Rebuilding Human Leukocyte Antigen Class II–Restricted Minor Histocompatibility Antigen Specificity in Recall Antigen-Specific T Cells by Adoptive T Cell Receptor Transfer: Implications for Adoptive Immunotherapy Free

Minor histocompatibility antigens (mHag) are MHC-bound polymorphic peptides derived from intracellular proteins. With these properties, mHags constitute the main targets of donor T cells mediating the curative graft-versus-tumor (GvT) effect and graft-versus-host disease (GvHD) after human leukocyte antigen (HLA)-matched allogeneic stem cell (SC) transplantation (allo-SCT; ref. 1). From the clinical point of view, a major breakthrough of the past decade was the discovery of mHags expressed exclusively on hematopoietic cells (2–4). In several in vitro studies, cytotoxic T cells (CTL) directed against hematopoietic mHags effectively eliminated mHag⁺ leukemia and myeloma cells, but spared nonhematopoietic cells (5–8). These studies suggested the possibility to separate GvT from GvHD by treatment of recipients with donor T cells directed at hematopoietic mHags. Subsequently, we and others have shown the feasibility of this concept by ex vivo generation of mHag-specific CTLs from unprimed SC donors using dendritic cells (DC) pulsed with mHag peptides or transduced with mHag cDNA (9–12). Nevertheless, such DC-based in vitro CTL induction protocols from unprimed donors require repeated stimulations of T cells in long-term cultures and are therefore not always successful for the generation of large quantities of mHag-specific T cells within a reasonable time period (12). Furthermore, long-term ex vivo culturing may also impair the in vivo survival capacity of adoptively transferred T cells. Finally, because such protocols require biochemical identification of mHags, they cannot be applied for the ex vivo generation of donor CTLs against hematopoietic mHags defined only by cellular assays. One such novel mHag antigen is recognized by the HLA-DPB1*0401–restricted cytotoxic CD4⁺ T-cell clone 3AB11 isolated from a multiple myeloma patient after HLA-identical sibling SCT. We have previously shown the hematopoietic-restricted and myeloma-associated expression of this mHag by determining the target cell specificity of clone 3AB11 (13).

Toward tackling the current drawbacks of the “mHag-specific adoptive immunotherapy,” in particular for HLA class II–restricted mHags, we here investigated the feasibility and efficacy of introducing the HLA class II–restricted mHag-specific TCR into donor's preactivated T cells via TCR transfer. Although several previous studies simply introduced HLA class I– or class II–restricted TCRs into randomly expanded T cells (14–17), we refrained to follow such a strategy because adoptive transfer of randomly activated donor T cells into recipients involves a high risk for GvHD induction. Considering requirements such as ease of generation and expandability, we chose to introduce mHag-specific TCRs into recall antigen [tetanus toxoid (TT)]–specific T cells. To test the feasibility of our hypothesis, we inserted TCR α- and β-chains of the HLA-DPB1*0401–restricted, mHag-specific T-cell clone 3AB11 into a TT-specific clone derived from the original SC donor. After purification, TCR double-transduced cells were compared with the parent mHag- and TT-specific clones for antigen-specific activity, cytokine secretion profile, and cytotoxic activity, and the TCR-transduced cells were analyzed for their efficacy of in vitro expansion in a TT- or mHag-specific fashion.

Cell culture and T-cell clones. EBV-LCLs and the multiple myeloma (MM) cell line UM9 (13) were cultured in RPMI 1640 (Invitrogen), supplemented with 10% fetal bovine serum (FBS; Integro BV) and antibiotics (100 units/mL penicillin, 100 μg/mL streptomycin; both Life Technologies). The amphotropic Phoenix packaging cell line (ϕ-NX-A) was cultured in DMEM (Invitrogen) containing 10% FBS and antibiotics.

The mHag-specific clone 3AB11. The isolation and functional characterization of HLA class II–restricted, mHag-specific CD4⁺ T-cell clone 3AB11 were reported previously (13). Briefly, 3AB11 was isolated from the peripheral blood mononuclear cells (PBMC) of a MM patient during the development of acute GvHD grade III and a strong GvT response after allo-SCT from his HLA-identical brother. 3AB11 was cytotoxic and produced IFN-γ in response to mHag⁺ MM cells, B cells, T cells, monocytes, but not to nonhematopoietic stromal cells derived from mHag⁺ individuals.

