene product. The HSV-TK safety switch has been shown to be effective in patients who developed GvHD after DLI (9, 10) and these studies provide the first proof of principle that a safety-switch can be effective in the clinical management of GvHD. Despite this, the HSV-TK safety switch has a number of limitations. First, the toxicity of the HSV-TK gene product is based on the enzymatic production of the toxic metabolite ganciclovir triphosphate upon administration of ganciclovir. However, ganciclovir is also frequently used to treat CMV infections in patients who undergo allogeneic stem cell transplantation (allo-SCT), resulting in an unwanted elimination of the genetically modified cells. More importantly, unwanted elimination of the transferred T cells as a consequence of immune responses toward the HSV-TK gene product has been observed in a substantial fraction of patients, likely limiting the use of this safety switch to patients who are immune suppressed at the time of T cell infusion (11, 12). As a first possible nonimmunogenic alternative safety switch, a human CD20 molecule has been validated in preclinical studies (13, 14). Exposure of patients to anti-CD20 mAb could then be used to induce killing of T cells that express this safety switch, but would in clinical practice also lead to an unwanted and prolonged loss of B cells.

As a second, nonimmunogenic safety switch system for clinical use, fusion proteins composed of human proapoptotic molecules linked to modified human FK506-binding proteins (FKBPF36V) have been developed. These safety switches can be activated by injection of a chemical inducer of dimerization (CID), consisting of a dimer of two synthetic variants of FK506. The CID is lipid-permeable and binds at subnanomolar concentrations to FKBPF36V domains, while the affinity for endogenous FKBP is 1000-fold lower (15). Binding of the CID leads to clustering of the linked proapoptotic molecules and thereby induces their activation. A safety switch based on this design and consisting of a modified human caspase 9 molecule fused to FKBPF36V (iCasp9M) has shown substantial potential, with elimination of 99% of T cells expressing high levels of this iCasp9M safety switch by a single dose of the CID (16). Because the iCasp9M safety switch is fully composed of human sequences, immunogenicity of this safety switch is likely to be minimal. In addition, contrary to safety switches that rely on upstream apoptosis initiators, the caspase 9-based suicide switch should be relatively insensitive to alterations in the expression levels of cellular inhibitors of apoptosis. In line with this, the iCasp9M safety switch is functional in a T cell line expressing high levels of c-FLIP and Bcl-xL, whereas the function of a Fas-based safety switch is blocked in these cells (16). Because of these reasons, the iCasp9M safety switch appears an attractive candidate for the pharmacological control of infused T cell populations. However, its effectiveness in the amelioration or blockade of cell therapy-induced autoimmune pathology has not been determined.

In this study, we aimed to test the value of the pharmacologically inducible caspase 9 (iCaspase9)-based safety switch to control cell therapy-induced autoimmunity, using a mouse model for severe and acute autoimmune diabetes. In this model, we show that autoreactive T cells containing the iCasp9M safety switch can be rapidly eliminated in vivo. This blockade of an ongoing autoimmune attack suffices to halt the development of an otherwise lethal type I diabetes.

C57BL/6 (B6), RIP-OVAhigh mice (17), and F1 offspring of Ly5.1+C57BL/6 with OT-I mice were obtained from the Experimental Animal Department of The Netherlands Cancer Institute. All animal experiments were performed in accordance with institutional and national guidelines and were approved by the Experimental Animal Committee of The Netherlands Cancer Institute.

The iCasp9M safety switch (16) and enhanced GFP (eGFP) genes, separated by an internal ribosomal entry sequence (IRES) were cloned into the pMX retroviral vector (18) to obtain pMX-iCasp9M-IRES-GFP. As a control, a pMX vector solely containing the eGFP gene was used (pMX-GFP). B6 splenocytes were modified by retroviral transduction as described previously (19). For transduction of Ly5.1+ OT-I T cells, a modified protocol was used in which splenocyte cultures of OT-I TCR-transgenic mice were supplemented with 5–10% purified CD4+ cells to enhance transduction efficiencies (our unpublished observations). To this purpose, spleens from RIP-OVAhigh mice (that lack detectable OVA-specific CD8+ T cells) were harvested and leukocytes were purified over a Lympholyte-M (Cedarlane Laboratories) gradient. Splenocytes (5 × 107/ml) were incubated with PE-labeled anti-CD4 (1 μg/ml; BD Pharmingen) for 20 min at room temperature in complete medium (IMDM; Invitrogen Life Technologies) supplemented with 8% FCS, 10 μM 2-ME, and 100 U/ml penicillin, 100 μg/ml streptomycin). Cells were washed and incubated with anti-PE beads (Miltenyi Biotec) according to the manufacturer’s protocol. Positive selection was performed by autoMACS (Miltenyi Biotec) according to the manufacturer’s guidelines. Following addition of purified CD4+ cells to the Ly5.1+ OT-I T cells, retroviral transduction was performed as described (19).

Transduction efficiency was measured 24 h after transduction by flow cytometric analysis of GFP expression within the CD8+ cell population. For the measurement of T cell responses, peripheral blood samples were taken at the indicated days posttransfer and treated as previously described (4). Samples were stained with anti-CD8α mAbs (PE or allophycocyanin conjugated), when necessary in combination with PE-conjugated anti-Ly5.1 mAb (all mAbs obtained from BD Pharmingen) and analyzed by flow cytometry. Propidium iodide (Sigma-Aldrich) was used to select for live cells. Data acquisition and analysis was done on a FACSCalibur (BD Biosciences) with CellQuest Pro software. For isolation of GFP+CD8+ T cells, retrovirally transduced splenocytes (50 × 106/ml) were stained with PE-labeled anti-CD8 (0.4 μg/ml; BD Pharmingen). Cells were subsequently sorted on a FACS Aria (BD Biosciences) (filter for GFP 585/30 nm; filter for PE 585/42 nm) using DiVa Software. In each experiment a postsort analysis was performed and ≥95% of the sorted population was GFPbright.

For in vitro experiments, CID (AP20187; ARIAD Pharmaceuticals) was added to transduced splenocytes at the indicated concentrations. In vitro depletion of transduced T cells was determined by flow cytometric analysis of GFP expression in CD8+ cells 24 h post-CID administration. For in vivo experiments, 50 μg of CID diluted in 200 μl of carrier solution (22.5% PEG400, 1.25% Tween 80) was injected once i.p. at the indicated day post-cell transfer. In vivo depletion of transduced T cells was determined by flow cytometric analysis of GFP expression in CD8+ cells in peripheral blood samples at the indicated time points post-cell transfer.

