Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Biol Blood Marrow Transplant. Author manuscript; available in PMC 2007 October 23.
Published in final edited form as:
PMCID: PMC2040267

Inducible caspase 9 suicide gene to improve the safety of allodepleted T cells after haploidentical stem cell transplantation


Addback of donor T cells following T cell-depleted stem cell transplantation (SCT) can accelerate immune reconstitution and be effective against relapsed malignancy. After haploidentical SCT, a high risk of graft-versus-host disease (GVHD) essentially precludes this option, unless the T cells are first depleted of alloreactive precursor cells. Even then, the risks of severe GVHD remain significant. To increase the safety of the approach and thereby permit administration of larger T cell doses, we used a suicide gene, inducible caspase 9 (iCasp9), to transduce allodepleted T cells, permitting their destruction should administration have adverse effects. We made a retroviral vector encoding iCasp9 and a selectable marker (truncated CD19). Even after allodepletion (using anti-CD25 immunotoxin), donor T cells could be efficiently transduced, expanded, and subsequently enriched by CD19 immunomagnetic selection to >90% purity. These engineered cells retained anti-viral specificity and functionality, and contained a subset with regulatory phenotype and function. Activating iCasp9 with a small-molecule dimerizer rapidly produced >90% apoptosis. Although transgene expression was downregulated in quiescent T cells, iCasp9 remained an efficient suicide gene, as expression was rapidly upregulated in activated (alloreactive) T cells. We have demonstrated the clinical feasibility of this approach after haploidentical transplantation by scaling up production using clinical grade materials.


Donor T cell infusion is an effective strategy for conferring anti-viral and anti-tumor immunity following allogeneic stem cell transplantation1-3. This can be particularly useful in T cell depleted transplantation, where immune reconstitution is delayed. In haploidentical transplantation, the need to accelerate immune reconstitution is most pressing; here, profound immune deficiency as a consequence of vigorous T cell depletion and MHC-incompatibility, results in high rates of infectious complications and disease relapse4,5. Unfortunately however, addback of unmanipulated donor T cells is unlikely to be feasible in the haploidentical setting because graft-versus-host disease (GVHD) can occur after addback of as few as 3×104 CD3+ cells /kg6. This problem can be partially overcome by selective depletion of alloreactive cells, for example by using immunotoxins directed to activation markers on alloreactive cells7-9. We, and others, have previously shown that addback of allodepleted T cells at doses between 1 to 8×105 cells /kg is associated with a low incidence of GVHD and significantly accelerates T cell recovery and reconstitutes anti-viral immunity7,8. However, disease relapse remains high in these series, and since the estimated frequency of tumor-reactive precursors is 1 to 2 logs less than frequency of viral-reactive precursors10,11, much greater dose escalation is likely required to reconstitute anti-tumor immunity.

While dose escalation of allodepleted T cells may be desirable, it may not be safe. The risk of GVHD increases with increasing T cell dose12, and the maximum dose that can be safely infused in any given individual cannot be predicted with certainty. Once established, severe GVHD unresponsive to frontline therapy has a poor prognosis. Hence, although severe GVHD occurs infrequently, the fact that it is unpredictable and may be fatal compromises dose intensity in all patients. Suicide gene-modification of T cells circumvents this biological uncertainty: effective T cell doses can be administered to all patients safe in the knowledge that any GVHD that develops can be effectively controlled by activation of the suicide gene mechanism.

One of the most widely used suicide genes is Herpes simplex virus thymidine kinase (HSVtk). This enzyme mediates the conversion of ganciclovir to ganciclovir triphosphate which is toxic to dividing cells; administration of ganciclovir efficiently eliminates HSVtk-modified T cells and abrogates acute GVHD13-15. Although providing proof of concept of suicide gene therapy, HSVtk has a number of drawbacks, the most important of which is immunogenicity: being a foreign protein, HSVtk is a target for CD4 and CD8 T cell-mediated immune response, which results in premature elimination of HSVtk-modified cells16. Other drawbacks of HSVtk include restriction of killing to dividing cells, the unintended elimination of gene-modified cells when ganciclovir is used for treatment of cytomegalovirus (CMV) reactivation, and ganciclovir resistance resulting from truncated HSVtk formed from cryptic splice donor and acceptor sites17.

We investigated the suitability of an alternative suicide gene, inducible caspase 9 (iCasp9)18. iCasp9-mediated suicide is based on conditional dimerization of pro-apoptotic molecules18,19, that are constructed from human proteins and therefore less likely to be immunogenic. The mechanism of killing allows the safe use of ganciclovir, and is independent of cell proliferation.

We now show the feasibility of engineering allodepleted T cells with an iCasp9 suicide gene transfer, and demonstrate the functionality of the modified T cells, and the scalability of the process.


Generation of allodepleted T cells

Allodepleted cells were generated from healthy volunteers as previously described7,20. In brief, peripheral blood mononuclear cells (PBMCs) from healthy donors were co-cultured with irradiated recipient Epstein Barr virus (EBV)-transformed lymphoblastoid cell lines (LCL) at responder-to-stimulator ratio of 40:1 in serum-free medium (AIM V; Invitrogen, Carlsbad, CA). After 72 hours, activated T cells that expressed CD25 were depleted from the co-culture by overnight incubation in RFT5-SMPT-dgA immunotoxin21. Allodepletion was considered adequate if the residual CD3+CD25+ population was <1% and residual proliferation by 3H-thymidine incorporation was <10%7.