The TT-specific T-cell clone N_TT3CB10. This clone was derived from the HLA-identical SC donor using standard limiting dilution techniques. Briefly, PBMCs were cultured in the presence of TT (15 LF/mL; NVI) and received 60 IU/mL interleukin 2 (IL-2; Chiron B.V.) at day 5. Proliferating cells were cloned at day 6 by limiting dilution at 0.3 cell per well using a feeder cell-cytokine-antigen mixture consisting of 30 Gy irradiated PBMCs from three random donors (10⁶ cells/mL), 50 Gy irradiated EBV-transformed B cells (EBV-LCLs) from two random donors (10⁵ cells/mL), phytohemagglutinin (100 ng/mL; Murex Diagnostics), TT (1.5 LF/mL), and IL-2 (120 IU/mL). Culture medium was RPMI 1640 supplemented with 10% pooled human serum (HS), glutamine, and antibiotics. T-cell clones, expanded by the addition of IL-2 (120 IU/mL), were tested for TT specificity after a second round of stimulation with the above-described feeder cell cytokine mixture. Selected TT-specific clones were further expanded by restimulation with the above-described feeder cell-cytokine-antigen mixture and were cryopreserved until use. In some experiments, T cells were stimulated in a truly mHag- or TT-specific fashion using only mHag⁺-EBV-LCLs or TT-pulsed (7.5 LF/mL for 48-72 h.) mHag⁻ EBV-LCLs as antigen presenting cells (APC), respectively.

Cloning of mHag-specific TCR α- and β-chains into retroviral vectors. cDNA derived from clone 3AB11 was amplified in reverse transcription-PCR using primers covering the entire repertoire of known TCR α- and β-chains. Two in-frame gene transcripts, TCR AV21S1 and TCR BV3S1, were identified and independently cloned into two separate pMX retroviral vectors containing the truncated nerve growth factor receptor (ΔNGF-R) and the green fluorescent protein (GFP) as marker genes, respectively. The genes were inserted downstream to the internal ribosome entry sequence, which ensures coexpression of the transgene and the marker gene. The vectors were then transfected into ϕ-NX-A packaging cells, and the retroviral supernatants were generated as described elsewhere (18). Functionality of the retroviral constructs encoding the TCR α- and β-chains was confirmed by transduction of the TCR-deficient Jurkat clone 76, followed by analysis of TCR α/β cell surface expression (data not shown).

Retroviral transduction of TT-specific T cells. Before retroviral transductions, TT-specific T cells were activated with the above-described feeder-cytokine mix for 48 h. Activated TT cells were then transduced with TCR α and TCR β vectors using RetroNectin (Takara)–coated plates as described elsewhere with some minor adjustments (18). Briefly, a mixture of TCR α and TCR β vectors (0.75 mL each) was incubated in RetroNectin–coated 24-well plates. After 1 h, 0.5 mL of the viral supernatant was replaced with 1 mL of T-cell suspension (1 × 10⁶ cells/mL) in RPMI supplemented with 10% HS and 240 IU/mL IL-2. After 24 h incubation at 37°C and 5% CO₂, the T cells were harvested, and the retroviral transduction was repeated using fresh viral supernatants. Sixteen hours after the second transduction, the cells were washed and expanded with 10% HS and 120 IU/mL IL-2 in RPMI. Retroviral transduction efficiency was determined by fluorescence-activated cell sorting (FACS) analysis 48 h after last transduction. TCR double-transduced cells were then purified by FACS sorting of GFP⁺ and ΔNGF-R⁺ cells using a FACS-Aria sorter (Becton Dickinson).

Immunophenotyping. Cell surface expression of various receptors were determined by FACS analysis using FITC-, PE-, or APC-conjugated monoclonal antibodies (mAb).All mAbs except anti–ΔNGF-R (clone 20.4, culture supernatant) and PE-conjugated goat–anti-mouse immunoglobulin (SBA) were purchased from BD Biosciences. Where indicated, propidium iodide (PI) was added as viable dye. Flow cytometry was done using a FACS-Calibur sorter; the data were analyzed using the CellQuest software (both BD Biosciences).

Cytokine secretion profile of T-cell clones. To induce cytokine secretion, parental and TCR-transduced T-cell clones were stimulated in a mHag- or TT-specific fashion using appropriate EBV-LCLs as APCs. For TT-specific stimulations, mHag⁻ EBV-LCLs were pulsed with 7.5 LF/mL TT for 48 to 72 h. Where indicated, EBV-LCLs were preincubated with anti–HLA-DP, -DQ or -DR antibodies (10 μg/mL; BD Biosciences) for 1 h at 37°C. T cells and APCs were then coincubated at a 1:1 ratio in a final volume of 200 μL in U-bottomed 96-well plates for 24 h. The cytokines released in the culture supernatants were determined using standard sandwich ELISA kits [IL-2, IL-4, IL-5, IL-13, tumor necrosis factor-α (TNF-α), IFN-γ (Biosource), IL-10 (Sanquin)]. The SE of the triplicate measurements never exceeded 10%.