In cell transfer experiments, indicated numbers of Ly5.1+, OT-I T cells were adoptively transferred into RIP-OVAhigh mice. Mice were subsequently vaccinated either by i.p. administration of 1 × 106 PFU of rVV-OVA, a recombinant vaccinia strain that expresses OVA (20) or by intranasal administration of 1000 PFU of influenza A/WSN/33 (WSN)-OVA(I) (21). Alternatively, mice underwent irradiation-induced host conditioning by 5 Gy total body irradiation with a radiobiology constant potential x-ray unit (Pantak HF-320; Pantak Limited), 1 day before adoptive cell transfer. To create a more proinflammatory environment, mice were s.c. injected with 20 μg of CpG (oligodeoxynucleotide (ODN)) 1826, a 20-mer containing two CpG motifs (TTCATGACGTTCCTGACGTT), at the indicated time point. Where indicated, blood glucose levels were monitored by Accu-Check Compact (Roche Diagnostics) measurement. Mice were considered diabetic when blood glucose levels were ≥20 mM/L. The severity of diabetes was determined by the extent of weight loss. Diabetic mice were either sacrificed when experiencing a weight loss of ≥5 g of their initial body weight, when mice became subconscious, or when mice displayed clinical signs of severe discomfort such as a hunched back and reduced activity.

IHC was conducted on pancreata sampled in buffered formalin. For Ag retrieval, sections were pretreated with 0.1 M citrate (pH 6.0; 95–100°C). Endogenous peroxidases were inactivated by incubation with 3% H2O2 in methanol. Sections were preincubated with PBS/4% BSA/5% normal goat serum. As a primary Ab, rabbit anti-murine CD3 (clone, SP7; 1/50 dilution; Neomarkers) was used. Anti-CD3 staining was visualized by a three-step immunoenzymatic procedure. First, biotin labeled goat-anti-rabbit Igs (1/1000 dilution; DakoCytomation) were applied, followed by HRP-labeled avidin-biotin complex (ABC; DakoCytomation). Finally, 3,3-diaminobenzidine-tetrahydrochoride (Sigma-Aldrich) was used as a substrate chromagen and slides were counterstained with hematoxylin. Images were acquired using an Axiocam HR digital camera and processed with Axiovision 4 software (Carl Zeiss Vision).

The iCasp9M safety switch consists of a modified human caspase 9 molecule of which the caspase recruitment domain is removed to prevent physiological dimerization. This truncated caspase 9 molecule is genetically coupled to a modified FKBP (FKBPF36V). Upon administration of a bivalent ligand of FKBPF36V, dimers of the caspase9-FKBPF36V fusion protein are formed, leading to caspase 9 activation (16). To address whether iCasp9M can function as a conditional apoptotic switch in murine T cells, splenocytes were retrovirally transduced with a vector encoding the iCasp9M safety switch plus a GFP reporter. Control cells were transduced with a retrovirus encoding GFP only. Twenty-four hours after transduction, 19.3% of CD8+ cells transduced with iCasp9M retrovirus showed high GFP expression (Fig. 1,A, middle panel) and 28.4% of CD8+ cells transduced with the control retrovirus showed high GFP expression (Fig. 1,A, right panel). To determine the sensitivity of murine T cells for triggering of the iCasp9M safety switch, the retrovirally transduced splenocytes were cultured for 24 h in presence of CID at various concentrations (range 0–100 nM). Subsequently, the percentage of GFP-expressing cells was calculated by comparing GFP expression in cultures exposed to CID to that in cultures not exposed to CID (Fig. 1,B). As is the case for human T cells (16), a single administration of CID at ≥1 nM results in the elimination of ∼98% of the iCasp9-GFPbright cells within 24 h (Fig. 1 C). The percentage of GFP-expressing cells in cultures transduced with the GFP control construct was unaffected by increasing doses of CID, demonstrating that the elimination of GFPbright cells is due to the induced dimerization of the modified caspase 9 molecules and not by some unrelated toxicity of the CID itself. Notably, iCasp9M-IRES-GFP-transduced T cells expressing intermediate levels of GFP are only moderately sensitive to CID administration, indicating that a critical expression level of iCaspase9 is required for the induction of apoptosis (16).

FIGURE 1.

Murine T cells expressing high levels of iCasp9M are eliminated after CID administration. A, Murine B6 splenocytes were transduced with pMX iCasp9M-IRES-GFP (middle) or with pMX-GFP (right). After 24 h, transduction efficiency was determined by measurement of GFP expression. Numbers in the left upper corners represent the percentage of GFP+CD8+ cells of total CD8+ cells. B and C, Twenty-four hours after transduction, CID was administered to the transduced cell populations at the indicated concentrations (range 0–100 nM). Twenty-four hours after CID administration, cells were analyzed for GFP expression to determine the sensitivity toward the CID in vitro. The numbers in the left upper corners (B) or on the y-axis (C) reflect the percentage of GFPbright cells in CID-exposed cultures, relative to the percentage of GFPbright cells in cultures not exposed to CID. Error bars in C represent the SD of three cultures per concentration of CID.

FIGURE 1.

Murine T cells expressing high levels of iCasp9M are eliminated after CID administration. A, Murine B6 splenocytes were transduced with pMX iCasp9M-IRES-GFP (middle) or with pMX-GFP (right). After 24 h, transduction efficiency was determined by measurement of GFP expression. Numbers in the left upper corners represent the percentage of GFP+CD8+ cells of total CD8+ cells. B and C, Twenty-four hours after transduction, CID was administered to the transduced cell populations at the indicated concentrations (range 0–100 nM). Twenty-four hours after CID administration, cells were analyzed for GFP expression to determine the sensitivity toward the CID in vitro. The numbers in the left upper corners (B) or on the y-axis (C) reflect the percentage of GFPbright cells in CID-exposed cultures, relative to the percentage of GFPbright cells in cultures not exposed to CID. Error bars in C represent the SD of three cultures per concentration of CID.

Close modal

RIP-OVAhigh mice express OVA in the insulin-producing β-cells of the pancreas. Whereas the endogenous T cell repertoire is immune tolerant for OVA (4), adoptive transfer of T cells expressing the OT-I TCR that recognizes the class I-restricted epitope OVA257–264 followed by vaccination leads to destruction of the β cells, resulting in a severe form of autoimmune diabetes (4, 17). The induction of diabetes in this model is extremely rapid, with blood glucose levels changing from normoglycemic into >20 mM/L within 24 h. Development of diabetes results in a weight loss of >1 g (>5%) in the same time span and death within <1 wk after disease onset if mice are left untreated. To assess whether T cells that express high levels of the conditional safety switch are functional in vivo, Ly5.1+ OT-I TCR-transgenic T cells were retrovirally transduced with the iCasp9M safety switch, sorted 24 h posttransduction for GFP expression and transferred into Ly5.1RIP-OVAhigh recipients. Subsequently, the mice were vaccinated with a recombinant influenza strain expressing OVA (inflova). Analysis of peripheral blood samples revealed that Ag-specific T cells that express the iCasp9M safety switch do proliferate upon Ag encounter in vivo (Fig. 2,A). Importantly, during this Ag-driven proliferation, there was no evidence of preferential outgrowth of either GFPdull or GFP-negative Ly5.1+ cells (Fig. 2,B). This indicates that during a physiological in vivo T cell response, the iCasp9M safety switch shows no basal toxicity. Consistent with this, recipients of iCasp9M modified T cells all became diabetic within 2 wk post-T cell infusion (Fig. 2 C).

