Optimum in vitro expansion of human antigen-specific CD8⁺ T cells for adoptive transfer therapy

Abstract

Increasing evidence suggests that adoptive transfer of antigen-specific CD8⁺ T cells could represent an effective strategy in the fight against chronic viral infections and malignancies such as melanoma. None the less, a major limitation in the implementation of such therapy resides in the difficulties associated with achieving rapid and efficient expansion of functional T cells in culture necessary to obtain the large numbers required for intravenous infusion. Recently, the critical role of the cytokines interleukin (IL)-2, IL-7 and IL-15 in driving T cell proliferation has been emphasized, thus suggesting their use in the optimization of expansion protocols. We have used major histocompatibility complex (MHC) class I/peptide multimers to monitor the expansion of antigen-specific CD8 T lymphocytes from whole blood, exploring the effect of antigenic peptide dose, IL-2, IL-7 and IL-15 concentrations on the magnitude and functional characteristics of the antigen-specific CD8⁺ T cells generated. We show here that significant expansions of antigen-specific T cells, up to 50% of the CD8⁺ T cell population, can be obtained after a single round of antigen/cytokine (IL-2 or IL-15) stimulation, and that these cells display good cytolytic and interferon (IFN)-γ secretion capabilities. Our results provide an important basis for the rapid in vitro expansion of autologous T cells from the circulating lymphocyte pool using a simple procedure, which is necessary for the development of adoptive transfer therapies.

Introduction

Cytolytic T lymphocytes (CTL) are known to play a major role in the protection against several pathogens, including viruses, bacteria and parasites [1]. More recently, their implication in the fight against malignancies has also become obvious, in particular from animal model studies, in which CTL responses were involved in the rejection of large tumours [2–4]. However, in humans, natural immunity often fails to halt tumour progression, resulting too often in the development of metastases and the death of the patients. Immunotherapy of cancer through the induction of protective CTL responses in patients represents a considerable challenge for the scientific community. Cancer vaccines based on defined tumour antigens have led to the induction of measurable antigen-specific CTL responses in vaccinated individuals [5–8]. None the less, the magnitude of T cell induction remains modest and its impact on tumour progression limited. While more potent vaccine formulations may improve clinical outcome in the future, adoptive transfer of large numbers of antigen-specific CTLs has emerged as a promising approach to readily increase anti-tumour immunological response in cancer patients [9,10]. Increasing evidence from preclinical and clinical studies indicate that such therapy is feasible, safe and well tolerated and, importantly, that it may lead to significant clinical benefits. In the context of viral infections, adoptive transfer of human immunodeficiency virus (HIV)-specific CTL has resulted in a transient reduction of HIV-infected CD4⁺ T cells in patients [11], and the infusion of Epstein–Barr virus (EBV)- or cytomegalovirus (CMV)-specific CTL has been shown to provide a consequential help in prevention and treatment of EBV- and CMV-associated complications, respectively, of bone marrow transplantation [12,13]. Recently, strategies of adoptive transfer have also proved useful in the context of cancer, for instance, resulting in regression of patients’ metastatic melanoma [14,15]. There is a strong hope that the treatment of malignancies with identified antigenic reactivity would be improved significantly if adoptive transfer therapy strategies could be implemented in the clinic. The main limitations for the establishment of standard protocols relate mainly to technical issues, such as the in vitro amplification of specific CTL to reach the numbers of cells required for intravenous infusions. Although tumour-specific T lymphocytes can be recovered from small fragments of tumour in the case of metastatic melanoma [16], practical considerations make the CD8⁺ T lymphocytes present in blood the source of choice. However, the frequency of tumour antigen-specific CD8 T lymphocytes in blood in the vast majority of cancer patients is low and often below the detection limit of currently used monitoring assays, such as fluorescent major histocompatibility complex (MHC) class I/peptide tetramers, or antigen-triggered cytokine release assays [17]. Moreover, the potential complexity and risk associated with prolonged manipulation of human material in vitro represent further difficulties. There is a need, therefore, to develop a simple and efficient method to expand rapidly functional CTL, which could be performed easily in clinical laboratories.