⁵¹Cr release–based 5-h cytotoxicity assays. Serial dilutions of effector T cells were incubated with ⁵¹Cr (50 μCi; Na₂⁵¹CrO₄)-labeled target cells in a final volume of 200 μL RPMI/10% FBS in 96-well U-bottomed microtiter plates (Costar 3799) at 37°C. Spontaneous and maximal ⁵¹Cr releases were determined by incubating target cells with medium alone and with 0.1% Triton X-100 in PBS, respectively. After 5 h, cell-free supernatants were harvested to determine their ⁵¹Cr content by γ-counting on a Cobra autogamma betaplate reader (Packard). The mean percentage of specific lysis in triplicate wells was calculated as follows: % specific lysis = [(mean experimental release − mean spontaneous release)/(mean maximal release − mean spontaneous release)] × 100%. The SE of triplicate cultures never exceeded 15%.

FACS-based 48-h cytotoxicity assays. To determine the lytic activity of T-cell clones on B cells and on MM cells at long term, we slightly modified a previously described FACS-based cytotoxicity assay (19). T-cell clones and target cells (EBV-LCLs or MM cells) were coincubated at different ratios in a final volume of 200 μL RPMI/10% FBS at 37°C. After 48 h, the wells were resuspended thoroughly, 150 μL of cell suspension was drawn and double stained with CD3-FITC plus CD19-APC or CD138-APC antibodies depending on the target cell. After 20 min incubation at room temperature, 2000 Flow-count fluorospheres (Beckman Coulter B.V.) and PI (0.1 mg/mL) were added. The samples were measured immediately on a FACS-Calibur flow cytometer to determine the counts of viable CD19⁺ or CD138⁺ cells. In each sample, the counts of CD19/CD138⁺ target cells were normalized to the counts of fluorospheres, and the % specific lysis induced by T-cell clones was then calculated as follows: % specific lysis = [(normalized counts of control viable target cells) − (normalized counts of coincubation viable target cells)]/[normalized counts of control viable target cells] × 100%. The SE of duplicate cultures never exceeded 15%.

Selection of a donor-derived TT-specific T-cell clone for retroviral TCR transfer. Because virtually all individuals receive tetanus vaccinations, we expected to easily generate TT-specific T cell from the healthy SC donor. Indeed, a 6-day stimulation protocol followed by limiting dilution revealed several CD4⁺ TT-specific T-cell clones. To show the proof of principle of our hypothesis, a TT-specific, easily expandable T-cell clone, N_TT3CB10, was selected for TCR transfer. This clone recognized TT in an HLA-DRB1*1501–restricted fashion because its TT-specific IFN-γ response was blocked by anti–HLA-DR antibodies and was only present against HLA-DRB1*1501⁺ APCs (Fig. 1A and B). Clone N_TT3CB10 showed also significant TT-dependent cytotoxic activity against TT-pulsed donor-derived EBV-LCLs (Fig. 1C).

Dual antigen specificity of TT-specific T cells transferred with mHag-TCR. Clone N_TT3CB10 was retrovirally transduced with the TCR α- and β-chains of mHag-specific T-cell clone 3AB11. After a brief expansion period, the TCR α and β double-transduced cells were enriched to >88% purity by FACS sorting of cells expressing GFP and ΔNGF-R. However, as illustrated in Fig. 2A, isolation of mHag-specific T cells required a second sorting of cells with the highest expression levels of GFP and ΔNGF-R. Only these cells produced significant levels of IFN-γ in response to mHag⁺ EBV-LCLs. These results were consistent with previous studies (20) and suggested that multiple copies of the introduced TCR chains were required for adequate surface expression of mHag-specific TCR heterodimers. In a detailed analysis, the latter cells responded also significantly toward the mHag⁺ MM cell line UM9, but showed no response against mHag⁻ EBV-LCLs from donor (Fig. 2B), illustrating the successful induction of mHag specificity in TT-specific T cells by TCR transfer. TCR-transduced cells also produced high levels of IFN-γ against TT-loaded donor EBV-LCLs, demonstrating that they maintained TT specificity (Fig. 2B).