FIGURE 2.

T cells expressing the iCasp9M safety switch function in vivo. Ly5.1+ OT-I T cells were retrovirally transduced with pMX iCasp9M-IRES-GFP. Twenty-four hours after transduction, CD8+ T cells were sorted for GFP expression. A total of 2 × 104 Ly5.1+ OT-I GFPbright cells were adoptively transferred into Ly5.1 RIP-OVAhigh mice, followed by intranasal vaccination with 1000 PFU inflova. From day 3 on, the percentage of GFP+Ly5.1+CD8+ T cells and blood glucose levels were analyzed daily in peripheral blood. A, Kinetics of GFP+CD8+ cell responses in peripheral blood. Circles represents values per mouse; bars represent average values. B, Analysis of GFP expression in Ly5.1+CD8+ T cells in peripheral blood. Shown are dot plots (gated on CD8+ cells) of blood samples of one representative mouse on 6 subsequent days. The number in the left upper corner represents the percentage of GFP+ cells within the CD8+ T cell population. C, Blood glucose levels were measured to determine the induction of diabetes. Mice were considered diabetic when blood glucose levels were ≥20 mM/L.

FIGURE 2.

T cells expressing the iCasp9M safety switch function in vivo. Ly5.1+ OT-I T cells were retrovirally transduced with pMX iCasp9M-IRES-GFP. Twenty-four hours after transduction, CD8+ T cells were sorted for GFP expression. A total of 2 × 104 Ly5.1+ OT-I GFPbright cells were adoptively transferred into Ly5.1 RIP-OVAhigh mice, followed by intranasal vaccination with 1000 PFU inflova. From day 3 on, the percentage of GFP+Ly5.1+CD8+ T cells and blood glucose levels were analyzed daily in peripheral blood. A, Kinetics of GFP+CD8+ cell responses in peripheral blood. Circles represents values per mouse; bars represent average values. B, Analysis of GFP expression in Ly5.1+CD8+ T cells in peripheral blood. Shown are dot plots (gated on CD8+ cells) of blood samples of one representative mouse on 6 subsequent days. The number in the left upper corner represents the percentage of GFP+ cells within the CD8+ T cell population. C, Blood glucose levels were measured to determine the induction of diabetes. Mice were considered diabetic when blood glucose levels were ≥20 mM/L.

Close modal

Having established that mouse T cells expressing the iCasp9M safety switch proliferate upon Ag encounter in vivo and perform effector functions, we wished to determine whether the activity of iCasp9M-modified T cells can be halted by CID administration. To first develop a preclinical parameter that would report on an ongoing autoimmune attack, we analyzed whether the presence of OT-I T cells in peripheral blood samples can be used as a measure of an ongoing autoimmune attack within the islets of Langerhans. To this purpose, RIP-OVAhigh mice received an adoptive transfer of 5 × 104 Ly5.1+OT-I T cells, followed by vaccination with rVV-OVA. Control mice were only vaccinated. To determine the percentage of OT-I T cells and blood glucose levels, blood was sampled daily from day 3 on. Mice were killed to harvest pancreatic tissue for IHC when OT-I cells could first be detected in peripheral blood, but blood glucose levels were still normal. In four of four recipients of OT-I T cells with detectable OT-I T cell responses in peripheral blood (average T cell response of 2.7%), a clear infiltrate of CD3+ cells could be detected by IHC in the islets of Langerhans. As a control, no CD3+ cells could be detected in pancreatic sections of mice that had not received OT-I T cells (Fig. 3, Table I). Furthermore, in line with the fact that mice are still normoglycemic when OT-I T cell responses are first detected in peripheral blood, the majority of β cells in sections of these mice are still intact (Fig. 3). Notably, recipients of OT-I T cells that were not sacrificed at the time point at which OT-I T cell responses became detectable in peripheral blood all became diabetic within 24 h after the first detection of Ag-specific T cells (Table I). In pancreatic sections of these mice, a dense infiltrate of CD3+ T cells was apparent throughout the whole islet (Fig. 3, Table I). These data indicate that the detection of OT-I T cells in peripheral blood can be used as a diagnostic indicator for an ongoing but incomplete β cell attack, thereby allowing one to determine the feasibility of halting such an attack by activation of a conditional safety switch. Importantly, as clinically manifest diabetes occurs within 24 h after the first detection of OT-I T cells in peripheral blood (Table I), the result of such an activation needs to be swift.

FIGURE 3.

T cell responses in peripheral blood are indicative of an ongoing autoimmune attack in the islets of Langerhans of RIP-OVAhigh mice. Ly5.1RIP-OVAhigh mice received an adoptive transfer of 5 × 104 Ly5.1+, OT-I T cells, followed by i.p. infection with 1 × 106 PFU rVV-OVA. Control mice were only infected with rVV-OVA. From day 3 on, the percentage of Ly5.1+CD8+ T cells and blood glucose levels were analyzed in peripheral blood. One group of mice (n = 4) was killed when OT-I cells could first be detected, and blood glucose levels were still normal in these mice. A second group of mice (n = 4) was killed, once diabetes has developed. Pancreata were subsequently harvested to determine the presence of CD3+ cells by IHC. Islets of Langerhans of a control mouse (top), a mouse killed at the first time point when detectable frequencies of the adoptively transferred cells are present (middle), and of a mouse killed after the development of diabetes (bottom). Arrows indicate CD3+ T cells as detected by IHC. Original magnification: ×40. In this and subsequent experiments, rVV-OVA rather than “inflova” was used for vaccination, because inflova-induced pneumonia can complicate the long-term monitoring of diabetes (M. A. de Witte, unpublished observations).

FIGURE 3.

T cell responses in peripheral blood are indicative of an ongoing autoimmune attack in the islets of Langerhans of RIP-OVAhigh mice. Ly5.1RIP-OVAhigh mice received an adoptive transfer of 5 × 104 Ly5.1+, OT-I T cells, followed by i.p. infection with 1 × 106 PFU rVV-OVA. Control mice were only infected with rVV-OVA. From day 3 on, the percentage of Ly5.1+CD8+ T cells and blood glucose levels were analyzed in peripheral blood. One group of mice (n = 4) was killed when OT-I cells could first be detected, and blood glucose levels were still normal in these mice. A second group of mice (n = 4) was killed, once diabetes has developed. Pancreata were subsequently harvested to determine the presence of CD3+ cells by IHC. Islets of Langerhans of a control mouse (top), a mouse killed at the first time point when detectable frequencies of the adoptively transferred cells are present (middle), and of a mouse killed after the development of diabetes (bottom). Arrows indicate CD3+ T cells as detected by IHC. Original magnification: ×40. In this and subsequent experiments, rVV-OVA rather than “inflova” was used for vaccination, because inflova-induced pneumonia can complicate the long-term monitoring of diabetes (M. A. de Witte, unpublished observations).

