MART-1–Specific Melanoma Tumor-Infiltrating Lymphocytes Maintaining CD28 Expression Have Improved Survival and Expansion Capability Following Antigenic Restimulation In Vitro

Yufeng Li, Shujuan Liu, Jessica Hernandez, Luis Vence, Patrick Hwu and Laszlo Radvanyi
J Immunol January 1, 2010, 184 (1) 452-465; DOI: https://doi.org/10.4049/jimmunol.0901101

Abstract

We determined how CD8+ melanoma tumor–infiltrating lymphocytes (TILs) isolated from two distinct phases of expansion in preparation for adoptive T cell therapy respond to melanoma Ag restimulation. We found that TILs isolated after the rapid expansion protocol (REP) phase, used to generate the final patient TIL infusion product, were hyporesponsive to restimulation with MART-1 peptide-pulsed dendritic cells, with many CD8+ T cells undergoing apoptosis. Telomere length was shorter post-REP, but of sufficient length to support further cell division. Phenotypic analysis revealed that cell-surface CD28 expression was significantly reduced in post-REP TILs, whereas CD27 levels remained unchanged. Tracking post-REP TIL proliferation by CFSE dilution, as well as sorting for CD8+CD28+ and CD8+CD28− post-REP subsets, revealed that the few CD28+ TILs remaining post-REP had superior survival capacity and proliferated after restimulation with MART-1 peptide. An analysis of different supportive cytokine mixtures during the REP found that a combination of IL-15 and IL-21 facilitated comparable expansion of CD8+ TILs as IL-2, but prevented the loss of CD28 expression with improved responsiveness to antigenic restimulation post-REP. These results suggest that current expansion protocols using IL-2 for melanoma adoptive T cell therapy yields largely CD8+ T cells unable to persist and divide in vivo following Ag contact. The few CD8+CD28+ T cells that remain may be the only CD8+ TILs that ultimately survive to repopulate the host and mediate long-term tumor control. A REP protocol using IL-15 and IL-21 may greatly increase the number of CD28+ TILs capable of long-term persistence.

Tumor vaccines against defined melanoma Ags, such as gp100 and MART-1, have not yet been successful in the treatment of human metastatic melanoma. A critical reason for this may be the inhibition of CD8+ CTL function and induction of T cell energy in the tumor microenvironment or lack of adequate activation of tumor-reactive T cells in vivo. Adoptive cell therapy (ACT) is an alternative approach that removes tumor-infiltrating lymphocytes (TILs) from this suppressive tumor microenvironment and expands and activates CD8+ and CD4+ T cells in vitro, differentiating them into potent antitumor effector cells that can be reintroduced back into the patient (1, 2). In fact, in recent clinical trials, combining TIL infusion with a single cycle of high-dose IL-2 therapy has demonstrated clinical response rates as high as 51% in patients who underwent prior lymphodepletion induced by cyclophosphamide and fludarabine (3–6). Although ACT using expanded TILs has shown great promise, there are still a number of outstanding issues with this type of therapy that limit its widespread application. Furthermore, although clinical response rates are quite remarkable for Stage IV melanoma, the frequency of actual complete response is still low (<10% of treated patients) and, in many patients, tumor regression is transient in nature (3).

One of the key problems that may limit tumor regression and long-term durable clinical responses in ACT is the persistence of TILs following infusion (7, 8). Long-term TIL persistence has been correlated with accumulation in vivo of T cells from the infused TIL population having an effector memory phenotype, characterized by the expression (or re-expression) of CD27 and CD28 (7). However, the factors regulating or facilitating this TIL survival and persistence are not known, and, moreover, the possible fates of TILs shortly after infusion into patients, when T cells make contact with tumor Ags, reactivating cells through the TCR, has not been addressed systematically. The current approach for ACT in melanoma involves expanding TILs from small tumor fragments or tumor digests by exposure to IL-2 for a period of 4–5 wk. This is followed by acute stimulation of TILs by CD3 ligation with anti-CD3 in the presence of an excess number of irradiated feeder cells and IL-2 for 2 wk, a process called the rapid expansion protocol (REP). The REP greatly expands the number of TILs, usually by 1000- to 3000-fold, before infusion of this final product into the patient. The REP not only expands TILs (up to 100 billion or more), but also drives further T cell differentiation and phenotypic changes that can affect the survival and proliferative capacity of T cells in vivo shortly after infusion, as these cells are reactivated by cytokines and through APCs presenting melanoma Ags (3). How TILs respond to melanoma Ag restimulation after the REP and how this relates to specific phenotypic changes during the REP has not been studied thoroughly.