Cytokine profiles of the mHag-TCR–transduced and the original T-cell clones. To determine whether TCR-transduced cells maintained effector functions of the original mHag-specific and TT-specific T cells, we first compared their cytokine secretion profiles. The original mHag-specific and TT-specific T-cell clones displayed a Th1-like cytokine profile indicated by the production of high levels of IFN-γ, but little or no IL-4 in response to antigen-specific or anti-CD3/CD28–mediated stimulations (Table 1). Both cells produced also significant levels of TNF-α (>1,000 pg/mL) and IL-10 (100-600 pg/mL), but only minimal levels of IL-2, IL-5, and IL-13 (data not shown). Similarly, TCR-transduced cells produced IFN-γ in response to mHag-, TT- or nonspecific stimulations, revealing the maintenance of the original Th1-like cytokine profile (Table 1). TNF-α and IL-10 production levels of the TCR-transduced cells were also similar to the parental cells.

Cytotoxic activity of the mHag-TCR–transduced and the original T-cell clones. From the immunotherapeutic point of view, perhaps the most important effector function of both CD8⁺ as well as CD4⁺ mHag-specific T cells is the cytotoxic activity against relevant target cells, in particular tumor cells. Therefore, next to determining the cytokine profiles, we also compared TCR-transduced cells with the parental T cells for cytotoxic capacity against patient or donor EBV-LCLs and against a mHag⁺ MM cell line UM9. Because the cytotoxic activity of CD4⁺ cells may not always be visible in short assays, we tested the cytotoxic activities in 5-h as well as in 48-h assays. As expected, the original TT-specific clone displayed significant TT-specific cytotoxic activity in 5-h assays. The lysis levels of TT-pulsed target cells increased to ∼75% in 48 h (Fig. 3A). The mHag-specific T-cell clone 3AB11 lysed two mHag⁺ EBV-LCLs and MM cells in 5-h assays (Fig. 3B, top). However, there were subtle differences. Although the mHag⁺ EBV-LCLs derived from an unrelated patient were lysed about 50%, EBV-LCLs of the original MM patient and the mHag⁺ UM9 myeloma cells were lysed about 20%. Nonetheless, in 48-h assays, the lysis levels of all targets were significantly higher and reached to maximum levels of 85% (Fig. 3B, bottom). The TCR-transduced cells displayed both mHag-specific and TT-specific cytotoxic activity similar to the parent cells. Most importantly, the lysis of recipient's EBV-LCLs was 45% to 60% and MM UM9 cells 40% in 48-h assays (Fig. 3C, bottom), indicating that at long term, the TCR-transduced cells exerted effective cytotoxic activity against mHag⁺ malignant B or plasma cells.

Efficient expansion of the mHag-TCR–transduced clone with the original TT antigen. One of the important reasons for inserting a TCR into recall antigen-specific T cells was the assumption that the recall antigen specificity could be exploited for efficient ex vivo expansion of TCR-transferred T cells. We therefore investigated the feasibility and efficacy of expanding TCR-transduced TT-specific cells in a TT-specific fashion and compared this with expansion in a mHag-specific fashion. To this end, the TCR-transduced cells were stimulated either by TT-loaded mHag⁻ EBV-LCLs or by unloaded mHag⁺ EBV-LCLs. The cell cultures were supported by the addition of IL-2; the stimulations were repeated at 14-day intervals, and the expanded cells were tested for mHag- and TT-specific cytotoxic activities. In independent experiments, the TT-specific expansion rates of cells varied between 10- and 100-fold, sufficient for the generation of large numbers of cells for adoptive immunotherapy within 4 weeks. A representative experiment is depicted in Fig. 4; four weeks of expansion of the TCR-transferred T cells in a TT-specific fashion revealed an effective and 15-fold better expansion as compared with mHag-specific stimulation (Fig. 4A). More important, TCR-transduced cells stimulated in a TT-specific fashion showed cytotoxic activity against mHag⁺ target cells at all time points, illustrating the feasibility of expansion of TCR-transferred cells via the intrinsic TT-specific TCR without reducing their immunotherapeutic functions (Fig. 4B). On the other hand, cells stimulated via the transduced mHag-specific TCR did not expand at all and showed poor specific lysis, suggesting a functional deficit when using the introduced TCR for expansion (Fig. 4C).

We evaluated the feasibility of transfer of a HLA class II–restricted, mHag-specific TCR into recall antigen-specific CD4⁺ CTLs. Our results show that TT-specific T cells can be readily isolated from peripheral blood, and that they are suitable target cells for adoptive TCR transfer. Upon transduction with mHag-specific TCRs, the cells display dual TT and mHag specificity and maintain the cytokine profile of the original TT- and mHag-specific T cells. Importantly, these dual-specific T cells displayed effective mHag-dependent cytotoxic activity against myeloma cells and could be successfully expanded through triggering via the TT-specific TCR.