Close modal
Table I.

Analysis immune response and glucose levels in peripheral blood and CD3+ cells in pancreas in RIP-OVAhigh mice after vaccination with rVV-OVA+/− transfer of OT-I T cells

Animal No.TreatmentAnalysis PBLaDay KilledPresence CD3+ Cells in Pancreasb
Day 3Day 4Day 5
Transfer OT-I T cells + vaccination rvv-OVA 0.1 (N) 2.1 (N) 8.1 (24.3) 
Transfer OT-I T cells + vaccination rvv-OVA 0.1 (N) 2.0 (N) 23.0 (33.1) 
Transfer OT-I T cells + vaccination rvv-OVA 0.1 (N) 0.9 (N) 21.8 (31.2) 
Transfer OT-I T cells + vaccination rvv-OVA 0.1 (N) 2.9 (N)  +/− 
Transfer OT-I T cells + vaccination rvv-OVA 0.2 (N) 2.5 (N)  +/− 
Transfer OT-I T cells + vaccination rvv-OVA 0.2 (N) 3.0 (N)  +/− 
Transfer OT-I T cells + vaccination rvv-OVA 0.2 (N) 2.5 (N)  +/− 
Transfer OT-I T cells + vaccination rvv-OVA 0.1 (N) 2.4 (N) 13.6 (26.2) 
Vaccination rvv-OVA 0.1 (N) 0.2 (N) 0 (N) − 
10 Vaccination rvv-OVA 0.1 (N) 0.1 (N) 0 (N) − 
Animal No.TreatmentAnalysis PBLaDay KilledPresence CD3+ Cells in Pancreasb
Day 3Day 4Day 5
Transfer OT-I T cells + vaccination rvv-OVA 0.1 (N) 2.1 (N) 8.1 (24.3) 
Transfer OT-I T cells + vaccination rvv-OVA 0.1 (N) 2.0 (N) 23.0 (33.1) 
Transfer OT-I T cells + vaccination rvv-OVA 0.1 (N) 0.9 (N) 21.8 (31.2) 
Transfer OT-I T cells + vaccination rvv-OVA 0.1 (N) 2.9 (N)  +/− 
Transfer OT-I T cells + vaccination rvv-OVA 0.2 (N) 2.5 (N)  +/− 
Transfer OT-I T cells + vaccination rvv-OVA 0.2 (N) 3.0 (N)  +/− 
Transfer OT-I T cells + vaccination rvv-OVA 0.2 (N) 2.5 (N)  +/− 
Transfer OT-I T cells + vaccination rvv-OVA 0.1 (N) 2.4 (N) 13.6 (26.2) 
Vaccination rvv-OVA 0.1 (N) 0.2 (N) 0 (N) − 
10 Vaccination rvv-OVA 0.1 (N) 0.1 (N) 0 (N) − 
a

Numbers indicate percentage of Ly5.1+CD8+ T cells. Numbers in parentheses indicate blood glucose levels (mM/L); N = glucose level <12 mM/L.

b

Definitions: −, no CD3+ cells; +/−, some islands show infiltrations, some do not; +, all islands show infiltrations.

To assess the feasibility of abrogating an ongoing autoimmune attack, RIP-OVAhigh mice received an adoptive transfer of 2 × 104 iCasp9M-transduced Ly5.1+ OT-I T cells sorted for high GFP expression (Fig. 4,A) and were subsequently vaccinated with rVV-OVA. At the moment the percentage of Ly5.1+ T cells in peripheral blood exceeded 5% (either day 5 or 6, Fig. 4, B and C), mice were either left untreated or were given an infusion of CID (50 μg). A single infusion of CID resulted in a ∼85% reduction in the number of GFP+Ly5.1+ T cells as compared with untreated mice (average T cell responses of 1.3 and 8.2% at day 7 posttransfer). The remaining Ly5.1+ population displayed low to intermediate expression of the GFP marker gene, also demonstrating that in vivo, a threshold of iCasp9M expression is required to render T cells sensitive to CID administration (Fig. 4, B and C). Indeed, when gated on GFPbright cells, in vivo depletion of the adoptively transferred cells was close to complete (∼98%, average T cell responses of 0.12 and 5.4% at day 7 posttransfer; Fig. 4, B and C). A similar efficiency of elimination of adoptively transferred cells was also seen when mice were treated after having already reached the diabetic state (data not shown).

FIGURE 4.

Function of iCasp9M safety switch in an autoimmune diabetes model. A, Ly5.1+OT-I T cells were retrovirally transduced with pMX iCasp9M-IRES-GFP. Twenty-four hours after transduction, CD8+ T cells were sorted for high GFP expression (indicate by rectangle). A total of 2 × 104 Ly5.1+OT-I GFPbright cells were adoptively transferred into Ly5.1RIP-OVAhigh mice, followed by i.p. vaccination with 1 × 106 PFU rVV-OVA. From day 3 on, the percentage of GFP+Ly5.1+CD8+ T cells and blood glucose levels were analyzed daily in peripheral blood. Mice either received 50 μg of CID i.p. when the percentage of Ly5.1+ cells was ≥5% of total CD8+ cells (indicated by CID), or were left untreated (control). B, Analysis of GFP expression in Ly5.1+CD8+ T cells in peripheral blood. Shown are dot plots of blood samples of one mouse per treatment group on 3 subsequent days. The bold number in the right upper corner represents the percentage of GFPbright cells of total CD8+ T cells; the number in parentheses represents the percentage of total GFP+ cells within the CD8+ T cell population. C, Kinetics of GFP+CD8+ cell responses (top panels) or GFPbrightCD8+ cell responses (bottom panels) in peripheral blood. Circles represent values per mouse; bars represent average values. •, Samples taken before administration of CID; ○, samples taken after administration of CID. D, Blood glucose levels were measured to determine the induction of diabetes. Mice were considered diabetic when blood glucose levels were ≥20 mM/L. E, The severity of diabetes as indicated by the extent of weight loss. Mice were sacrificed when weight loss reached or exceeded 5 g of the initial body weight (starting weight 20–25 g). F, Ly5.1RIP-OVAhigh mice received either 2 × 104 (left panels) or 2 × 105 (right panels) Ly5.1+OT-I GFPbright cells and were subsequently vaccinated by i.p. administration of 1 × 106 PFU rVV-OVA. Peripheral blood samples were analyzed as in B, data are depicted as in C. Note that the suppression of self-reactive T cell responses is independent of the original T cell dose.