As reported in this study, we determined the response of MART-1 peptide–specific CD8+ TILs to antigenic restimulation before and after the REP and how this relates to changes in T cell phenotype and effector function. MART-1 is a major Ag in melanoma expressed in >95% of patients, and CD8+ T cells specific for this Ag are commonly found in TIL preparations and can be reliably tracked using tetramers (9–11). Targeting MART-1 has also been used in vaccine clinical trials and adoptive T cell therapy using TCR-transduced PBMCs can mount antitumor responses against melanoma (9, 11). Thus, the fate of MART-1–specific CD8+ T cells can have important ramifications for ACT. TILs from HLA-A2.1+ melanomas were stimulated under optimal conditions using HLA-A2.1+ MART-1 peptide–pulsed mature dendritic cells (DCs) and added IL-2, mimicking a possible scenario in vivo in which infused TILs can make contact with their cognate melanoma Ags. Interestingly, post-REP TILs were found to be hyporesponsive to peptide restimulation, manifested as slow entry into the cell cycle and increased apoptosis. In contrast, under identical conditions, pre-REP MART-1–reactive TILs expanded well, with little induction of apoptosis. Phenotypic and functional analysis of the two TIL populations using a number of different T cell differentiation markers found that CD28 was markedly downmodulated in post-REP cells, whereas CD27 and CD57 levels showed no statistically relevant changes in expression. Sorting of CD28+ and CD28− post-REP CD8+ T cells revealed that only the few remaining CD28+ TILs retained proliferative capacity in response to MART-1 peptide restimulation. Sorted CD28− post-REP TILs did not re-express CD28. We also went on to study whether loss of CD28 during the REP can be prevented with different supportive cytokine mixtures using IL-21 and/or IL-15. Interestingly, a synergy between IL-21 and IL-15 was found with IL-15 driving TIL expansion and IL-21 maintaining CD28 expression. Our results suggest that loss of CD28 but not CD27 expression marks diminished proliferative and survival potential of post-REP melanoma TILs generated using current expansion protocols with IL-2. This has implications for how TIL persistence may be regulated in vivo and suggests that only the few remaining younger, less-differentiated CD28+ TILs are capable of long-term survival and expansion in vivo. These results also explain recent clinical observations in ACT patients showing that CD28 expression predominantly marks a long-term persistent TIL population in patients that have a durable response to ACT (12). Expansion of TILs using IL-15 together with IL-21 instead of IL-2 may help overcome these current shortcomings in ACT.

Materials and Methods

Patient TIL samples and initial TIL expansion

TILs for laboratory studies were obtained from patients with HLA-A0201+ (HLA-A2.1+) Stage IV melanoma who were undergoing surgery at The University of Texas M. D. Anderson Cancer Center according to an Institutional Review Board–approved protocol and following patient consent. The tumor samples were cut into 3- to 5-mm2 pieces and cultured in 2 ml of TIL culture medium (TIL-CM) consisting of RPMI 1640, 10% human Ab serum, 1 mM glutamine, 1 mM sodium pyruvate, 50 μM 2-mercaptoethanol, 1× penicillin-streptomycin, and 20 μg/ml gentamicin (Invitrogen, Carlsbad, CA) in 24-well plates. IL-2 at a concentration of 6000 U/ml was added to all wells. The dividing TIL lines were fed with fresh TIL-CM containing 3000 U/ml IL-2 (Proleukin, Novartis, East Hanover, NJ) and subcultured 1:1 in TIL-CM with 6000 U/ml IL-2 to maintain viable cell density of 0.5–2 × 106/ml.

TIL REP

TILs, initially expanded from tumor fragments as above, were harvested after 5 wk and stained for CD8 expression and recognition of the HLA-A2.1 MART-1 peptide tetramer and other T cell differentiation markers (see below). These TILs were designated as pre-REP and were restimulated with MART-1 peptide as described below or subjected to continued expansion using the REP originally designed by Riddell and Greenberg (13) that is used currently for ACT for Stage IV melanoma. Briefly, 1.3 × 105 pre-REP TILs in TIL-CM were added to upright T-25 flasks containing 30 ng/ml OKT3 (Abbott Labs, Abbott Park, IL) and 26 × 106 allogeneic irradiated (5000 cGy) PBMC feeder cells obtained by pooling PBMC from four normal donors (obtained from the Gulf Coast Regional Blood Bank, Houston, TX) in 20 ml of TIL-CM. On day 2, 10 ml of AIM V medium (Invitrogen) was added together with 6000 U/ml IL-2 (final concentration). The TILs were expanded for another 12 d and diluted as needed with AIM V and IL-2 to keep the viable cell density in the range of 1–4 × 106/ml. The post-REP TILs were isolated and washed in TIL-CM and rested for 3–6 h before restimulation as described below. In some experiments, IL-2 was substituted with IL-15 at 100 ng/ml (R&D Systems, Minneapolis, MN) or IL-21 at 100 ng/ml (BioVision, Mountainview, CA) or a combination of IL-15 (100 ng/ml) together with and IL-21 (100 ng/ml), at all steps in the REP.

Flow cytometric Abs and analysis and sorting of expanded TILs

Expanded TILs were routinely stained for human T cell differentiation markers using fluorochrome-conjugated mAb recognizing CD3, CD4, CD8, CD27, CD28, CD57, and CD62L obtained from BD Biosciences (San Jose, CA) or eBiosciences (La Jolla, CA). TILs were stained with HLA-A2.1 MART-1 peptide (ELAGIGILTV) tetramer (Beckman Coulter, Fullerton, CA) to track changes in the MART-1–specific CD8+ subpopulation. In some experiments, pre-REP and post-REP TILs were also stained for CTLA-4, CD25, CD122, CD132, granzyme B, IFN-γ, and Ki67 (all from BD Biosciences) and PD-1 (BioLegend, San Diego, CA). Apoptosis was monitored by 7-amino-actinomycin D (7-AAD; Sigma-Aldrich, St. Louis, MO) and Annexin V (BD Biosciences) staining of CD8+ and CD8+MART-1 tetramer+ subpopulations. The stained cells were acquired by using a BD FACScanto II flow cytometry analyzer and FACSDiva software (BD Biosciences). Data were analyzed by FlowJo software (TreeStar, San Carlos, CA). In some experiments, the TILs were subjected to FACS using a FACSAria sorter (BD Biosciences) to isolate CD8+CD28+ and CD8+CD28− subpopulations and CD8+CD27+ from CD8+CD27− subpopulations.