Here, through the study of the T cell expansion in vitro, we have worked on the optimization of culture conditions to enable a maximum yield of antigen-specific T cells available for adoptive transfer therapy. Despite the fact that mature dendritic cells may be particularly efficient at expanding T cells in vitro, their use remains complex and their incorporation into adoptive transfer procedures may not be wished. The cytokines interleukin (IL)-2 and, more recently, IL-7 and IL-15 have raised considerable interest for their potent effect on activation and expansion of T cells [18–20]. They therefore present a strong potential for the optimization of in vitro expansion, although their effects seem to vary according to the type of cells affected (e.g. IL-15 on memory cells or IL-7 on naive cells). This may be particularly relevant because the majority of circulating Melan-A/Mart-1_26–35A27L-specific CD8⁺ T cells from HLA-A2 healthy donors, as well as from metastatic melanoma patients, are naive as defined by their surface markers and functional quiescence [21–23], in contrast with influenza (Matrix Protein_58–66), EBV (BLF1_280–289) and CMV (p65_495–503)-specific cells, which are primarily antigen-experienced CD8 T cells [24–26]. Using MHC class I/peptide fluorescent multimers, we have analysed in detail the influence of cytokine and antigen concentrations on the expansion from healthy donors’ PBMC of tumour or viral antigen-specific CD8⁺ T cells with naive or antigen-experienced characteristics (Melan-A, influenza, EBV and CMV). Using high doses of cytokines, ample expansions of tumour-reactive CD8⁺ T cells can be obtained following a single round of in vitro stimulation which, at this stage, may constitute a valuable source of T cells of controlled specificity for adoptive transfer.

Methods

Reagents

Fluorescent-labelled antibodies specific for CD1a, CD8, CD14, CD27, CD28, CD45RA, CD54, CD80, CD83, CD86, goat anti-rat and interferon (IFN)-γ were obtained from BD PharMingen (San Diego, CA, USA), IL-2 (Proleukin) from Roche Pharma (Basel, Switzerland), IL-7, IL-15 and IL-4 from R&D Systems (Abingdon, UK) and granulocyte-macrophage colony stimulating factor (GM-CSF) from Novartis Pharma Schweiz AG (Bern, Switzerland). Phycoerythrin (PE)-labelled HLA-A2/peptide multimers were synthesized as described [27]. The Influenza Matrix Protein_58–66 (GILGFVFTL), the Melan-A_26–35A27L (ELAGIGILTV), the Melan-A_27–35 (AAGIGILTV), the Melan-A_26–35 (EAAGIGILTV), the CMVpp65_495–503 (NLVPMVATV), the EBV BLF1_280–289 (GLCTLVAML) and the HIV pol 1589 (IVGAETFYV) peptides were synthesized using F-moc chemistry at the peptide synthesis facility, Institute of Biochemistry, University of Lausanne, Switzerland.

Donors and cells

HLA-A2 healthy donors were selected based on the presence in ex-vivo staining of measurable frequencies of one or more multimer⁺ CD8⁺ T cell specificities. The mean frequency of Melan-A_26−35A27L multimer⁺ CD8 T cells was 0·13 ± 0·06%, that of influenza MP_58–66 0·19 ± 0·16, that of EBV BLF1_280–289 0·52 ± 0·49% and that of CMV pp65_495–503 0·49 ± 0·62%. The detection limit ranged between 0·01 and 0·02% of gated CD8 T cells on the basis of background staining observed in HLA-A2 negative lymphocytes (not shown). PBMCs were obtained by Ficoll-Hypaque (Amersham Pharmacia Biotech, Uppsala, Sweden) density gradient centrifugation of heparinized peripheral blood and cryopreserved in RPMI-1640, 20% fetal calf serum (FCS) and 10% dimethylsulphoxide (DMSO) and stored in liquid nitrogen.