FIGURE 4.

Function of iCasp9M safety switch in an autoimmune diabetes model. A, Ly5.1+OT-I T cells were retrovirally transduced with pMX iCasp9M-IRES-GFP. Twenty-four hours after transduction, CD8+ T cells were sorted for high GFP expression (indicate by rectangle). A total of 2 × 104 Ly5.1+OT-I GFPbright cells were adoptively transferred into Ly5.1RIP-OVAhigh mice, followed by i.p. vaccination with 1 × 106 PFU rVV-OVA. From day 3 on, the percentage of GFP+Ly5.1+CD8+ T cells and blood glucose levels were analyzed daily in peripheral blood. Mice either received 50 μg of CID i.p. when the percentage of Ly5.1+ cells was ≥5% of total CD8+ cells (indicated by CID), or were left untreated (control). B, Analysis of GFP expression in Ly5.1+CD8+ T cells in peripheral blood. Shown are dot plots of blood samples of one mouse per treatment group on 3 subsequent days. The bold number in the right upper corner represents the percentage of GFPbright cells of total CD8+ T cells; the number in parentheses represents the percentage of total GFP+ cells within the CD8+ T cell population. C, Kinetics of GFP+CD8+ cell responses (top panels) or GFPbrightCD8+ cell responses (bottom panels) in peripheral blood. Circles represent values per mouse; bars represent average values. •, Samples taken before administration of CID; ○, samples taken after administration of CID. D, Blood glucose levels were measured to determine the induction of diabetes. Mice were considered diabetic when blood glucose levels were ≥20 mM/L. E, The severity of diabetes as indicated by the extent of weight loss. Mice were sacrificed when weight loss reached or exceeded 5 g of the initial body weight (starting weight 20–25 g). F, Ly5.1RIP-OVAhigh mice received either 2 × 104 (left panels) or 2 × 105 (right panels) Ly5.1+OT-I GFPbright cells and were subsequently vaccinated by i.p. administration of 1 × 106 PFU rVV-OVA. Peripheral blood samples were analyzed as in B, data are depicted as in C. Note that the suppression of self-reactive T cell responses is independent of the original T cell dose.

Close modal

Notably, the rapid elimination of the adoptively transferred cells requires the combination of CID and iCasp9M, as the administration of CID to mice that received GFP-transduced cells, or the administration of carrier to mice that received iCasp9M transduced cells were both without effect (Fig. 5).

FIGURE 5.

Effectiveness and specificity of the iCasp9M-CID safety switch system. Ly5.1+OT-I T cells were retrovirally transduced with pMX iCasp9M-IRES-GFP or with pMX-GFP as a control. Twenty-four hours after transduction, CD8+ T cells were sorted for high GFP expression. A total of 1 × 104 Ly5.1+OT-I GFPbright cells were adoptively transferred into Ly5.1RIP-OVAhigh mice, followed by i.p. vaccination with 1 × 106 PFU rVV-OVA. From day 3 on, the percentage of GFP+Ly5.1+CD8+ cells and blood glucose levels were analyzed daily in peripheral blood. When the percentage of Ly5.1+ cells was ≥5% of total CD8+ cells, mice either received 50 μg of CID i.p. (left and middle panel) or an equal volume of carrier. Shown are the kinetics of GFPbrightCD8+ cell responses in peripheral blood. Circles represent values per mouse; bars represent average values. •, Samples taken before administration of CID or carrier; ○, samples taken after administration of CID or carrier.

FIGURE 5.

Effectiveness and specificity of the iCasp9M-CID safety switch system. Ly5.1+OT-I T cells were retrovirally transduced with pMX iCasp9M-IRES-GFP or with pMX-GFP as a control. Twenty-four hours after transduction, CD8+ T cells were sorted for high GFP expression. A total of 1 × 104 Ly5.1+OT-I GFPbright cells were adoptively transferred into Ly5.1RIP-OVAhigh mice, followed by i.p. vaccination with 1 × 106 PFU rVV-OVA. From day 3 on, the percentage of GFP+Ly5.1+CD8+ cells and blood glucose levels were analyzed daily in peripheral blood. When the percentage of Ly5.1+ cells was ≥5% of total CD8+ cells, mice either received 50 μg of CID i.p. (left and middle panel) or an equal volume of carrier. Shown are the kinetics of GFPbrightCD8+ cell responses in peripheral blood. Circles represent values per mouse; bars represent average values. •, Samples taken before administration of CID or carrier; ○, samples taken after administration of CID or carrier.

Close modal

One hundred percent of the mice that were left untreated developed type I diabetes within the first week post-cell transfer. Furthermore, all mice experienced a severe weight loss (up to 7 g within 5 days upon diabetes development), and had to be sacrificed within a week after the onset of diabetes (Fig. 4,E, top panel). In contrast, in mice that were treated by CID administration at the moment when an autoreactive T cell response became apparent in peripheral blood, there was a marked amelioration of diabetes development. Four of six mice developed a transient diabetic state, characterized by a reduced weight loss and return to normoglycemia within a month, and the remaining two mice stayed normoglycemic. Importantly, long-term survival and stable normoglycemia was 100% in this treatment group, indicating that the one time disruption of the autoaggressive T cell response was sufficient to control cell therapy-induced disease (Fig. 4, D and E, bottom panel). To test whether induction of a proinflammatory environment could break this state of self-tolerance, mice received CpG ODN 2 wk postadministration of CID. In none of the animals was an expansion of the Ly5.1+GFPdull OT-I T cell compartment observed and all animals remained normoglycemic throughout the period of observation (data not shown). To address whether the efficient deletion of a self-Ag-reactive T cell population is feasible regardless of the original T cell dose, mice were infused with either 2 × 104 or 2 × 105 iCasp9M-transduced Ly5.1+OT-I T cells and were subsequently vaccinated with rVV-OVA. OT-I T cell responses developed significantly more rapidly in recipients of the high T cell dose (p < 0.01 at day 5). However, peak T cell responses were essentially indistinguishable, consistent with the prior observation that at higher precursor frequencies, the magnitude of CD8 T cell responses does not correlate with T cell input (22). Importantly, regardless of the original T cell dose, suppression of the OT-I T cell response by CID administration was highly efficient (Fig. 4 F).

In clinical trials of adoptive T cell therapy (ACT) (23), (non)myeloablative irradiation or chemotherapy rather than vaccination is often used to promote engraftment of the infused T cells. Due to the fact that vaccination- and lymphopenia-induced T cell proliferation are driven by distinct mechanisms, the resultant T cell populations show clear differences with respect to both functional properties and persistence. Specifically, while vaccination results in the rapid emergence of a highly differentiated pool of effector T cells (24), T cell populations induced by host conditioning display properties of memory T cells (25), translating into an enhanced capacity for long-term persistence. Consistent with the less differentiated state of T cell populations generated through lymphopenia-induced proliferation, ACT with OT-I T cells rarely leads to the induction of type I diabetes in RIP-OVAhigh mice (data not shown).