Generation of DCs and TIL restimulation with MART-1 peptide–pulsed mature DCs

Human DCs were generated from monocytes obtained from HLA-A2.1+ normal donors (obtained from the Gulf Coast Regional Blood Bank) in T-75 flasks after plastic adherence. The DCs were differentiated with 1000 U/ml GM-CSF and 1000 U/ml IL-4 (R&D Systems) in IMDM with Glutamax, 2% human Ab serum, 1× penicillin-streptomycin, and 50 μM 2-mercaptoethanol (designated as DC-CM). After 5 d, the DCs were matured through exposure to a mixture of IL-1β TNF-α, IL-6, and PGE2 (ITIP) (14). The mature DCs were irradiated at 2000 cGy and pulsed with 3 μg/ml of MART-1 peptide for 90 min. TILs (2 × 106 TIL in 24-well plates) were restimulated by adding 2 × 105 peptide-pulsed mature DCs in TIL-CM with 100 U/ml IL-2 and culturing for 7 to 8 d. IL-2 (to 100 U/ml) was added on day 4.

Measurement of TIL cell division and effector cell activity

Expansion of TILs following restimulation was determined by counting viable cells after trypan blue staining using a hemocytometer. In some experiments, cell division was determined by CFSE dilution in prelabeled pre-REP or post-REP TILs. For these CFSE experiments, TILs were washed in Dulbecco’s PBS and resuspended in Dulbecco’s PBS with 1 μM of CFSE (Molecular Probes-Invitrogen, Carlsbad, CA) for 5–7 min and then washed three times in TIL-CM. Cell division was monitored by using the FITC channel in a FACScanto II flow cytometer (BD Biosciences). CTL activity of TILs was monitored by a flow cytometry-based assay measuring the cleavage of caspase 3 in MART-1 peptide–pulsed T2 target cells as readout (15). Briefly, the T2 cells were labeled with DDAO-SE (Molecule Probes-Invitrogen), washed, and then pulsed with MART-1 peptide or a control HLA-A2.1-binding HIV rev peptide (SLYNTVATL), both at 5 μg/ml, for 1 h. The pulsed T2 cells were incubated with TILs at different effector:target ratios for 3 to 4 h and then fixed, permeabilized, and stained with a PE-conjugated anticleaved caspase 3 rabbit mAb. Staining for IFN-γ production by TILs coincubated with MART-1 peptide–pulsed T2 cells for 5 to 6 h was done by using an intracellular cytokine staining protocol. GolgiStop (BD Biosciences) was added 1 h into the coincubation time. In each case, stained cells were analyzed in a FACScanto II flow cytometer (BD Biosciences).

Fluorescence telomere length assay

Quantitative flow-fluorescence in situ hybridization (FISH) analysis was used to measure the average length of telomere repeats at chromosome ends in individual T cells, as previously described (8). FITC-conjugated telomere probe (FITC-OO-CCCTAACCCTAACCCTAA, O indicating a molecule linking FITC to DNA sequence) was obtained from Dako (Carpinteria, CA). FITC-labeled fluorescent calibration beads (Quantum TM-24; Bangs Laboratories, Fishers, IN) were used to convert telomere fluorescence data to molecules of equivalent soluble fluorescence units (8). Length in base pairs was calculated by the equation: bp = molecules of equivalent soluble fluorescence × 0.495 (8). Aliquots of the K562 cell line, which has a stable telomere length between 4.0 and 4.5 kb, were used each time the assay was run to normalize telomere lengths. Nonspecific sequence probes were included as negative controls.

Microarray experiment and analysis

Post-REP TILs were freshly harvested from the flasks and sorted after anti-CD4 and anti-CD28 staining into two subsets, CD4−CD28+ and CD4−CD28−. After a 3-h rest period to shed staining Abs after sorting, TILs were processed with the RNeasy kit (Qiagen, Valencia, CA) to obtain RNA for microarray analysis. Samples (1 μg) of RNA were subjected to reverse transcription and probe blotting using an Illumina kit, which uses a human Ref6 chip (Illumina, San Diego, CA). Image acquisition and data processing, conducted with BeadStudio software (Illumina), generated a set of genes (detected with a p value < 0.01) as well as data from supervised group analysis. Microarray data sets were explored further using a heatmap server (http://noble.gs.washington.edu/prism/) and Onto-Tool (12; http://vortex.cs.wayne.edu/projects.htm). Real-time PCR was conducted to validate key genes of interest.