Plasmid and retrovirus

SFG.iCasp9.2A.ΔCD19 consists of inducible caspase 9 (iCasp9) linked, via a cleavable 2A-like sequence, to truncated human CD19 (ΔCD19)(Fig. 1A). iCasp9 consists of a human FK506-binding protein (FKBP12; GenBank AH002 818) with an F36V mutation, connected via a Ser-Gly-Gly-Gly-Ser linker to human caspase 9 (CASP9; GenBank NM 001229)18. The F36V mutation increases the binding affinity of FKBP12 to the synthetic homodimerizer, AP20187 or AP190322. The caspase recruitment domain (CARD) has been deleted from the human caspase 9 sequence because its physiological function has been replaced by FKBP12, and its removal increases transgene expression and function18. The 2A-like sequence encodes an 20 amino acid peptide from Thosea Asigna insect virus, which mediates >99% cleavage between a glycine and terminal proline residue23, resulting in 19 extra amino acids in the C terminus of iCasp9, and one extra proline residue in the N terminus of CD19. ΔCD19 consists of full-length CD19 (GenBank NM 001770) truncated at amino acid 333 (TDPTRRF), which shortens the intracytoplasmic domain from 242 to 19 amino acids, and removes all conserved tyrosine residues that are potential sites for phosphorylation24,25.

Figure 1
Generation of iCasp9 suicide gene-modified selectively allodepleted cells

A stable PG13 clone producing Gibbon ape leukemia virus (Gal-V) pseudotyped retrovirus was made by transiently transfecting Phoenix Eco cell line (ATCC product #SD3444; ATCC, Manassas, VA) with SFG.iCasp9.2A.ΔCD19. This produced Eco-pseudotyped retrovirus. The PG13 packaging cell line (ATCC) was transduced three times with Eco-pseudotyped retrovirus to generate a producer line that contained multiple SFG.iCasp9.2A.ΔCD19 proviral integrants per cell. Single cell cloning was performed, and the PG13 clone that produced the highest titer was expanded and used for vector production.

Retroviral transduction

Culture medium for T cell activation and expansion consisted of 45% RPMI 1640 (Hyclone, Logan, UT), 45% Clicks (Irvine Scientific, Santa Ana, CA) and 10% fetal bovine serum (FBS; Hyclone). Allodepleted cells were activated by immobilized anti-CD3 (OKT3; Ortho Biotech, Bridgewater, NJ) for 48 hours before transduction with retroviral vector (Fig. 1B). In small-scale experiments, non-tissue culture-treated 24-well plates (Becton Dickinson, San Jose, CA) were coated with OKT3 1μg/ml for 2 to 4 hours at 37°C. Allodepleted cells were added at 1×106 cells per well. At 24 hours, 100U/ml of recombinant human interleukin-2 (IL-2)(Proleukin; Chiron, Emeryville, CA) was added. Retroviral transduction was performed 48 hours after activation. Non-tissue culture-treated 24-well plates were coated with 3.5μg/cm2 recombinant fibronectin fragment (CH-296;Retronectin; Takara Mirus Bio, Madison, WI) and the wells loaded twice with retroviral vector-containing supernatant at 0.5ml per well for 30 minutes at 37°C, following which OKT3-activated cells were plated at 5×105 cells per well in fresh retroviral vector-containing supernatant and T cell culture medium at a ratio of 3:1, supplemented with 100U/ml IL-2. Cells were harvested after 2 to 3 days and expanded in the presence of 50U/ml IL-2.

Scaling-up production of gene-modified allodepleted cells

Scale-up of the transduction process for clinical application used non-tissue culture-treated T75 flasks (Nunc, Rochester, NY), which were coated with 10ml of OKT3 1μg/ml or 10ml of fibronectin 7μg/ml at 4°C overnight. We also used fluorinated ethylene propylene bags corona-treated for increased cell adherence (2PF-0072AC, American Fluoroseal Corporation, Gaithersburg, MD). Allodepleted cells were seeded in OKT3-coated flasks at 1×106 cells/ml. 100U/ml IL-2 was added the next day. For retroviral transduction, retronectin-coated flasks or bags were loaded once with 10ml of retrovirus-containing supernatant for 2 to 3 hours. OKT3-activated T cells were seeded at 1×106 cells/ml in fresh retroviral vector-containing medium and T cell culture medium at a ratio of 3:1, supplemented with 100U/ml IL-2. Cells were harvested the following morning and expanded in tissue-culture treated T75 or T175 flasks in culture medium supplemented with 50 to 100U/ml IL-2 at a seeding density of 5 to 8×105 cells/ml.

CD19 immunomagnetic selection

Immunomagnetic selection for CD19 was performed 4 days after transduction. Cells were labeled with paramagnetic microbeads conjugated to monoclonal mouse anti-human CD19 antibodies (Miltenyi Biotech, Auburn, CA) and selected on MS or LS columns in small scale experiments and on a CliniMacs Plus automated selection device in large scale experiments. CD19-selected cells were expanded for a further 4 days and cryopreserved on day 8 post transduction. These cells were referred to as “gene-modified allodepleted cells”.