Expansion of antigen-specific T cell lines

For unfractionated PBMC stimulation, 10⁵ cells per well were cultivated in medium (RPMI-1640 with glutamax, 1%l-glutamine, 1% non-essential amino-acids, 1% sodium pyruvate, 1% penicillin–streptomycin, 0·1% 2-mercaptoethanol) complemented with human serum (8%) in 96-well round-bottomed plates in the presence of various concentrations of peptide. For dendritic cell (DC) stimulation 3 × 10³ purified CD8⁺ cells were co-cultivated with autologous DC (DC : T ratio 1 : 10) in 96-well plates with soluble peptide. Cytokines were added after 2 days and fresh medium on days 5 and 7. Substantial variability in T cell expansion could be observed between donors, which was not dependent on initial frequencies of antigen-specific precursors, in particular concerning microcultures of naive antigen-specific CD8 T cells with low initial frequency, such as Melan-A-specific cells (< 0·2% of CD8⁺ PBMC population). In order to palliate the problem of variability within microcultures, each condition was set up in 12 replicates, eventually pooled in three replicates before analysis to provide average values (thus, four replicates were analysed each time). Cultures were harvested on day 9. Cells were then stained with specific PE-multimers and surface antibodies. To establish expansion of antigen-specific CD8 T cells, relative frequencies of multimer⁺ per 100 CD8 T cells in wells were calculated using the CellQuest software; moreover, 10⁴ fluorescent beads (Immuno Brite, Beckman Coulter, Nyon, Switzerland) were also added to each sample before cytometric analysis to enable calculation of the total number of multimer⁺ cells as follows: (no. of multimer⁺ cells analysed/no. of beads analysed) × 10⁴.

Generation of dendritic cells

Monocyte-derived dendritic cells (MDDC) were generated from frozen PBMC, as described previously [28]. Briefly, autologous dendritic cells were prepared from the adherent monocyte fraction by culture for 5 days with IL-4 (500 U/ml) and GM-CSF (1000 U/ml) and matured with lipopolysaccharide (LPS) (5 µg/ml, Sigma, St Louis, MO, USA) for 2 additional days. Those cells which displayed the characteristic phenotype of mature dendritic cells (MHC class II high, CD80 high, CD86 high, data not shown) were used immediately to present antigenic peptide to purified autologous CD8⁺ T lymphocytes in a ratio of 1 : 10. CD8⁺ lymphocytes were purified from PBMC by positive selection with magnetic anti-CD8 beads (Miltenyi Biotec, Bergisch Gladbach, Germany).

Functional assays and phenotype analysis

For cytotoxic assays, sorted specific cell lines were cultured with allogeneic feeders and phytohaemagglutinin (PHA) for 10 days and the cytotoxic activity was measured by ⁵¹Cr release assay. Cells were tested for its lytic activity against T2, Me 260 and Me 290 cell lines pulsed or not with the peptide indicated. The percentage of specific lysis was calculated as described previously [29]. For intracellular staining, cells were stimulated with 1 µ M of specific peptide for 4 h at 37°C (after 1 h, 10 µg/ml Brefeldin-A were added). After washing with phosphate buffered saline (PBS), 0·2% bovine serum albumin (BSA) and 5 mM ethylenediamine tetraacetic acid (EDTA), cells were fixed with PBS, 2% glucose and 1% formaldehyde, 5 m M NaN₃ for 10 min at room temperature, and stained with anti-IFN-γ antibody in PBS, 0·1% saponin for 30 min. For the phenotypic analysis of antigen-specific populations, cells were stained for 45 min with fluorescent multimers at room temperature in PBS, 0·2% BSA and 5 m M EDTA followed by labelling with fluorescent conjugates of antibodies to CD8, CD45RA, CD27 or CD28 for 20 min at 4°C.