These results show the suitability of this strategy for ex vivo generation and expansion of donor-derived HLA class II–restricted mHag-specific T cells. To our knowledge, our study is the first attempt to rebuild the T-cell specificity toward an HLA class II–restricted potential immunotherapeutic antigen into TT-specific T cells. Unlike several previous studies, we deliberately refrained from using randomly activated donor T cells as targets for adoptive TCR due to a number of reasons. First of all, using randomly activated donor T cells as targets for therapeutic TCRs in an allo-SCT setting introduces a high risk for severe GvHD. Because recipient and donor are mismatched for several mHags, GvHD might be mediated by insertion of TCRs into immunogenic mHag-specific T cells. Theoretically, such a risk can be avoided if recipient's cells are used as targets for TCR transfer. Unfortunately, however, in several cases, like in our case, it may not be possible to access recipient T cells after SCT. Besides this logistic drawback, inserting TCRs into polyclonally activated T cells may introduce other problems such as the risk for introducing TCRs into potential immune deviating Th-2 cells or regulatory T cells. Illustrating this possibility, a recent study reported that upon TCR transduction into randomly activated T cells, the cytokine profile of the original T cells was altered from a Th1-like into Th2-like profile with high production of IL-13 (18). In contrast, targeting TCRs into antigen-specific T cells with a known cytokine profile may reduce such risks. Furthermore, the latter strategy may also reduce the risk of formation of undesired self-reactive TCRs through hybrid formation between endogenous and transduced TCR chains (21). More importantly, the transfer of specific receptors into T cells with known virus specificity or allospecificity may permit the maintenance of the dual antigen-specific T cells via triggering through endogenous TCRs, as shown by a number of previous studies (22–25). As suggested in these studies, the use of cytomegalovirus- and EBV-specific T cells as targets for TCR transfer can permit continuous in vivo stimulation of transduced cells by these latent viruses (22, 23). In the case of influenza-specific T cells, however, it seems possible to boost TCR-transduced cells by influenza vaccinations (25). Although we acknowledge these reports, because we primarily aimed at finding a suitable target cell for HLA class II–restricted TCRs, we chose CD4⁺ TT-specific T cells because the expression of the CD4 molecule is generally required to increase the avidity of TCR-MHC class II/peptide interactions (26). In fact, our results reveal that cytotoxic TT-specific cells can serve as appropriate targets for TCR transfer: after the retroviral transduction of mHag-specific TCRs, the cells displayed dual antigen specificity and could be effectively expanded via stimulation through TT-specific TCRs, without losing the dual antigen specificity. In fact, our results suggest that TT-specific stimulation is probably the best way to expand these cells because stimulation through transgenic mHag-specific TCRs resulted only poor expansion of the cells, which also displayed diminished mHag-specific effector functions.

Naturally, one obvious advantage of inserting the therapeutic mHag-specific TCRs into TT-specific cells is the possibility to boost these dual-specific T cells in vivo by vaccination of the recipients with TT. Because a TT-specific recall response can readily be maintained by TT boosters in healthy individuals, we think that a similar strategy may also be sufficient for the in vivo persistence of dual-specific T cells in recipients of allo-SCT. However, it may also be considered to booster the recipients with TT-loaded donor-DC vaccinations if simple TT boosters seem insufficient to ensure persistence of dual-specific T cells.

In this study, we inserted the mHag-specific TCRs in a well-defined, TT-specific T-cell clone to show the proof of the principle of our strategy. We acknowledge, however, that in a clinical setting, it may be more convenient to use polyclonal TT-specific T cells rather than T-cell clones as targets for TCR transfer. Therefore, a possible improvement of our methodology can be the direct selection of IFN-γ–producing TT-specific T cells using clinically applicable immunomagnetic bead-based IFN-γ capture assays (7). Selection of IFN-γ–secreting cells may also have the advantage that TCRs will be preferentially targeted into CTLs because it is known that IFN-γ secretion and cytotoxic activity of CD4⁺ cells are well correlated (27, 28).

In conclusion, we here present a feasible strategy to generate immunotherapeutic T cells directed against HLA class II–restricted mHags. Because this strategy is suitable for hematopoietic mHags that could be defined only by cellular assays, it may expand the application arena of “mHag-specific” immunotherapy for leukemia and myeloma patients without the need for biochemical or genetic identification of mHags.

We thank Dr. Saskia Ebeling for supervision of TCR cloning.