Immunophenotyping and pentamer analysis

Flow cytometric analysis (FACSCalibur and CellQuest software; Becton Dickinson) used the following antibodies: CD3, CD4, CD8, CD19, CD25, CD27, CD28, CD45RA, CD45RO, CD56 and CD62L. CD19-PE (Clone 4G7; Becton Dickinson) was found to give optimum staining and was used in all analysis. Non-transduced control was used to set the negative gate for CD19. An HLA-pentamer, HLA-B8-RAKFKQLL (Proimmune, Springfield, VA) was used to detect T cells recognizing an epitope from EBV lytic antigen (BZLF1). HLA-A2-NLVPMVATV pentamer was used to detect T cells recognizing an epitope from CMV-pp65 antigen.

Interferon-γ ELISPOT assay for anti-viral response

Interferon-γ ELISPOT for assessment of responses to EBV, CMV and adenovirus antigens was performed as previously described20. Gene-modified allodepleted cells cryopreserved at 8 days post transduction were thawed and rested overnight in complete medium without IL-2 prior to use as responder cells. Cryopreserved PBMCs from the same donor were used as comparators. Responder cells were plated in duplicate or triplicate in serial dilutions of 2×105, 1×105, 5×104 and 2.5×104 cells per well. Stimulator cells were plated at 1×105 per well. For response to EBV, donor-derived EBV-LCLs irradiated at 40Gy were used as stimulators. For response to adenovirus, donor-derived activated monocytes infected with Ad5f35 adenovirus were used. In brief, donor PBMCs were plated in X-Vivo 15 (Cambrex, Walkersville, MD) in 24-well plates overnight, harvested the next morning, infected with Ad5f35 at MOI of 200 for 2 hours, washed, irradiated at 30Gy, and used as stimulators. For anti-CMV response, a similar process using Ad5f35 adenovirus encoding the CMV pp65 transgene (Ad5f35-pp65) at an MOI 5000 was used. Specific spot-forming units (SFU) were calculated by subtracting SFU from responder-alone and stimulator-alone wells from test wells. Response to CMV was the difference in SFU between Ad5f35-pp65 and Ad5f35 wells.

EBV-specific cytotoxicity

Gene-modified allodepleted cells were stimulated with 40Gy-irradiated donor-derived EBV-LCL at a responder:stimulator ratio of 40:1. After 9 days, the cultures were restimulated at a responder:stimulator ratio of 4:1. Restimulation was performed weekly as indicated. After two or three rounds of stimulation, cytotoxicity was measured in a 4-hour 51Cr-release assay, using donor EBV-LCL as target cells and donor OKT3 blasts as autologous controls. NK activity was inhibited by adding 30-fold excess of cold K562 cells.

Induction of apoptosis with chemical inducer of dimerization, AP20187

Suicide gene functionality was assessed by adding a small molecule synthetic homodimerizer, AP20187 (Ariad Pharmaceuticals; Cambridge, MA), at 10nM final concentration the day following CD19 immunomagnetic selection. Cells were stained with annexin V and 7-amino-actinomycin (7-AAD)(BD Pharmingen) at 24 hours and analyzed by flow cytometry. Cells negative for both Annexin V and 7-AAD were considered viable, cells that were annexin V positive were apoptotic, and cells that were both annexin V and 7-AAD positive were necrotic. The percentage killing induced by dimerization was corrected for baseline viability as follows: Percentage killing = 100% - (%Viability in AP20187-treated cells ÷ %Viability in non-treated cells).

Assessment of transgene expression following extended culture and reactivation

Cells were maintained in T cell medium containing 50U/ml IL-2 until 22 days after transduction. A portion of cells was reactivated on 24-well plates coated with 1μg/ml OKT3 and 1μg/ml anti-CD28 (Clone CD28.2, BD Pharmingen, San Jose, CA) for 48 to 72 hours. CD19 expression and suicide gene function in both reactivated and non-reactivated cells were measured on day 24 or 25 post transduction.

In some experiments we also cultured the cells for 3 weeks post transduction but simulated them with 30G-irradiated allogeneic PBMC at responder:stimulator ratio of 1:1. After 4 days of co-culture, a portion of cells was treated with 10nM AP20187. Killing was measured by annexin V/7-AAD staining at 24 hours, and the effect of dimerizer on bystander virus-specific T cells was assessed by pentamer analysis on AP20187-treated and untreated cells.

Regulatory T cells

We analyzed CD4, CD25 and Foxp3 expression in gene-modified allodepleted cells using flow cytometry. For human Foxp3 staining we used the eBioscience (San Diego, CA) staining set with an appropriate rat IgG2a isotype control. These cells were co-stained with surface CD25-FITC and CD4-PE. Functional analysis was performed by co-culturing CD4+25+ cells selected after allodepletion and gene modification with carboxyfluorescein diacetate N-succinimidyl ester (CFSE)-labeled autologous PBMC. CD4+25+ selection was performed by first depleting CD8+ cells using anti-CD8 microbeads (Miltenyi Biotec, Auburn, CA), followed by positive selection using anti-CD25 microbeads (Miltenyi Biotec, Auburn, CA). CFSE-labeling was performed by incubating autologous PBMC at 2×107/ml in phosphate buffered saline containing 1.5μM CFSE for 10 minutes. The reaction was stopped by adding an equivalent volume of FBS and incubating for 10 minutes at 37°C. Cells were washed twice before use. We stimulated CFSE-labeled PBMCs with OKT3 500ng/ml and 40G-irradiated allogeneic PBMC feeders at PBMC: allogeneic feeders ratio of 5:1. We cultured the cells with or without an equal number of autologous CD4+25+ gene-modified allodepleted cells. After 5 days of culture, cell division was analyzed by flow cytometry; CD19 was used to gate out non-CFSE-labeled CD4+CD25+ gene-modified T cells.