To test whether triggering of the iCasp9 M safety switch can also be used to halt lymphopenia-induced T cell expansion, RIP-OVAhigh recipients were irradiated and then received 1 × 105 Ly5.1+ iCasp9M-bright OT-I T cells. At day 6 post transfer, mice either received the chemical inducer of dimerization or were left untreated. In all recipients of the CID a marked and rapid reduction in the percentage of GFP+, Ly5.1+ OT-I T cells was observed as compared with untreated mice (average T cell responses of 7.6% and 25.5% at day 10 post transfer; Fig. 6, top panel). Furthermore, the reduction in GFPbright T cell numbers was close to complete (∼97%, average T cell responses of 0.3% and 10.6% at day 10 post transfer; Fig. 6, bottom panel). Importantly, this single administration was sufficient to maintain OT-I T cell responses at this low level throughout the course of the experiment. These data indicate that the iCasp9M safety switch can also be used to significantly reduce lymphopenia-induced T cell responses.

FIGURE 6.

Suppression of lymphopenia-induced T cell proliferation by the iCasp9M-safety switch. Ly5.1+ OT-I T iCasp9M-modified CD8+ T cells were generated as in the top panel of Fig. 4. A total of 1 × 105 Ly5.1+OT-I GFPbright cells were adoptively transferred into Ly5.1RIP-OVAhigh mice, which had received a sublethal dose of irradiation the previous day. From day 3 on, the percentage of GFP+Ly5.1+CD8+ T cells was analyzed in peripheral blood. Mice either received 50 μg of CID i.p. on day 6 posttransfer (indicated by CID), or were left untreated (control). Shown are the kinetics of GFP+CD8+ cell responses (top panels) or GFPbrightCD8+ cell responses (bottom panels) in peripheral blood. Circles represent values per mouse; bars represent average values. •, Samples taken before administration of CID; ○, samples taken after administration of CID.

FIGURE 6.

Suppression of lymphopenia-induced T cell proliferation by the iCasp9M-safety switch. Ly5.1+ OT-I T iCasp9M-modified CD8+ T cells were generated as in the top panel of Fig. 4. A total of 1 × 105 Ly5.1+OT-I GFPbright cells were adoptively transferred into Ly5.1RIP-OVAhigh mice, which had received a sublethal dose of irradiation the previous day. From day 3 on, the percentage of GFP+Ly5.1+CD8+ T cells was analyzed in peripheral blood. Mice either received 50 μg of CID i.p. on day 6 posttransfer (indicated by CID), or were left untreated (control). Shown are the kinetics of GFP+CD8+ cell responses (top panels) or GFPbrightCD8+ cell responses (bottom panels) in peripheral blood. Circles represent values per mouse; bars represent average values. •, Samples taken before administration of CID; ○, samples taken after administration of CID.

Close modal

The development of treatment-induced autoimmunity remains a major concern in T cell-based immunotherapy. Adoptive cell therapy in the form of HLA-matched allo-SCT/donor lymphocyte infusions is an accepted strategy for the treatment of a number of hematological malignancies (26). However, in addition to the desired GvL effect, this therapy frequently also results in the development of GvHD, and the development of GvHD has been a major cause for morbidity and mortality upon allo-SCT/DLI. Unfortunately, efforts to segregate the GvL and GvHD effects of DLI have to date not been successful (27).

A second form of ACT in which the inclusion of safety switch systems may be of value involves the transfer of autologous gene-modified T cells. Such gene modification may either involve the introduction of chimeric receptor genes or full-length TCR genes, with the aim to convey tumor specificity onto the genetically modified T cell population. Several scenarios in which TCR- or CR-modified T cells can induce autoimmunity by recognition of self-Ags on healthy tissue are conceivable (8). First, TCR gene transfer may result in recognition of unknown self-peptide MHC complexes (self-pMHC), for instance through the formation of mixed dimers of endogenous and exogenous TCR chains, or by activation of ignorant self-reactive T cells. Although this type of “off-target” autoimmunity has not been seen in mouse studies (4) or in a first clinical trial in melanoma patients (5), this clearly does not exclude the possibility that such side effects will occur when other TCRs are used, or when conditioning regimens or adjuvant treatments are modified. In addition to these off-target effects, the potential occurrence of on-target autoimmunity by CR- or TCR-modified T cells may be a reason for concern, in particular for target Ags that are also expressed in vital tissues (28, 29).

Safety switch systems such as HSV-TK, CD20, Myc-tagged transgenes, and iCasp9M (9, 13, 16, 30) that can be used for the conditional elimination of infused T lymphocytes in either of the above clinical settings should meet three conditions. First implementation of the system in a clinical setting should be readily feasible. Second, introduction of the safety switch should have no deleterious effect on the function or survival of the infused cell product. Third, upon triggering, the elimination of the safety switch-expressing cells should be rapid, efficient, and selective.

With regard to the implementation in a clinical setting, the generation of cells that show uniform expression of the safety switch will be essential for most if not all clinical applications of genetically encoded safety switch systems. Furthermore, such expression may need to exceed a certain threshold level, as is the case for iCasp9M-induced apoptosis and also for complement-directed cytotoxicity via CD20 (31). It is noted that even though transduced T cells were sorted for GFPbright expression in the current experiments (Fig. 4,A), Ly5.1+GFPdull cells were apparent in peripheral blood, and these cells were refractory to CID treatment (Fig. 4 B). Interestingly, the presence of this CID-resistant population did not result in lethal diabetes, even when mice were challenged with CpGs 2 wk post-CID infusion. The absence of diabetes despite the presence of a residual population of infused T cells may form a reflection of the fact that development of diabetes in this model is cell dose dependent (32). In addition, it seems possible that the CID-resistant population is enriched in quiescent T cells. Prior studies have demonstrated that long terminal repeat-driven transgene expression in retrovirally modified T cells is decreased in quiescent cells (33, 34) and analysis of ex vivo blood samples indicates that the average forward and sideward scatter of GFPbright cells is indeed larger than that of GFPdull cells (p < 0.005 for both parameters; data not shown). It may therefore be useful to determine whether the presence of such marker-genedull populations upon ACT can be reduced by inclusion of gene elements that promote maintenance of transgene expression in quiescent cells (35).

Although the CD20 and Myc-tagged systems inherently offer the possibility for the selection of gene-modified cells on magnetic bead-based systems, for the iCasp9M system such selection will require the additional incorporation of a nonimmunogenic marker gene such as the truncated nerve growth factor receptor (36). In particular for obtaining the required cell doses for clinical application, inclusion of such a marker gene does seem an essential next step.