Results

High doses of IL-2 drive strong expansion of both naive and antigen-experienced CD8⁺ T cells

The optimum procedure to expansion of specific T cells for adoptive transfer should be brief and require as little as possible manipulation of the patient's cells. Amplification of antigen-specific T cell populations from PBMCs is usually performed through repeated in-vitro stimulation with optimal length antigenic peptides in the presence of IL-2. Low doses of IL-2 (between 10 and 50 U/ml) have been used traditionally to avoid the activation/expansion of lymphokine-activated killer cells, as revealed in chromium release assays that were commonly employed to monitor specific T cell expansion. We first determined, at the customary low IL-2 concentration (10 U/ml), the optimal peptide concentration needed for maximum expansion of antigen-specific CTL. The use of fluorescent multimers allowed daily monitoring of specific T cell expansion on small aliquots of single microcultures. Following a single round of stimulation, maximum expansion was usually attained after 9 days of culture, regardless of the antigen concerned (data not shown). While 0·1–1 µ M of peptide was needed for the expansion of Flu-, EBV- or CMV-specific T cells, Melan-A-specific T cells needed higher concentrations of peptide (10 µ M) for a significant expansion to be observed (Fig. 1a), illustrating the distinct antigenic requirement between antigen-experienced T cells and non-primed naive T cells. It is notable that, even at this concentration, the expansion of Melan-A-specific T cells was suboptimal, as their frequency reached after peptide stimulation remained usually low compared to that of memory viral antigen-specific T cells in the same donors.

Fig. 1.

Strong in vitro expansion of antigen-specific CD8⁺ T cells in the presence of high concentrations of interleukin (IL)-2. (a) Expansion of memory or naive CD8⁺ T cells from a single donor in low IL-2 concentrations (10 U/ml). Following one round of stimulation with peptide (Flu/GILGFVFTL, CMV/NLVPMVATV, EBV/GLCTLVAML and ELA/ELAGIGILTV) at different concentrations and 9 days in culture, cells were stained with relevant tetramers to assess expansion. One representative donor of three is shown with percentages of tetramer staining cells in the CD8⁺ T cells (left panel) and corresponding absolute numbers (right panel). (b) Influence of IL-2 titration on antigen-specific T cell expansion. Peripheral blood mononuclear cells (PBMC) were incubated for 9 days with 0·1 µ M (Flu, EBV, CMV) or 10 µ M (ELA) peptide in the presence of various IL-2 concentrations, followed by tetramer analysis. Reduced-size graphs show numbers of tetramer positive cells (log₁₀). Results from two donors are shown (squares or lozenges). (c) Expansion of antigen-specific CD8⁺ T cells with high IL-2 concentrations (1000 U/ml). One representative donor out of two is shown. (d) Expansion of ELA-specific CD8⁺ T cells with peptide and IL-2 titration. One experiment of four is shown. (e) Expansion of ELA-specific CD8⁺ T cells following priming of naive cells with autologous dendritic cells. CD8⁺ T cells were mixed with mature dendritic cells loaded with various concentrations of antigen (+ IL-2 10 U/ml, left panel) or various concentrations of IL-2 (+ antigen 0·1 µM, right panel) (plain squares and circles). Expansions obtained in parallel from PBMC without dendritic cells (DCs) but in the presence of IL-2 1000 U/ml and 10 µ M peptide are shown in comparison (open squares and circles). Reduced-size graphs show numbers of tetramer positive cells (log₁₀). Results from two donors are shown (squares or circles).

We next measured the effect of various IL-2 concentrations in the stimulation protocol. A positive influence of increasing IL-2 dose (from 100 to 1000 U/ml) was particularly evident in the case of Melan-A-specific CD8⁺ T lymphocyte growth, both in percentages as well as in net numbers of multimer⁺ CD8⁺ T cells (Fig. 1b). At high IL-2 concentrations (1000 U/ml) the expansion of these cells, originally naive, could match those of antigen-experienced cells, which also reacted to IL-2 concentrations, but more moderately (Fig. 1c). The double titration of peptide and IL-2 concentrations showed that the use of high concentrations of IL-2 (100–1000 U/ml) supported a robust expansion of Melan-A-specific cells at lower peptide concentrations (0·1–1 µ M) (Fig. 1d). None the less, highest multimer⁺ cell yields were generally obtained with high IL-2 doses and peptide concentrations, which were therefore subsequently used in the assays.