Statistical analysis

Paired, 2-tailed Student’s t test was used to determine the statistical significance of differences between samples. All data are represented as mean ± 1 standard deviation.


Selectively allodepleted T cells could be efficiently transduced with iCasp9 and expanded

Selective allodepletion was performed in accordance with procedures established for our recently completed clinical protocol7, by co-culturing 3/6 to 5/6 HLA-mismatched PBMC and lymphoblastoid cell lines (LCL). RFT5-SMPT-dgA immunotoxin was applied after 72 hours of co-culture and reliably produced allodepleted cells with <10% residual proliferation (mean 4.5 ± 2.8%; range 0.74 to 9.1%; 10 experiments) and containing <1% residual CD3+CD25+ cells (mean 0.23 ± 0.20%; range 0.06 to 0.73%; 10 experiments), thereby fulfilling the release criteria for selective allodepletion, and serving as starting materials for subsequent manipulation.

Allodepleted cells activated on immobilized OKT3 for 48 hours could be efficiently transduced with Gal-V pseudotyped retrovirus vector encoding SFG.iCasp9.2A.ΔCD19. Transduction efficiency assessed by FACS analysis for CD19 expression 2 to 4 days after transduction was 53% ± 8%, with comparable results for small-scale (24-well plates) and large-scale (T75 flasks) transduction (55 ± 8% versus 50% ± 10% in 6 and 4 experiments, respectively). Cell numbers contracted in the first 2 days following OKT3 activation such that only 61 ± 12% (range 45 to 80%) of allodepleted cells were recovered on the day of transduction (Fig. 2). Thereafter, the cells showed significant expansion, with a mean 94 ± 46 -fold expansion (range 40 to 153) over the subsequent 8 days, resulting in a net 58 ± 33 -fold expansion. Cell expansion in both small- and large-scale experiments was similar, with net expansion of 45 ± 29 fold (range 25 to 90) in 5 small-scale experiments and 79 ± 34 fold (range 50 to 116) in 3 large-scale experiments.

Figure 2
Allodepleted cells could be successfully expanded following transduction

ΔCD19 enables efficient and selective enrichment of transduced cells on immunomagnetic columns

The efficiency of suicide gene activation will depend not only on the functionality of the suicide gene itself, but also on the selection system used to enrich for gene-modified cells. Due to limitations of selectable markers currently in use, we investigated whether ΔCD19 enabled the selection of gene-modified cells with sufficient purity and yield, and whether selection had any deleterious effects on subsequent cell growth. Small-scale selection was performed according to manufacturer’s instruction; however, we found that large-scale selection was optimum when we used 10μl of CD19 microbeads per 1.3×107 cells. FACS analysis was performed at 24 hours after immunomagnetic selection to minimize interference from anti-CD19 microbeads. The purity of the cells after immunomagnetic selection was consistently greater than 90%: mean percentage of CD19+ cells was 98.3 ± 0.5% (n=5) in small-scale selections and 97.4 ± 0.9% (n=3) in large-scale CliniMacs selections (Fig.3). The absolute yield of small- and large-scale selections were 31 ± 11 % and 28 ± 6% respectively; after correction for transduction efficiency, the mean recovery of transduced cells was 54 ± 14 % in small-scale and 72 ± 18 % in large-scale selections. The selection process did not have any discernable deleterious effect on subsequent cell expansion: in 4 experiments, the mean cell expansion over 3 days following CD19 immunomagnetic selection was 3.5 fold for the CD19 positive fraction versus 4.1 fold for non-selected transduced cells (p=0.34) and 3.7 fold for non-transduced cells (p=0.75).

Figure 3
Suicide gene-modified cells could be enriched to high purity by CD19 immunomagnetic selection

Immunophenotype of gene-modified allodepleted cells

We immunophenotyped the final cell product (gene-modified allodepleted cells that had been cryopreserved 8 days after transduction). This final cell product contained both CD4 and CD8 cells, with CD8 cells predominant (Table 1) at 62% ±11% CD8+ versus 23% ± 8% CD4+. The majority of cells were CD45RO+ and had the surface immunophenotype of effector memory T cells. Expression of memory markers, including CD62L, CD27 and CD28, was heterogeneous. Approximately 24% of cells expressed CD62L, a lymph node-homing molecule predominantly expressed on central memory cells.

Table 1
Comparison of immunophenotype of gene-modified allodepleted cells and unmanipulated PBMCa

Gene-modified allodepleted cells retained antiviral repertoire and functionality

The ability of end-product cells to mediate antiviral immunity was assessed by interferon-γ ELISPOT, cytotoxicity assay, and pentamer analysis. These cryopreserved gene-modified allodepleted cells were used in all analyses, since they were representative of the product we propose to administer in the clinical study. Interferon-γ secretion in response to adenovirus, CMV or EBV antigens presented by donor cells was preserved although there was a trend towards reduced anti-EBV response in gene-modified allodepleted cells versus unmanipulated PBMC (Fig. 4A).