Statistical analysis

Data were analyzed by Microsoft Excel (Microsoft, Redmond, WA). The Student t test was used to analyze statistical differences.

Results

Post-REP TILs are hyporesponsive to Ag restimulation

Post-REP melanoma TILs must survive and be able to undergo cell division after infusion into lymphodepleted patients to repopulate the empty lymphoid compartment. Thus, melanoma-specific CD8+ TIL subpopulations should have the capacity to selectively expand and persist in the host when restimulated by APC-presenting melanoma Ags. These cells should have a proliferative advantage over the bulk T cell population if they are to effectively mediate antitumor effector cell responses. We established an in vitro system to address these possibilities by tracking the fate of the MART-1 peptide–specific CD8+ T cell population in pre-REP and post-REP TILs after restimulation with MART-1 peptide–pulsed mature DCs. HLA-A2.1-matched DCs were generated with GM-CSF and IL-4, followed by 1 to 2 d of maturation using ITIP. Screening for CD83, CD86, CD80, CD70, and HLA-A2.1 revealed their expression at expected levels for DC matured by this mixture (Supplemental Fig. 1).

We first tested a number of pre-REP HLA-A2.1+ TIL cultures and confirmed that restimulation with the peptide-pulsed HLA-A2.1-matched DC selectively expanded the CD8+MART-1 tetramer+ T cell population over a background level of expansion (Fig. 1A). We then tested a set of post-REP TIL lines in the same fashion and found that, in contrast, CD8+MART-1–specific T cells were hyporesponsive during the 7-d restimulation period, as determined by tetramer staining (Fig. 1B). This was reflected by a low level of CFSE dilution (Fig. 1C) and a lack of increase of total viable CD8+MART-1 tetramer+ cells (Fig. 1D) in the different patient TIL lines shown. A lack of appreciable proliferation after DC restimulation was also found in the bulk (Ag-nonspecific) CD8+ TILs, as shown by CFSE dilution and counts of total viable cells. Thus, this hyporesponsiveness of post-REP MART-1–specific T cells was due not to an overwhelming expansion of nonspecific T cells, but to the lack of proliferative capacity of the MART-1–specific subpopulation. Fig. 2 summarizes results of seven pre-REP and post-REP melanoma TIL lines showing the fold changes in total T cells and MART-1 peptide–specific CD8+ T cells after the 7-d restimulation period. In both cases, post-REP TILs had substantially lower levels of T cell expansion than pre-REP TILs.

FIGURE 1.
FIGURE 1.

MART-1–specific TILs preferentially expand when restimulated with MART-1 peptide–pulsed DCs, and post-REP TILs are hyporesponsive to this form of restimulation. Mature DCs were generated and pulsed with MART-1 peptide as described in Materials and Methods and then used to activate TILs (1:10 ratio). The cultures were incubated for 7 d. IL-2 (200 U/ml) was added to all cultures to maintain viability. A, Preferential increase in the percentage of MART-1 tetramer+ TILs (pre-REP) by flow cytometry analysis before and after the 7-d restimulation with MART-1 peptide–pulsed normal donor HLA-A2.1+ DCs. Changes in MART-1 tetramer staining (B) and CFSE dilution (C) in pre-REP versus post-REP TILs 7 d after restimulation with MART-1 peptide–pulsed DCs. D, Total number of CD8+MART-1 tetramer+ T cells in post-REP melanoma TILs before and after restimulation with MART-1 peptide–pulsed DCs.

FIGURE 2.
FIGURE 2.

Pre-REP TILs have better proliferative ability than post-REP TILs in response to restimulation with Ag-pulsed DCs. Multiple lines of TILs were expanded from different patients and restimulated at the pre-REP or post-REP stage with MART-1 peptide–pulsed DCs. The fold increase in the bulk CD8+ TIL population (A) and the CD8+MART-1 tetramer+ subset (B) 7 d after restimulation is shown for seven different TIL lines. Each point represents a different TIL sample; the horizontal bars show the averages. The p value was calculated using the Student t test.

Post-REP restimulated TILs have stronger cytolytic ability

In addition to T cell expansion, the acquisition of cytolytic ability is also crucial for effective antitumor T cell responses during ACT. CD8+ T cells exert their CTL function by inflammatory cytokine secretion and direct killing by granzyme B and perforin release. We assessed whether post-REP TILs retained their Ag-specific cytolytic ability despite the loss of proliferative capacity following antigenic restimulation. Pre-REP and post-REP TILs were restimulated with peptide-pulsed DCs as before and tested for CTL activity against MART-1 peptide–pulsed T2 cells after 7 d using a previously described FACS-based caspase 3 cleavage assay (15). Post-REP TILs had a greater cytolytic ability after restimulation on a per-cell basis than pre-REP TILs (Fig. 3A). This was correlated with an increase in the level of granzyme B expression in the MART-1 tetramer+ population and ability to produce IFN-γ in response to MART-1 peptide–pulsed targets (Fig. 3B). These data indicate that post-REP TILs, upon MART-1 restimulation, can differentiate into more potent CTL than pre-REP TILs, despite having a low proliferative potential after restimulation with Ag.

FIGURE 3.
FIGURE 3.