With regard to the second issue, the effect of safety switch expression on the function and survival of the infused cell product; the current data show that introduction of the iCasp9M safety switch has no detectable effect on the in vivo potential of the gene-modified cells in a murine model system. Furthermore, immunogenicity of the gene-modified cells is likely to be minimal, because of the lack of substantial amounts of nonhuman sequence. This offers a substantial advantage over the HSV-TK system, where the survival of gene-modified cells may be limited due the immunogenicity of the HSV-TK gene product (11), and where ganciclovir treatment of CMV infection can also lead to the untimely elimination of the infused cell population.

Finally, with regard to the speed, effectiveness, and selectivity of the different safety switch systems: while anti-CD20 treatment leads to a prolonged depletion of B cells up to 6 mo (37), administration of CID has no detrimental effects other than on the targeted gene-modified cells. Furthermore, while the effectiveness of all four systems is likely to be comparable, T cell depletion via iCasp9M may be somewhat more rapid due to the direct activation of an effector caspase. If such a difference in kinetics can be confirmed in a direct comparison, this may be considered an advantage in settings where a rapid elimination of the infused cell population is required.

In summary, this study has demonstrated in a murine in vivo model that a conditional caspase 9-based safety switch is effective and rapid, and that the triggering of this switch can stall an ongoing autoimmune attack that is lethal if untreated. Provided that genetically modified cells with a homogeneous and high iCasp9M can readily be generated or selected, these properties, combined with the high selectivity and low immunogenicity of this system, form a clear incentive to use this safety switch in clinical trials.

We thank Marly van den Boom, Frank van Diepen, and Anita Pfauth for cell sorting, Ji-Ying Song for reviewing and imaging of the IHC slides, and Gavin Bendle for review of the manuscript. We thank ARIAD Pharmaceuticals for the generous supply of AP20187 and M. J. van Stipdonk (Leiden University Medical Center, Leiden, The Netherlands) for CpG ODN.

The authors have no financial conflict of interest.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1

This work was supported by the European Union FP6 Program ATTACK and Koningin Wilhelmina Fonds Grant 2003-2860.

4

Abbreviations used in this paper: DLI, donor lymphocyte infusion; GvL, graft-vs-leukemia; GvHD, graft-vs-host disease; CR, chimeric receptor; TK, thymidine kinase; allo-SCT, allogeneic stem cell transplantation; CID, chemical inducer of dimerization; FKBP, FK506-binding protein; IRES, internal ribosomal entry site; iCaspase9, inducible caspase 9; ODN, oligodeoxynucleotide; IHC, immunohistochemistry; ACT, adoptive T cell therapy.