Figure 4Figure 4
Gene-modified allodepleted cells retained anti-viral repertoire and functionality

Cytotoxicity was assessed using donor-derived EBV-LCL as targets. Gene-modified allodepleted cells that had undergone 2 or 3 rounds of stimulation with donor-derived EBV-LCL could efficiently lyse virus-infected autologous target cells (Fig. 4B). Since EBV-LCLs were used as antigen-presenting cells during selective allodepletion, it was possible that EBV-specific T cells could be significantly depleted when the donor and recipient were haploidentical. We therefore included three experiments using unrelated HLA-haploidentical donor-recipient pairs, and showed that cytotoxicity against donor-derived EBV-LCL was retained. The results were corroborated by pentamer analysis for T cells recognizing HLA-B8-RAKFKQLL, an EBV lytic antigen (BZLF1) epitope, in two informative donors following allodepletion against HLA-B8 negative haploidentical recipients (Fig 4C). The RAK-pentamer positive population was retained and could expand following stimulation with donor EBV-LCL. Together, these results provided evidence that gene-modified allodepleted cells retained significant anti-viral functionality.

Regulatory T cells in the Gene-modified allodepleted cell population

We used flow cytometry and functional analysis to discover whether regulatory T cells was retained in our allodepleted, gene modified, T cell product. We found a Foxp3+ CD4+25+ population (Fig. 5A), and following immunomagnetic separation, the CD4+CD25+ enriched fraction demonstrated suppressor function when co-cultured with CFSE-labeled autologous PBMC in the presence of OKT3 and allogeneic feeders (Fig. 5B). Hence, allodepleted T cells may re- acquire regulatory phenotype even after exposure to a CD25 depleting immunotoxin.

Figure 5
Regulatory T cells could be isolated from gene-modified end-product cells despite initial allodepletion using CD25-immunotoxin

Gene-modified allodepleted cells were efficiently and rapidly eliminated by addition of chemical inducer of dimerization

The day following immunomagnetic selection, 10nM of the chemical inducer of dimerization, AP20187, was added to induce apoptosis, which appeared within 24 hours. FACS analysis with annexin V and 7-AAD staining at 24 hours showed that only 5.5 ± 2.5% of AP20187-treated cells remained viable, whereas 81.0 ± 9.0 % of untreated cells were viable (Fig.6A). Killing efficiency after correction for baseline viability was 92.9 ± 3.8%. Large-scale CD19 selection produced cells that were killed with similar efficiency as small-scale selection: mean viability with and without AP20187, and percentage killing, in large and small scale were 3.9%, 84.0%, 95.4% (n=3) and 6.6%, 79.3%, 91.4% (n=5) respectively. AP20187 was non-toxic to non-transduced cells: viability with and without AP20187 were 86 ± 9% and 87 ± 8% respectively (n=6).

Figure 6Figure 6
Gene-modified allodepleted cells were rapidly and efficiently eliminated by AP20187. Transgene expression and killing efficiency diminished with extended culture but were completely restored upon T cell reactivation

Transgene expression and function decreased with extended culture but were restored upon cell reactivation

To assess the stability of transgene expression and function, we maintained cells in T cell culture medium and low dose IL-2 (50U/ml) until 24 days after transduction. By day 24, surface CD19 expression fell from 98 ± 1% to 88 ± 4% (p<0.05) with a parallel decrease in mean fluorescence intensity (MFI) from 793 ± 128 to 478 ± 107 (p<0.05) (Fig.6B). Similarly, there was a significant reduction in suicide gene function: residual viability was 19.6 ± 5.6% following treatment with AP20187; after correction for baseline viability of 54.8 ± 20.9%, this equated to killing efficiency of only 63.1 ± 6.2%.

To discover whether the decrease in transgene expression with time was due to reduced transcription following T cell quiescence or to elimination of transduced cells, we reactivated a portion of cells on day 22 post transduction with OKT3 and anti-CD28 antibody. At 48 to 72 hours (day 24 or 25 post transduction), OKT3/aCD28-reactivated cells had significantly higher transgene expression than non-reactivated cells. CD19 expression increased from 88 ± 4% to 93 ± 4% (p <0.01) and CD19 MFI increased from 478 ± 107 to 643 ± 174 (p<0.01). More importantly, suicide gene function also increased significantly from 63.1 ± 6.2% killing efficiency to 84.6 ± 8.0% (p<0.01). Furthermore, killing efficiency was completely restored if the cells were immunomagnetically sorted for the activation marker CD25: killing efficiency of CD25 positive cells was 93.2 ± 1.2%, which was the same as killing efficiency on day 5 post transduction (93.1 ± 3.5%)(Fig. 6C). Killing of the CD25 negative fraction was 78.6 ± 9.1%.

Of note, virus-specific T cells are spared when dimerizer is used to deplete gene-modified cells that have been re-activated with allogeneic PBMC, rather than by non-specific mitogenic stimuli. After 4 days reactivation with allogeneic cells, Figs 7A and B show that treatment with AP20187 spares (and thereby enriches) viral reactive subpopulations, as measured by the proportion of T cells reactive with HLA pentamers specific for peptides derived from EBV and CMV.