Restimulated post-REP TILs exhibit stronger cytolytic activity than pre-REP TILs. A, CTL activity of restimulated pre-REP and post-REP TILs against MART-1 peptide–pulsed T2 cells using a caspase 3 cleavage CTL assay. The data shown are normalized by dividing the percentage caspase 3 cleavage in the targets by the percentage of CD8+MART-1 tetramer+ T cells added to the CTL assay (1:1 effector:target ratio) to calculate killing per added TIL. Statistical difference was calculated by using Student t test. B, Granzyme B staining of a representative TIL line shows a higher level of staining (MFI for granzyme B) in post-REP TILs. Post-REP TILs synthesized more IFN-γ in response to MART-1 peptide–pulsed DC restimulation as determined by an intracellular cytokine staining assay. Results in B are representative of a minimum of three separate TIL lines tested.

Pre-REP TILs exhibited better long-term persistence in culture after Ag restimulation

In addition to following pre-REP and post-REP TILs over a 7-d peptide restimulation assay, we also continued to track the expansion of MART-1–specific CD8+ TILs for up to 28 d after restimulation to determine their long-term expansion or persistence capacities. Both pre-REP and post-REP TILs were restimulated for 1 wk by MART-1 peptide–pulsed mature DCs. Following this restimulation, the TILs were replated in culture medium with IL-2 and cultured for another 3 wk with IL-2 feeding every 4 d. Supplemental Table I shows the changes in total number and percentage of CD8+MART-1 tetramer+ pre-REP and post-REP T cells in five different patient TIL lines over 28 d after MART-1 peptide restimulation. Overall, pre-REP, MART-1–specific TILs expanded much better during the first 7 d and cells from 3 of the lines continued to expand up to the 28-d time point, with the other two having a drop in MART-1–specific CD8+ T cell recovery and percentage. This is in contrast to pre-REP TILs in which a expansion of MART-1–specific CD8+ T cells was found in only one out of the five TIL lines tested, with most having undetectable MART-1–specific cells at 28 d (Supplemental Table I).

Most MART-1–reactive post-REP TILs fail to enter cell cycle normally and undergo delayed apoptosis after restimulation with MART-1 peptide

We went on to investigate possible reasons accounting for the lack of MART-1 CD8+ T cell expansion in post-REP TILs following restimulation with peptide-pulsed DCs. We first measured the ability of post-REP TILs to enter the cell cycle and determined the level of apoptosis at different times following restimulation with MART-1 peptide. Cell cycle progression was determined by costaining CD8+MART-1 tetramer+ TILs for Ki67 and DNA content using 7-AAD. As shown in Fig. 4A, post-REP TILs mostly failed to enter the cell cycle after the 7-d restimulation period, whereas a significant fraction of pre-REP TILs entered the cell cycle, as determined by Ki67+ staining. Ki67+ cells in the S phase and G2/M phase of the cell cycle, as determined by DNA content, could be clearly seen, whereas very few such cells were found among post-REP TILs. These results correlate with the differences in CFSE dilution described earlier. Measurement of apoptosis (Fig. 4C) in the MART-1 peptide–specific CD8+ and bulk CD8+ T cell populations following peptide-restimulation found that post-REP TILs had an increase in Annexin V+, 7-AAD+ cells (67.06% and 65.84%, respectively, in the two TIL lines shown) after 4 d in comparison with pre-REP TIL (3.93% and 12.98%, respectively, in the two TIL lines shown). Thus, most post-REP TILs failed to enter the cell cycle after antigenic restimulation, and many of these cells later underwent apoptosis (Fig. 4B, 4D). It must be noted that this cell death was not due to cytokine deprivation after day 4, because the cultures were fed with 200 U/ml IL-2 after restimulation and again on day 4.

FIGURE 4.
FIGURE 4.

Cell cycle entry is reduced and apoptosis increased in post-REP TILs after restimulation with Ag. Pre-REP and post-REP TILs including MART-1 tetramer+ CD8+ T cells were stimulated with peptide-pulsed DCs in the presence of IL-2 as already described. The cultures were assayed for cell cycle entry and progression using staining for Ki67 expression and 7-AAD for DNA content after fixation and permeabilization on day 7 of stimulation (A and C). A shows flow cytometry plots of Ki67 versus 7-AAD performed on day 7 after restimulation of two representative TIL lines at the pre- and post-REP stages. B shows the time course of changes in Ki67 staining in these two pre-REP or post-REP CD8+ TILs after restimulation. C, Post-REP CD8+MART-1 tetramer+ TILs underwent apoptosis after restimulation with peptide, as measured by uptake of 7-AAD and Annexin V staining of unfixed cells. D, Time course of changes in 7-AAD+ and Annexin V+ CD8+MART-1 tetramer+ T cells in pre-REP and post-REP TILs after restimulation in two TIL lines.