Melan-A-specific CD8⁺ T cells were also stimulated using autologous monocyte-derived mature dendritic cells loaded with peptide. As reported previously [30], this procedure yielded a strong expansion of cells; moreover, the use of dendritic cells enables a decrease of the concentration of peptide (to 0·1 µ M) as well as IL-2 (to 100 U/ml) required for optimal expansion of multimer⁺ CD8 T cells (Fig. 1e). Nevertheless, the DC-driven T cell expansion was of similar scale to that induced in high IL-2 and peptide dose conditions. This suggests that IL-2 at low doses is limiting and the higher efficiency of expansion in the presence of high doses of IL-2 may not be due to indirect effects on other subpopulations of T cells. Considering the handling that implies the generation of autologous DC to stimulate T cells, culture conditions with high IL-2 and peptide concentrations appear to be the method of choice to generate simply and rapidly a high number of Melan-A-specific CD8⁺ T cells.

Beneficial influence of IL-15 on antigen-specific CD8⁺ T cell expansion in contrast to IL-7

IL-7 is known as a cytokine that supports naive T cell regeneration and proliferation and is thus often included in conjunction with low-dose IL-2 for the antigen-driven expansion of naive CD8 T cell precursors from blood. The effect of IL-7 on the expansion of naive CD8 T cell precursors was assessed on its own or in combination with IL-2. The addition of IL-7 (from 1 to 10 ng/ml) alone supported a modest but significant expansion of Melan-A multimer⁺ CD8 T cell expansion (Fig. 2a). Combined with low IL-2 concentrations (10 U/ml), IL-7 (at 10 ng/ml) did not have a measurable effect in the frequency of multimer⁺ CD8 T cells obtained with IL-2 alone, inducing only a small increase of 0·5–3-fold in the yield of multimer⁺ CD8 T cells (not shown). However, the combination of IL-7 with high-dose IL-2 (1000 U/ml) resulted in an IL-7 dose-dependent decline of the Melan-A-specific CD8 T cell expansion (Fig. 2a). This detrimental effect was observed in peptide-stimulated bulk PBMCs as well as in stimulation of CD8 T cells purified from PBMC with autologous mature dendritic cells (not shown). Overall, it appears that the use of IL-7 in conjunction with IL-2, compared to IL-2 alone, is not particularly adapted to an effective expansion of Melan-A-specific CD8⁺ T cells.

Fig. 2.

In vitro expansion of antigen-specific CD8⁺ T cells in the presence of interleukin (IL)-7 or IL-15. (a) ELA-specific CD8⁺ T cell expansion was assessed with various concentrations of IL-7 and in the presence or absence of IL-2 (1000 U/ml). One experiment of three is shown, with percentages of tetramer staining cells in the CD8⁺ T cells (left panel, dashed line corresponds to right y axis) and corresponding absolute numbers (right panel). (b) Effect of IL-15 titration on the expansion of memory (Flu) or naive (ELA) CD8⁺ T cells, with or without IL-2 (10 U/ml). (c) Comparison between IL-15 (5 ng/ml), IL-2 (1000 U/ml) and a combination of the two cytokines on antigen-driven expansion of Flu- or ELA-specific CD8⁺ T cells. Reduced-size graphs show numbers of tetramer positive cells (log₁₀). One experiment of two is shown.

Much attention has been given recently to IL-15, described as a potent proliferation inducing cytokine for memory T cells; its effect on naive T cell homeostasis is less known. We have studied the effect of IL-15 to assist antigen-driven expansion of CD8⁺ T cells, in comparison or in combination with IL-2. First, the titration of the IL-15 dose was analysed in the absence or in the presence of IL-2 at low concentration (10 U/ml): 5 ng/ml (equivalent to 2250 units/ml) of IL-15 were sufficient for the growth of memory cells (influenza-specific), regardless of IL-2 addition (Fig. 2b). Interestingly, the potent effect of IL-15 was also observed on naive CD8⁺ T cells, as the expansion of Melan-A-specific T cells was strongly triggered in the presence of the cytokine alone (Fig. 2b). IL-15 at 5 ng/ml was sufficient to support marked expansion of both antigen-experienced and naive multimer⁺ CD8⁺ circulating T cells. This attained the same extent as that reached when using 1000 U/ml of IL-2; the combination of high IL-2 concentration with IL-15 led to no further cellular expansion (Fig. 2c). The antigen-specific CD8 T cell numbers raised in an IL-15 dose-dependent way in all cultures. IL-15, like IL-2, therefore emerges as a solid candidate to drive tumour-specific CD8⁺ T cell in vitro expansion for adoptive therapy protocols.