Figure 7
Viral-specific T cells were partially retained after treatment of allostimulated cells with dimerizer


We have demonstrated the feasibility of engineering allogeneic T cells with two distinct safety mechanisms, selective allodepletion and suicide gene-modification. In combination, these modifications should enable addback of substantial numbers of T cells with anti-viral and anti-tumor activity, even after haploidentical transplantation. Our data show that the suicide gene, iCasp9, functions efficiently (>90% apoptosis after treatment with dimerizer) and that down-modulation of transgene expression that occurred with time was rapidly reversed upon T cell activation, as would occur when alloreactive T cells encountered their targets. We also showed that ΔCD19 was a suitable selectable marker that enabled efficient and selective enrichment of transduced cells to >90% purity, and that these manipulations had no discernable effects on the immunological competence of the engineered T cells with retention of antiviral activity, and regeneration of a CD4+CD25+Foxp3+ population with Treg activity.

Given that the overall functionality of suicide genes depends on both the suicide gene itself and the marker used to select the transduced cells, translation into clinical use requires optimization of both components, and of the method used to couple expression of the two genes. The two most widely used selectable markers in clinical practice each have drawbacks. Neomycin phosphotransferase (neo) encodes a potentially immunogenic foreign protein and requires a 7-day culture in selection medium, which not only increases the complexity of the system, but is also potentially damaging to virus-specific T cells26. The most widely used surface selection marker is ΔLNGFR, but recent report of leukemia in a mouse model raises concerns regarding its oncogenic potential27, despite its apparent clinical safety 28. Furthermore, ΔLNGFR selection is not widely available, because it is used almost exclusively in gene therapy. A number of alternative selectable markers have been suggested. CD34 has been well-studied in vitro29,30, but the steps required to optimize a system designed primarily for selection of rare hematopoietic progenitors, and more critically, the potential for altered in vivo T cell homing31, make CD34 sub-optimal for our purpose. As an alternative marker, we used ΔCD19, since clinical grade CD19 selection is readily available as a method for B-cell depletion of stem cell autografts. We demonstrated that ΔCD19 enrichment could be performed with high purity and yield and, more importantly, the selection process had no discernable effect on subsequent cell growth and functionality.

The effectiveness of suicide gene activation in ΔCD19-selected iCasp9 cells compared very favorably to that of neo- or ΔLNGFR-selected cells transduced to express the HSVtk gene. The earlier generations of HSVtk constructs provided 80-90% suppression of 3H-thymidine uptake14,32,33 and showed similar reduction in killing efficiency upon extended in vitro culture33, but were nonetheless clinically efficacious. Complete resolution of both acute and chronic GVHD has been reported with as little as 80% in vivo reduction in circulating gene-modified cells. These data supported the hypothesis that transgene down-modulation seen in vitro is unlikely to be an issue because activated T cells responsible for GVHD will upregulate suicide gene expression and will therefore be selectively eliminated in vivo14. Whether this effect is sufficient to allow retention of virus- and leukemia-specific T cells in vivo remains to be tested in a clinical setting. By combining in vitro selective allodepletion prior to suicide gene modification, we hope that the need to activate the suicide gene mechanism will be significantly reduced, thereby maximizing the benefits of addback T cell.

The high efficiency of iCasp9-mediated suicide seen in vitro has been replicated in vivo. In a SCID mouse-human xenograft model, more than 99% of iCasp9-modified T cells were eliminated after a single dose of dimerizer18. Safety testing has been performed on healthy human volunteers using AP1903, which has extremely close functional and chemical equivalence to AP20187, and would be used in a proposed clinical application. Maximal plasma level of 10 to 1275ng/ml AP1903 (equivalent to 7 to 892nM) was attained over a 0.01mg/kg to 1.0mg/kg dose range administered as a 2-hour intravenous infusion34. There were no significant adverse effects. After allowing for rapid plasma redistribution, the concentration of dimerizer used in vitro remains readily achievable in vivo.

We believe iCasp9 will prove to have several advantages over HSVtk, including the ability to use ganciclovir and related drugs without causing inadvertent T cell destruction, and a reduced risk of destructive immune responses against transduced cells since the iCasp9 is almost entirely human-derived. Although both our selectable marker (ΔCD19) and our suicide gene (iCasp9) are of human origin, one potential limitation of our approach is that linkage of expression of the selectable marker to the suicide gene uses a 2A-like cleavable peptide which is of non-human origin. However, this product is only 20 amino acids long, and is therefore less likely to be immunogenic than entirely non-human proteins such as neo or HSVtk.

It is essential to define the culture conditions that are optimal for maintaining the immunological competence of suicide gene-modified T cells, since phenotype, repertoire and functionality can all be affected by the stimulation used for polyclonal T cell activation35,36, the method for selection of transduced cells26, and duration of culture37. The addition of CD28 co-stimulation and the use of cell-sized paramagnetic beads to generate gene modified-cells that more closely resemble unmanipulated PBMC in terms of CD4:CD8 ratio, and expression of memory subset markers including lymph node homing molecules CD62L and CCR7, have been proposed as methods for improving the in vivo functionality of gene-modified T cells35,36. CD28 co-stimulation has also been reported to increase the efficiency of retroviral transduction and expansion35,36. Interestingly however, we found that the addition of CD28 co-stimulation had no impact on transduction of allodepleted cells (data not shown), and the degree of cell expansion seen in our study was higher compared to the anti-CD3 alone arm in these studies. Furthermore, iCasp9-modified allodepleted cells retained significant anti-viral functionality, and approximately one fourth retained CD62L expression. There was also regeneration of CD4+CD25+Foxp3+ regulatory T cells. The allodepleted cells we used as starting material for T cell activation and transduction may have been less sensitive to the addition of anti-CD28 antibody as co-stimulation. CD25-depleted PBMC / EBV-LCL co-cultures contained T cells and B cells that already express CD86 at significantly higher level than unmanipulated PBMC (data not shown) and may themselves provide co-stimulation. In addition, it has been reported that depletion of CD25+ regulatory T cells prior to polyclonal T cell activation with anti-CD3 enhances the immunological competence of the final T cell product37. In order to minimize the effect of in vitro culture and expansion on functional competence, we used a relatively brief culture period whereby cells were expanded for a total of 8 days post-transduction with CD19-selection being performed on day 4.