Post-REP TILs have decreased telomere length

Decreased telomere length has been observed in post-REP TILs, and the loss of persistence of many infused TILs in patients in vivo has been attributed to this decrease (8). We therefore measured telomere lengths by flow-FISH analysis in a panel of independent patient TIL lines. This analysis did not determine telomere lengths of the MART-1–specific CD8+ T cell population because not enough Ag-specific TILs could be isolated for the flow-FISH. Nevertheless, the results should reflect the changes in the MART-1–specific population due to the overall small variation of the flow-FISH peaks in the samples (Fig. 5A), suggesting that most T cell clones in the population had similar telomere lengths. As seen in Fig. 5B, telomere erosion occurred during the REP in most TIL lines; the average telomere length of the pre-REP TIL was 5.7 kb, whereas that of the post-REP TILs had an average loss of 2 kb down to 3.7 kb (p = 0.0004 using Student t test). Thus, a loss of telomeres in CD8+ T cells occurs during the REP; however, sufficient telomere length (3.7 kb) was retained that would support continued cell division.

FIGURE 5.
FIGURE 5.

Telomere length is shortened after the REP. Telomere lengths in nine independent TIL lines were determined before and after the REP using flow cytometric FISH assay. A shows how the telomere length was determined using flow-FISH analysis on a representative TIL line. The telomere lengths were measured in gated G1 phase cells using 7-AAD staining. Flow-FISH assay of K652 cells (stable telomere length) was used to calculate the telomere lengths. Telomere length for nine independent pre-REP versus post-REP TIL lines, with p value, is shown in B. Each point represents a different TIL sample; the heavy bar shows the average.

Loss of CD28 expression in CD8+ MART-specific TILs after the REP

We went on to look at markers of CD8+ T cell differentiation to determine whether any specific phenotypic changes correlated with the loss of post-REP TIL proliferation potential. Decreased cell cycle progression can result from either an increase in T cell–suppressive signaling molecules and coreceptors or a loss of costimulatory or activating receptors. Staining for CD3 (TCR), IL-2Rβ (CD122), IL-2Rγ (CD132), and IL-2Rα (CD25) did not reveal any significant differences between pre-REP and post-REP TILs (data not shown). We also looked at two negative costimulatory molecules, CTLA4 and PD1, and found no significant differences in the percentage of bulk or MART-1–specific CD8+ T cells expressing these molecules between pre-REP and post-REP cultures (data not shown). We stained for the costimulatory molecules CD27 and CD28, key markers of the state of CD8+ T cell differentiation (16, 17). As shown in Fig. 6A, CD8+ TILs exhibited a marked downregulation of CD28 expression (average of 58% CD28+ in the CD8+ in pre-REP TILs versus 28.8% CD28+ in the same subpopulation of post-REP TILs). CD27 expression was also somewhat decreased in the MART-1–specific CD8+ T cell population, but this was not statistically significant over the 15 different TIL lines tested. CD28 expression was also markedly lost in the post-REP CD8+ MART-1 tetramer+ population, and only a minor population retained significant CD28 expression (Fig. 6B). CD27 expression, however, was retained in most of the CD8+ MART-1 tetramer+ population (Fig. 6B). CD57, a marker for end-stage effector CD8+ T cells shown to be hyporesponsive to activation, increased slightly after the REP (Fig. 6A). Thus, the most consistent change in T cell phenotype in CD8+ and the MART-1–specific CD8+ TILs after the REP was a loss of CD28 expression, with a slight increase in the percentage of CD57+ cells and an insignificant change in CD27 expression.

FIGURE 6.
FIGURE 6.

Cell-surface CD28 expression is lost after the REP. Pre-REP and post-REP TILs from over 15 different lines were analyzed for CD27, CD28, and CD57 cell-surface expression in the CD8+ subset using flow cytometry. A, Percentage of positively staining CD8+ cells for each marker is shown for multiple matched pre- and post- REP TILs. Each point represents a different TIL sample; the heavy bar shows the average. The p values were calculated by the Student t test. B, Profound loss of cell-surface CD28 expression in the CD8+MART-1 tetramer+ TIL subset after the REP. Dot plots in B show analysis of CD27 versus CD28 staining in the gated CD8+MART-1 tetramer+ subset in two different matched pre- and post-REP TIL lines.

CD28+ post-REP TILs have improved persistence after restimulation with MART-1 peptide