Functional capacity and phenotype of in vitro expanded Melan-A-specific CD8⁺ T cells

In recent years, the relationship between T cell differentiation phenotype and functional characteristics has been emphasized, even though the factors driving differentiation and the significance of this process remain to be clarified. We compared the phenotypic changes of Melan-A-specific CD8⁺ T cells emerging from peptide antigen-driven expansion in different cytokine environments by examining the expression of the differentiation markers CD45RA, CD28 and CD27, which are down-regulated upon CD8 T cell differentiation. Starting from a naive phenotype ex vivo (CD45RA⁺), the majority of Melan-A-specific CD8⁺ T cells acquired a phenotype of antigen-experienced cells in culture (CD45RA⁻) (Fig. 3). Interestingly, the cytokine environment had an influence, with further differentiation driven by IL-7, IL-2 and IL-15 in this order (based principally on the expression of CD28 expression as no CD27 down-regulation was generally observed). Larger expansions seemed to be related to more pronounced cell differentiation, thus illustrating further the potent effect of IL-2 and IL-15 to drive expansion and differentiation of T cells, both naive and antigen-experienced.

Fig. 3.

Differentiation phenotype of ELA-specific CD8⁺ T cells expanded in vitro in the presence of interleukin (IL)-2, IL-7 and IL-15. CD45RA, CD28 and CD27 expression was analysed on whole CD8⁺ or ELA tetramer staining T cells cultured for 9 days with 10 µM ELA peptide and IL-2 (1000 U/ml) and/or IL-7 (10 ng/ml) and/or IL-15 (5 ng/ml). Results are representative of two independent experiments.

A caveat with a rapid and strong expansion of T cells may be the risk of generating dysfunctional cells, rendered either anergic or stunned during the process. We therefore assessed effector functions of Melan-A-specific CD8⁺ T cells following their in vitro expansion. To assess the cytotoxic capacity of Melan-A-specific CD8⁺ T cells, multimer⁺ CD8 T cells were tested in chromium release assays. As shown in Fig. 4a, the Melan-A-specific CD8 T cells expanded in 100 or 1000 U/ml of IL-2, were perfectly able to recognize and lyse target cells loaded with the enhanced peptide analogue (Melan-A_26−35A27L), as well as the native nona- and decapeptides (Melan-A_27–35 and Melan-A_26–35). Cells expanded in 10 U/ml of IL-2 could not be tested due to the low number of cells obtained after expansion. IL-15-driven expansions also resulted in cells displaying full cytotoxic capacities (Fig. 4b). All cells could kill naturally antigen-expressing cell lines (i.e. melanoma cell lines Me 290 HLA-A2+/Melan-A⁺) through HLA-A2 restricted recognition (no effect on Me 260 HLA-A2-/Melan-A⁺) (Fig. 4c). Moreover, the majority of Melan-A-specific cells could produce IFN-γ upon short stimulation with cognate peptides presented by T2 target cells (Fig. 4d). Overall, rapidly expanded Melan-A-specific CD8⁺ T cells exhibited CTL effector functions and could be defined as fully functional.

Fig. 4.