Finally, we have demonstrated that we can scale up production sufficiently to treat adult patients at doses of up to 107 cells/kg: allodepleted cells can be activated and transduced at 4×107cells per flask, and a minimum of 8-fold return of CD19-selected final cell product can be obtained on day 8 post-transduction, to produce at least 3×108 allodepleted gene-modified cells per original flask. The increased culture volume is readily handled in additional flasks or bags.

The allodepletion and iCasp9-modification we describe should significantly improve the safety of adding back T cells, particularly after haploidentical stem cell allografts. This should in turn enable greater dose-escalation, with a higher chance of producing a hoped for anti-leukemia effect. We will explore this possibility in a clinical study of our modified allodepleted T cells.


This work was supported in part by Specialized Centers for Cell-based Therapy grant NIH-NHLBI 1 U54 HL081007, and a Doris Duke Distinguished Clinical Scientist Award (H.E.H.).


Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.


1. Kolb H, Schattenberg A, Goldman J, et al. Graft-versus-leukemia effect of donor lymphocyte transfusions in marrow grafted patients. European Group for Blood and Marrow Transplantation Working Party Chronic Leukemia. Blood. 1995;86:2041–2050. [PubMed]
2. Papadopoulos EB, Ladanyi M, Emanuel D, et al. Infusions of Donor Leukocytes to Treat Epstein-Barr Virus-Associated Lymphoproliferative Disorders after Allogeneic Bone Marrow Transplantation. N Engl J Med. 1994;330:1185–1191. [PubMed]
3. Drobyski WR, Hessner MJ, Klein JP, et al. T-cell depletion plus salvage immunotherapy with donor leukocyte infusions as a strategy to treat chronic-phase chronic myelogenous leukemia patients undergoing HLA-identical sibling marrow transplantation. Blood. 1999;94:434–441. [PubMed]
4. Aversa F, Terenzi A, Tabilio A, et al. Full haplotype-mismatched hematopoietic stem-cell transplantation: a phase II study in patients with acute leukemia at high risk of relapse. J Clin Oncol. 2005;23:3447–3454. [PubMed]
5. Lang P, Greil J, Bader P, et al. Long-term outcome after haploidentical stem cell transplantation in children. Blood Cells, Molecules, and Diseases. 2004;33:281–287. [PubMed]
6. Lewalle P, Triffet A, Delforge A, et al. Donor lymphocyte infusions in adult haploidentical transplant: a dose finding study. Bone Marrow Transplant. 2003;31:39–44. [PubMed]
7. Amrolia PJ, Muccioli-Casadei G, Huls H, et al. Adoptive immunotherapy with allodepleted donor T-cells improves immune reconstitution after haploidentical stem cell transplantation. Blood. 2006;108:1797–1808. [PubMed]
8. Andre-Schmutz I, Le Deist F, Hacein-Bey-Abina S, et al. Immune reconstitution without graft-versus-host disease after haemopoietic stem-cell transplantation: a phase 1/2 study. Lancet. 2002;360:130–137. [PubMed]
9. Solomon SR, Mielke S, Savani BN, et al. Selective depletion of alloreactive donor lymphocytes: a novel method to reduce the severity of graft-versus-host disease in older patients undergoing matched sibling donor stem cell transplantation. Blood. 2005;106:1123–1129. [PubMed]
10. Molldrem JJ, Lee PP, Wang C, Champlin RE, Davis MM. A PR1-Human Leukocyte Antigen-A2 Tetramer Can Be Used to Isolate Low-Frequency Cytotoxic T Lymphocytes from Healthy Donors That Selectively Lyse Chronic Myelogenous Leukemia. Cancer Res. 1999;59:2675–2681. [PubMed]
11. Rezvani K, Grube M, Brenchley JM, et al. Functional leukemia-associated antigen-specific memory CD8+ T cells exist in healthy individuals and in patients with chronic myelogenous leukemia before and after stem cell transplantation. Blood. 2003;102:2892–2900. [PubMed]
12. Guglielmi C, Arcese W, Dazzi F, et al. Donor lymphocyte infusion for relapsed chronic myelogenous leukemia: prognostic relevance of the initial cell dose. Blood. 2002;100:397–405. [PubMed]
13. Bonini C, Ferrari G, Verzeletti S, et al. HSV-TK gene transfer into donor lymphocytes for control of allogeneic graft-versus-leukemia. Science. 1997;276:1719–1724. [PubMed]
14. Tiberghien P, Ferrand C, Lioure B, et al. Administration of herpes simplex-thymidine kinase-expressing donor T cells with a T-cell-depleted allogeneic marrow graft. Blood. 2001;97:63–72. [PubMed]