The data presented so far show that, of the key effector memory costimulatory markers (CD27 and CD28), melanoma TILs lose mostly CD28 expression, and to a lesser extent CD27 expression, following the REP. Although CD28 has been known for many years as a critical costimulatory signal for the productive activation of naive T cells and to prevent activation-induced cell death, several recent studies have shown that memory CD8+ T cells may still require CD28 costimulation for optimal reactivation and expansion (18, 19). With this in mind, we hypothesized that the few remaining CD28+MART-1–specific CD8+ T cells in post-REP TILs may exhibit better persistence and may retain the capacity to divide after restimulation with MART-1 peptide. Although most CD28 expression was lost, some post-REP TILs still retained enough CD28+ cells to allow us to track the fate of CD28+ and CD28−, as well as CD27+ and CD27−, MART-1 tetramer+CD8+ post-REP TILs after restimulation with peptide-pulsed DCs. For these experiments, we also used CD40L-activated and expanded HLA-A2.1+ human B cells as APCs, as these cells express high levels of CD70 (ligand for CD27) in addition to CD86 and CD80 (Supplemental Fig. 1), allowing us to determine the effects of both CD28 and CD27 costimulation on TIL restimulation. The DCs (ITIP-matured) expressed high levels of CD86 and CD80, but low levels of CD70 (Supplemental Fig. 1). Post-REP TILs having MART-1 tetramer+CD8+ T cells were labeled with CFSE and restimulated by HLA-A2.1+ MART-1 peptide–pulsed DCs or CD40L-activated B cells as APCs. After 5–7 d, the cells were stained for CD8, CD28, CD27, and MART-1 tetramer and analyzed by flow cytometry. As shown in Fig. 7A, no distinct difference was observed between the responses to the two different types of APCs used to stimulate the TILs. However, subset analysis (Fig. 7B) revealed a significant degree of CFSE dilution in the MART-1 tetramer+CD28+CD8+ T cells, whereas CFSE remained largely undiluted in the corresponding CD28- population. Interestingly, when CD27+ versus CD27− CD8+ T cells were analyzed in the same manner, no difference in CFSE dilution was observed within the MART-1 tetramer+ population. Annexin V staining of post-REP TILs after restimulation with MART-1 showed that CD28− TILs underwent significantly more apoptosis upon stimulation than CD28+ TILs (data not shown). Further tracking of post-REP MART-1–specific CD8+ TILs over a 10-d period after restimulation found that the CD28+ subpopulation lost CFSE staining and was the dominant subpopulation remaining after 10 d, whereas the CD28− subpopulation disappeared (Fig. 7C). Thus, the small fraction of CD8+CD28+ post-REP TILs exhibited a measurable proliferative advantage following antigenic restimulation and were more resistant to apoptosis.

FIGURE 7.
FIGURE 7.

Only CD28+ TILs remaining after the REP are capable of further cell division in response to restimulation with Ag. Post-REP TILs from an HLA-A2.1+ patient were restimulated with either MART-1 peptide–pulsed HLA-A2.1+ DCs (CD70low/−) or MART-1 peptide–pulsed CD40L-activated HLA-A2.1+ B cells (CD70+). A, FACS dot plots show the change in CD27+ versus CD28+ cells in the CD8+MART-1 tetramer+ population before and after restimulation with the DCs or B cells. B, Differential dilution of CFSE in the gated CD28+ versus CD28− and CD27+ versus CD27−MART-1 tetramer+ T cell subset 7 d after restimulation with peptide-pulsed B cells. C, CFSE dilution versus CD28 staining in the TIL line in B was tracked on days 5, 7, and 10 after restimulation, showing the preferential outgrowth of CD28+ post-REP T cells during culture. D, CD8+CD28+ and CD8+CD28− T cell subsets were sorted from four independent HLA-A2.1+ MART-1–reactive patient TIL lines using FACS followed by restimulation with MART-1 peptide–pulsed DCs and measurement of the fold-change in CD8+MART-1 tetramer+ T cell numbers after 7 d using viable cell counting and FACS staining.

To further explore whether maintenance of CD28 expression is associated with persistence and/or expansion of post-REP TILs during restimulation with Ag, we sorted CD8+CD28+ and CD8+CD28− T cells from post-REP TILs using FACS (Supplemental Fig. 2). The sorted CD8+ T cells were CFSE-labeled and restimulated with peptide-pulsed APCs, as before. After 7 d, CFSE dilution and the total number of recovered MART-1 tetramer+CD8+ T cells were determined. Analysis of five independent post-REP TIL lines found that the numbers of sorted MART-1 tetramer+CD28+CD8+ T cells remained stable or in other cases increased after the restimulation period, whereas the numbers of sorted CD28− cells decreased (Fig. 7D). Both the CD28+ and CD28− subpopulations had similar levels of CD27 expression after the sort (Supplemental Fig. 2); therefore, differences in CD27 costimulation or CD27 expression did not account for these differences in post-REP CD8+ T cell persistence or expansion. To confirm this, we sorted pre-REP TILs (which expand well after restimulation) into CD27+ and CD27− subsets using FACS and analyzed their capacity to expand following restimulation with MART-1 peptide–pulsed DCs or B cells. The sorted CD27+ and CD27− CD8+ pre-REP TILs both expressed CD28 at similar levels (Supplemental Fig. 2). In this case, however, both sorted cell populations expanded similarly in response to restimulation with MART-1 peptide (Supplemental Fig. 3). Thus, sustained and high-level CD28 expression, but not CD27 expression, delineates TILs that have a capacity to continue dividing after restimulation with cognate Ag.

We went on to test whether inhibiting CD28 costimulation by using anti-CD80 and anti-CD86 Abs, or by addition of CTLA4-Fc fusion protein, affected the persistence and/or further expansion of the CD28+ post-REP TILs after restimulation. The activity of these reagents was confirmed by their ability to block proliferation in a mixed lymphocyte response (data not shown). Blocking CD28 costimulation by either of these techniques did not, however, inhibit expansion of CD28+MART-1–specific TILs (data not shown). Thus, CD28 seems to be a marker for post-REP CD8+ TILs with better persistence or expansion capability, but may not be required or function as a costimulatory molecule at this stage.