1
Kolb, H. J., A. Schattenberg, J. M. Goldman, B. Hertenstein, N. Jacobsen, W. Arcese, P. Ljungman, A. Ferrant, L. Verdonck, D. Niederwieser, et al
1995
. Graft-versus-leukemia effect of donor lymphocyte transfusions in marrow grafted patients. European Group for Blood and Marrow Transplantation Working Party Chronic Leukemia.
Blood
86
:
2041
-2050.
2
Dudley, M. E., J. R. Wunderlich, P. F. Robbins, J. C. Yang, P. Hwu, D. J. Schwartzentruber, S. L. Topalian, R. Sherry, N. P. Restifo, A. M. Hubicki, et al
2002
. Cancer regression and autoimmunity in patients after clonal repopulation with antitumor lymphocytes.
Science
298
:
850
-854.
3
Yee, C., J. A. Thompson, D. Byrd, S. R. Riddell, P. Roche, E. Celis, P. D. Greenberg.
2002
. Adoptive T cell therapy using antigen-specific CD8+ T cell clones for the treatment of patients with metastatic melanoma: in vivo persistence, migration, and antitumor effect of transferred T cells.
Proc. Natl. Acad. Sci. USA
99
:
16168
-16173.
4
de Witte, M. A., M. Coccoris, M. C. Wolkers, M. D. van den Boom, E. M. Mesman, J. Y. Song, M. van der Valk, J. B. Haanen, T. N. Schumacher.
2006
. Targeting self-antigens through allogeneic TCR gene transfer.
Blood
108
:
870
-877.
5
Morgan, R. A., M. E. Dudley, J. R. Wunderlich, M. S. Hughes, J. C. Yang, R. M. Sherry, R. E. Royal, S. L. Topalian, U. S. Kammula, N. P. Restifo, et al
2006
. Cancer regression in patients after transfer of genetically engineered lymphocytes.
Science
314
:
126
-129.
6
Willemsen, R. A., R. Debets, P. Chames, R. L. Bolhuis.
2003
. Genetic engineering of T cell specificity for immunotherapy of cancer.
Hum. Immunol.
64
:
56
-68.
7
Lamers, C. H., S. Sleijfer, A. G. Vulto, W. H. Kruit, M. Kliffen, R. Debets, J. W. Gratama, G. Stoter, E. Oosterwijk.
2006
. Treatment of metastatic renal cell carcinoma with autologous T-lymphocytes genetically retargeted against carbonic anhydrase IX: first clinical experience.
J. Clin. Oncol.
24
:
e20
-e22.
8
Schumacher, T. N..
2002
. T-cell-receptor gene therapy.
Nat. Rev. Immunol.
2
:
512
-519.
9
Bonini, C., G. Ferrari, S. Verzeletti, P. Servida, E. Zappone, L. Ruggieri, M. Ponzoni, S. Rossini, F. Mavilio, C. Traversari, C. Bordignon.
1997
. HSV-TK gene transfer into donor lymphocytes for control of allogeneic graft-versus-leukemia.
Science
276
:
1719
-1724.
10
Tiberghien, P., C. Ferrand, B. Lioure, N. Milpied, R. Angonin, E. Deconinck, J. M. Certoux, E. Robinet, P. Saas, B. Petracca, et al
2001
. Administration of herpes simplex-thymidine kinase-expressing donor T cells with a T-cell-depleted allogeneic marrow graft.
Blood
97
:
63
-72.
11
Berger, C., M. E. Flowers, E. H. Warren, S. R. Riddell.
2006
. Analysis of transgene-specific immune responses that limit the in vivo persistence of adoptively transferred HSV-TK-modified donor T cells after allogeneic hematopoietic cell transplantation.
Blood
107
:
2294
-2302.
12
Riddell, S. R., M. Elliott, D. A. Lewinsohn, M. J. Gilbert, L. Wilson, S. A. Manley, S. D. Lupton, R. W. Overell, T. C. Reynolds, L. Corey, P. D. Greenberg.
1996
. T-cell mediated rejection of gene-modified HIV-specific cytotoxic T lymphocytes in HIV-infected patients.
Nat. Med.
2
:
216
-223.
13
Introna, M., A. M. Barbui, F. Bambacioni, C. Casati, G. Gaipa, G. Borleri, S. Bernasconi, T. Barbui, J. Golay, A. Biondi, A. Rambaldi.
2000
. Genetic modification of human T cells with CD20: a strategy to purify and lyse transduced cells with anti-CD20 antibodies.
Hum. Gene Ther.
11
:
611
-620.
14
Serafini, M., M. Manganini, G. Borleri, M. Bonamino, L. Imberti, A. Biondi, J. Golay, A. Rambaldi, M. Introna.
2004
. Characterization of CD20-transduced T lymphocytes as an alternative suicide gene therapy approach for the treatment of graft-versus-host disease.
Hum. Gene Ther.
15
:
63
-76.
15
Clackson, T., W. Yang, L. W. Rozamus, M. Hatada, J. F. Amara, C. T. Rollins, L. F. Stevenson, S. R. Magari, S. A. Wood, N. L. Courage, et al
1998
. Redesigning an FKBP-ligand interface to generate chemical dimerizers with novel specificity.
Proc. Natl. Acad. Sci. USA
95
:
10437
-10442.
16
Straathof, K. C., M. A. Pule, P. Yotnda, G. Dotti, E. F. Vanin, M. K. Brenner, H. E. Heslop, D. M. Spencer, C. M. Rooney.
2005
. An inducible caspase 9 safety switch for T-cell therapy.
Blood
105
:
4247
-4254.
17
Kurts, C., J. F. Miller, R. M. Subramaniam, F. R. Carbone, W. R. Heath.
1998
. Major histocompatibility complex class I-restricted cross-presentation is biased towards high dose antigens and those released during cellular destruction.
J. Exp. Med.
188
:
409
-414.
18
Kitamura, T..
1998
. New experimental approaches in retrovirus-mediated expression screening.
Int. J. Hematol.
67
:
351
-359.
19
Kessels, H. W., M. C. Wolkers, M. D. van den Boom, M. A. van der Valk, T. N. Schumacher.
2001
. Immunotherapy through TCR gene transfer.
Nat. Immunol.
2
:
957
-961.
20
Norbury, C. C., M. F. Princiotta, I. Bacik, R. R. Brutkiewicz, P. Wood, T. Elliott, J. R. Bennink, J. W. Yewdell.
2001
. Multiple antigen-specific processing pathways for activating naive CD8+ T cells in vivo.
J. Immunol.
166
:
4355
-4362.
21
Topham, D. J., M. R. Castrucci, F. S. Wingo, G. T. Belz, P. C. Doherty.
2001
. The role of antigen in the localization of naive, acutely activated, and memory CD8+ T cells to the lung during influenza pneumonia.
J. Immunol.
167
:
6983
-6990.
22
Badovinac, V. P., J. S. Haring, J. T. Harty.
2007
. Initial T cell receptor transgenic cell precursor frequency dictates critical aspects of the CD8+ T cell response to infection.
Immunity
26
:
827
-841.
23
Muranski, P., A. Boni, C. Wrzesinski, D. E. Citrin, S. A. Rosenberg, R. Childs, N. P. Restifo.
2006
. Increased intensity lymphodepletion and adoptive immunotherapy–how far can we go?.
Nat. Clin. Pract. Oncol.
3
:
668
-681.
24
Kaech, S. M., R. Ahmed.
2003
. Immunology: CD8 T cells remember with a little help.
Science
300
:
263
-265.
25
Goldrath, A. W., L. Y. Bogatzki, M. J. Bevan.
2000
. Naive T cells transiently acquire a memory-like phenotype during homeostasis-driven proliferation.
J. Exp. Med.
192
:
557
-564.
26
Copelan, E. A..
2006
. Hematopoietic stem-cell transplantation.
N. Engl. J. Med.
354
:
1813
-1826.
27
Bleakley, M., S. R. Riddell.
2004
. Molecules and mechanisms of the graft-versus-leukaemia effect.
Nat. Rev. Cancer
4
:
371
-380.
28
Lamers, C. H., S. Sleijfer, A. G. Vulto, W. H. Kruit, M. Kliffen, R. Debets, J. W. Gratama, G. Stoter, E. Oosterwijk.
2006
. Treatment of metastatic renal cell carcinoma with autologous T-lymphocytes genetically retargeted against carbonic anhydrase IX: first clinical experience.
J. Clin. Oncol.
24
:
e20
-e22.
29
Coccoris, M., M. A. de Witte, T. N. M. Schumacher.
2005
. Prospects and limitations of T cell receptor gene therapy.
Curr. Gene Ther.
5
:
583
-593.
30
Kieback, E., J. Charo, D. Sommermeyer, T. Blankenstein, W. Uckert.
2008
. A safeguard eliminates T cell receptor gene-modified autoreactive T cells after adoptive transfer.
Proc. Natl. Acad. Sci. USA
105
:
623
-628.
31
van Meerten, T., R. S. van Rijn, S. Hol, A. Hagenbeek, S. B. Ebeling.
2006
. Complement-induced cell death by rituximab depends on CD20 expression level and acts complementary to antibody-dependent cellular cytotoxicity.
Clin. Cancer Res.
12
:
4027
-4035.
32
Kurts, C., F. R. Carbone, M. Barnden, E. Blanas, J. Allison, W. R. Heath, J. F. Miller.
1997
. CD4+ T cell help impairs CD8+ T cell deletion induced by cross-presentation of self-antigens and favors autoimmunity.
J. Exp. Med.
186
:
2057
-2062.
33
Kohn, D. B., M. S. Hershfield, D. Carbonaro, A. Shigeoka, J. Brooks, E. M. Smogorzewska, L. W. Barsky, R. Chan, F. Burotto, G. Annett, et al
1998
. T lymphocytes with a normal ADA gene accumulate after transplantation of transduced autologous umbilical cord blood CD34+ cells in ADA-deficient SCID neonates.
Nat. Med.
4
:
775
-780.
34
Quinn, E. R., L. G. Lum, K. T. Trevor.
1998
. T cell activation modulates retrovirus-mediated gene expression.
Hum. Gene Ther.
9
:
1457
-1467.
35
Cooper, L. J., M. S. Topp, C. Pinzon, I. Plavec, M. C. Jensen, S. R. Riddell, P. D. Greenberg.
2004
. Enhanced transgene expression in quiescent and activated human CD8+ T cells.
Hum. Gene Ther.
15
:
648
-658.
36
Verzeletti, S., C. Bonini, S. Marktel, N. Nobili, F. Ciceri, C. Traversari, C. Bordignon.
1998
. Herpes simplex virus thymidine kinase gene transfer for controlled graft-versus-host disease and graft-versus-leukemia: clinical follow-up and improved new vectors.
Hum. Gene Ther.
9
:
2243
-2251.
37
Kimby, E..
2005
. Tolerability and safety of rituximab (MabThera).
Cancer Treat. Rev.
31
:
456
-473.