15. Bonini C, Ciceri F, Stanghellini MTL, et al. Rapid and Wide Immunereconstitution Obtained with HSV-TK Engineered Donor Lymphocyte Add-Backs Permits Long-Term Survival after haplo-HSCT. ASH Annual Meeting Abstracts. 2006;108:307.
16. Berger C, Flowers ME, Warren EH, Riddell SR. 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. 2006;107:2294–2302. [PubMed]
17. Garin MI, Garrett E, Tiberghien P, et al. Molecular mechanism for ganciclovir resistance in human T lymphocytes transduced with retroviral vectors carrying the herpes simplex virus thymidine kinase gene. Blood. 2001;97:122–129. [PubMed]
18. Straathof KC, Pule MA, Yotnda P, et al. An inducible caspase 9 safety switch for T-cell therapy. Blood. 2005;105:4247–4254. [PubMed]
19. Thomis DC, Marktel S, Bonini C, et al. A Fas-based suicide switch in human T cells for the treatment of graft-versus-host disease. Blood. 2001;97:1249–1257. [PubMed]
20. Amrolia PJ, Muccioli-Casadei G, Yvon E, et al. Selective depletion of donor alloreactive T cells without loss of antiviral or antileukemic responses. Blood. 2003;102:2292–2299. [PubMed]
21. Ghetie V, Thorpe P, Ghetie MA, Knowles P, Uhr JW, Vitetta ES. The GLP large scale preparation of immunotoxins containing deglycosylated ricin A chain and a hindered disulfide bond. J Immunol Methods. 1991;142:223–230. [PubMed]
22. Clackson T, Yang W, Rozamus LW, et al. Redesigning an FKBP-ligand interface to generate chemical dimerizers with novel specificity. Proc Natl Acad Sci U S A. 1998;95:10437–10442. [PubMed]
23. Donnelly MLL, Hughes LE, Luke G, et al. The ‘cleavage’ activities of foot-and-mouth disease virus 2A site-directed mutants and naturally occurring ‘2A-like’ sequences. J Gen Virol. 2001;82:1027–1041. [PubMed]
24. Zhou LJ, Ord DC, Hughes AL, Tedder TF. Structure and domain organization of the CD19 antigen of human, mouse, and guinea pig B lymphocytes. Conservation of the extensive cytoplasmic domain. J Immunol. 1991;147:1424–1432. [PubMed]
25. Fujimoto M, Poe JC, Inaoki M, Tedder TF. CD19 regulates B lymphocyte responses to transmembrane signals. Semin Immunol. 1998;10:267–277. [PubMed]
26. Sauce D, Bodinier M, Garin M, et al. Retrovirus-mediated gene transfer in primary T lymphocytes impairs their anti-Epstein-Barr virus potential through both culture-dependent and selection process-dependent mechanisms. Blood. 2002;99:1165–1173. [PubMed]
27. Li Z, Dullmann J, Schiedlmeier B, et al. Murine leukemia induced by retroviral gene marking. Science. 2002;296:497. [PubMed]
28. Bonini C, Grez M, Traversari C, et al. Safety of retroviral gene marking with a truncated NGF receptor. Nat Med. 2003;9:367–369. [PubMed]
29. Rettig MP, Ritchey JK, Meyerrose TE, Haug JS, DiPersio JF. Transduction and selection of human T cells with novel CD34/thymidine kinase chimeric suicide genes for the treatment of graft-versus-host disease. Mol Ther. 2003;8:29–41. [PubMed]
30. Fehse B, Richters A, Putimtseva-Scharf K, et al. CD34 splice variant: an attractive marker for selection of gene-modified cells. Mol Ther. 2000;1:448–456. [PubMed]
31. Lanza F, Healy L, Sutherland DR. Structural and functional features of the CD34 antigen: an update. J Biol Regul Homeost Agents. 2001;15:1–13. [PubMed]
32. Tiberghien P, Reynolds C, Keller J, et al. Ganciclovir treatment of herpes simplex thymidine kinase-transduced primary T lymphocytes: an approach for specific in vivo donor T-cell depletion after bone marrow transplantation? Blood. 1994;84:1333–1341. [PubMed]
33. Verzeletti S, Bonini C, Marktel S, et al. 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. 1998;9:2243–2251. [PubMed]
34. Iuliucci J, Oliver S, Morley S, et al. Intravenous safety and pharmacokinetics of a novel dimerizer drug, AP1903, in healthy volunteers. J Clin Pharmacol. 2001;41:870–879. [PubMed]
35. Bondanza A, Valtolina V, Magnani Z, et al. Suicide gene therapy of graft-versus-host disease induced by central memory human T lymphocytes. Blood. 2006;107:1828–1836. [PubMed]
36. Berger C, Blau CA, Clackson T, Riddell SR, Heimfeld S. CD28 costimulation and immunoaffinity-based selection efficiently generate primary gene-modified T cells for adoptive immunotherapy. Blood. 2003;101:476–484. [PubMed]
37. Mesel-Lemoine M, Cherai M, Le Gouvello S, et al. Initial depletion of regulatory T cells: the missing solution to preserve the immune functions of T lymphocytes designed for cell therapy. Blood. 2006;107:381–388. [PubMed]