Functional analysis of ELA-specific CD8⁺ T cells expanded in vitro in the presence of interleukin (IL)-2 and IL-15. (a) The lytic capacity of ELA-specific CD8⁺ T cells expanded with 100 or 1000 U/ml IL-2 was measured by chromium release assay using target T2 cells loaded with 1 µ M of the ELA/ELAGIGILTV, AA/AAGIGILTV, EAA/EAAGIGILTV or HIV/IVGAETFYV peptides. (b) The lytic capacity of ELA-specific CD8⁺ T cells expanded with IL-2 (1000 U/ml), IL-15 (5 ng/ml) or both cytokines combined was measured by chromium release assay using target T2 cells loaded with ELA, AA, EAA or HIV peptides. (c) The lytic capacity of ELA-specific CD8⁺ T cells expanded with IL-2 (1000 U/ml), IL-15 (5 ng/ml) or both cytokines combined was assessed using Melan-A⁺ (by immunohistochemistry) target tumour cell line either expressing HLA-A2 (Me 290) or not (Me 260) with or without ELA peptide. (d) Interferon (IFN)-γ secretion capacity of ELA-specific CD8⁺ T cells expanded with IL-2 (1000 U/ml), IL-15 (5 ng/ml) or both cytokines combined was assessed by intracellular cytokine staining following 5 h stimulation with T2 cells loaded with ELA, AA, EAA or HIV peptides.

Discussion

The implementation of CTL adoptive transfer therapies for the clinic requires the development of simple and swift procedures for in vitro T cell expansion, adapted to the obtaining of large numbers required for intravenous infusion. Such procedures may be particularly difficult to develop for the expansion of Melan-A-specific CD8⁺ T cells which, although found in substantial numbers in the blood, thus representing a potential source of anti-tumour CTL, display a naive phenotype in most donors, healthy or suffering from melanoma. For instance, the use of autologous dendritic cells has been proposed to be the most efficient way to expand T cells from naive precursors [30]. Because of the identification of IL-2 as a major T cell growth factor [31] this cytokine has been used to grow effector cells in many procedures, although usually at low IL-2 concentrations (10–50 U/ml) in order to avoid lymphokine-activated-killer (LAK) activity. In addition of IL-2, two other cytokines have now been shown to have an important role in CD8⁺ T cell generation and maintenance: IL-7 and IL-15. The present study of antigen and cytokine concentration influence on human antigen-specific CD8⁺ T cells has enabled the definition of optimum culture conditions leading to their expansion.

We show that, following one single round of antigenic stimulation, the use of high concentration of IL-2 on its own gives rise to a strong yield of antigen-specific CD8⁺ T cells, independently of their differentiation status in the circulating lymphocyte pool. Thus, the numbers of initial low-frequency Melan-A precursors detected in PBMC could be drastically increased with the use of 1000 U/ml of IL-2. When a high IL-2 dose is added, non-professional antigen-presenting cells (APCs) are able to be as efficient as dendritic cells for the stimulation of naive specific T cells, therefore providing a simpler method than using autologous dendritic cells. In animal models, in vivo survival and homeostatic proliferation of naive cells was shown to require IL-7 [32,33]

In humans, IL-7 seems to support the induction of a cytotoxic response; however, this effect requires the presence of IL-2 [34]. In our hands, IL-7 did indeed stimulate the development of CTL, but only weakly in comparison with IL-2; and interestingly, the addition of both cytokines led to a reduction in the total cell yield. Recently, it was shown that both IL-2 and IL-7 regulate the proliferation and death of CD4⁺ T cells by modulating Fas expression and susceptibility to FasL-induced cell death [35]. We have also observed the increase in Fas expression in the PBMC cells in the presence of both cytokines, evident between day 5 and day 7 in culture (data not shown). IL-2 and IL-7 could balance in this way the proliferation and apoptosis of T cell pools. The use of IL-7 is therefore not recommended for the expansion of T cells in vitro.

In contrast to the IL-7 relationship to naive cells, IL-15 has been defined more as a factor promoting proliferation of memory T cells [36,37]. In our hands, IL-15 seems to support the expansion of antigen-specific CD8 T cells to a similar extent to that observed with high-dose IL-2. Remarkably, this effect was observed not only with memory T cell precursors but also with naive T cell precursors from peripheral blood (Melan-A-specific). These results are in keeping with two recent studies, in mouse and in humans, showing that IL-15 indeed directly acts on naive CD8⁺ T cells promoting their expansion, differentiation and survival [38,39]. The IL-15 receptor is composed of three chains (α, β, γ), two of which are shared with the IL-2R (β and γ) and only the IL-15Rα being specific for IL-15. Naive T cells express neither IL-2Rα nor IL-15Rα, but they express the common chains β and γ. This suggests that IL-15 may be important in the initial stimulation of naive T cells upon antigen presentation by IL-15-secreting cells, such as macrophages and dendritic cells. It has been shown that IL-15Rα can be provided in trans by neighbouring non-T cells, thus assembling a functional IL-15R, whereby IL-15Rα subunit would bind to the β and γ chains of the receptor on the T cell surface and the incoming IL-15 cytokine [40]. IL-15 represents a particularly strong candidate to drive T cell expansion in vitro, which might be even more adapted than, and therefore preferred to, IL-2, as several studies in mouse models have stressed more advantageous features of IL-15 in comparison to IL-2. IL-15 promotes the maintenance of T cells and induces proliferation of CD8⁺ T cells showing poor effector function [41]. Furthermore, a recent study suggests that IL-15 enhances in vivo anti-tumour activity of tumour-reactive CTL [42].

While it is important to consider issues relative to applicability of the expansion procedure and quantity of expanded cells, it is also essential to take into consideration the quality of the CTLs to be reinfused, that is to say to look at their functional capacity. Dendritic cells have been reported as essential APCs to obtain functional CTLs from Melan-A-specific precursors, capable of recognizing and killing cancer cells [30]. None the less, we have shown here that, following expansions in the presence of IL-2 or IL-15 on their own (i.e. without need for additional DCs in the culture), Melan-A-specific CD8⁺ T cells differentiated fully into antigen-experienced cells and displayed good ability to kill target cells loaded with analogue or native Melan-A antigens as well as melanoma cell lines naturally expressing the antigen. In addition, these cells exhibited the capacity to secrete the effector cytokine IFN-γ upon stimulation. Comparable to more conventional in vitro expansion procedures with low-dose IL-2, the expanded Melan-A-specific T cells entered the resting phase by the end of the second week and required further restimulation to resume proliferation (data not shown). Moreover, Melan-A-specific CD8 T cell clonotypes were reported to greatly expand in vivo and persist at high frequency in peripheral blood for long periods of time after adoptive transfer of TIL expanded in the presence of high-dose IL-2 [14]. Despite being functional, it may also be important that the expanded antigen-specific CD8 T cells go through a step of enrichment before adoptive transfer, in order to shed the remaining bulk of the culture, including natural killer (NK) cells stimulated by IL-2 and IL-15 treatments, and thus to provide purified populations to the patient. The recent observation that high-dose IL-2 may also be efficient at promoting the expansion of regulatory T cells [43] emphasizes further the need of enriching the antigen-specific CD8 T cells. The MHC multimer technology represents the most adapted way to isolate such cells. However, the prolonged interaction between the T cell receptor and peptide-presenting HLA molecule results in activation of the cells, and usually in a significant loss by apoptosis of enriched cells in culture [44–46]. Recent developments have brought a solution to this problem: the use of reversible MHC multimers, which can be rapidly dissociated following T cell enrichment, enables the preservation of the functional status of the cells, so that these cells can be used in adoptive transfer therapy [47]. Overall, a rapid and simple protocol of T cell expansion based on the addition of IL-2 or IL-15 at high concentrations, which can easily serve as a basis for scaling-up, provides a method of choice to generate a strong yield of functional antigen-specific CTL. The development of a clinically applicable protocol of rapid in vitro expansion of antigen-specific CD8 T cells will require the scaling-up of the in-vitro culture conditions in a Good Manufacturing Practice environment. Preliminary experiments using 50-ml tubes as culture support that may mimic the U-bottomed microwell environment provided unsatisfactory results. Thus, this phase clearly needs optimization. Alternatively, it has been shown that large-scale expansion of tumour-infiltrating lymphocytes available for adoptive transfer immunotherapy can be achieved using a large number of the same microtitre plate format reported in this study [48].