Differential gene expression analysis by cDNA microarray analysis reveals significant differences between sorted CD8+CD28− and CD8+CD28+ post-REP TILs

To investigate the mechanism of why post-REP CD8+ TILs lose proliferation and survival capacity in relation to a loss of CD28 cell-surface expression, we performed cDNA microarray analysis on three distinct post-REP TIL lines. Each TIL sample was sorted for CD8+CD28+ and CD8+CD28− immediately after the REP and subjected to gene expression analysis using the Illumina human Ref6 chip after RNA isolation (Illumina). The microarray data has been deposited into the GEO database (Accession #GSE16517; http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE16517). The microarray results among the three different sorted post-REP TIL samples were highly concordant, with 12,430 genes expressed (overall p value < 0.01). Among these genes, those with a change in expression of >2-fold were clustered together (Fig. 8A). This analysis showed that a number of different genes coclustered with CD28− TILs and could be clearly separated from CD28+ TILs. Analysis of the most differentially expressed genes in the CD28− subset according to specific functional subsets (Fig. 8B) using pathway analysis software (Onto-Tools; http://vortex.cs.wayne.edu/projects.htm) revealed genes in eight specific pathways that significantly differed between CD28+ and CD28− TILs, including genes involved in Ag processing and presentation (28 of 88 genes differentially expressed), TCR signaling (38 of 93 genes differentially expressed), cell cycle (43 of 112 genes differentially expressed), p53 pathway (29 of 68 genes differentially expressed), and killer inhibitory receptor (KIR) pathway (47 of 131 genes differentially expressed). Interestingly, the KIR pathway genes exhibited the greatest difference between CD28− and CD28+ TILs (Fig. 8B). Table I shows a subset of the most differentially expressed genes in this analysis. Overall, the CD28− TILs had higher expression of cytolytic genes (granzyme family members, perforin, FasL, TNF) and cell cycle inhibitors (p15, p16, and p19). On the other hand, CD28+ TILs showed higher expression of a combined set of early effector and central memory markers, such as IL-7R and CCR7. Interestingly, PD-1, TP53BP2, IL-17RB, and IL-23A were more highly expressed on CD28+ TILs. Staining of post-REP TILs for CD8, CD28, and PD-1 confirmed this difference in PD-1 gene expression at the protein level. PD-1 expression was significantly higher in the CD8+CD28+MART-1 tetramer+ subset, with 26% of this subset staining positive for PD-1 after the REP, whereas only 14% of the post-REP CD8+CD28−MART-1 tetramer+ cells expressed PD-1 (Fig. 8C). Five days after MART-1 peptide restimulation, this difference in PD-1 expression between the two subsets persisted (Fig. 8C). Another striking difference between CD28− and CD28+ TILs was a markedly greater expression of inhibitory KIR receptors (20) on CD28− TILs, such as KIR2DL family members 1 through 5 (Table I). Interestingly, CD28+ TILs expressed more of a p53-binding protein, TP53BP2. As expected, the CD28 gene was highly overexpressed in the sorted CD28+ population (Table I). This also served as an indicator for the purity of the sorted populations used in the microarray analysis.

FIGURE 8.
FIGURE 8.

CD8+CD28+ and CD8+CD28− post-REP TILs exhibit different gene expression profiles, as determined by cDNA microarray analysis. Post-REP TILs from three patients were sorted for CD8+CD28− and CD8+CD28− T cells and gene expression analyzed using cDNA microarray as described in Materials and Methods. A, Heat map showing genes differentially expressed (>2-fold) and having statistically significant differences in expression between CD28+ and CD28− samples (p < 0.01). Onto-Tool software (http://vortex.cs.wayne.edu/projects.htm) was used to determine differences in gene expression in major cellular signaling or functional pathways. B, The signaling pathways or functions that had statistically significant differences between CD28− and CD28+ cells (impact factor >20) are shown, together with the number of genes with each difference out of the total number of genes represented in each pathway. C, Flow cytometry analysis of post-REP TILs for CD28 versus PD-1 staining in gated CD8+MART-1 tetramer+ T cells after the REP and 5 d after restimulation. The calculated percentage of PD-1+ and PD-1− cells as a fraction of the CD28+ or CD28− subset is depicted in the dot plots.

View this table:
Table I. Key immune regulatory genes differentially and not differentially expressed in sorted post-REP CD8+CD28+ versus CD8+CD28− melanoma TILs

Effects of alternative cytokines during the REP on CD28 expression and maintenance of TIL responsiveness to antigenic restimulation

We went on to test whether alternative cytokines other than IL-2, such as IL-15 and IL-21, can support TIL expansion during the REP and yield TILs with a different phenotype, especially in terms of preventing CD28 loss in the CD8+ T cells. First, we tested IL-15, known for its ability to support T cell proliferation and effector cell differentiation (21), in comparison with IL-2 in the REP. However, in six different TIL lines, no major differences were observed when IL-2 (IL-2 REP) or IL-15 (IL-15 REP) was used in the REP. The IL-15 REP expanded TILs to a similar extent as IL-2 (data not shown) with a similar CD8+ T cell phenotype observed after staining for CD27, CD28 (Fig. 9A), and CD57 expression (data not shown). Next, we tested IL-21 alone, or together with IL-15, in comparison with IL-2 and IL-15 alone in the REP. IL-21 has been shown to maintain CD28 expression in activated PBMCs and maintain a younger T cell phenotype (22). However, although IL-21 used alone in the REP preserved CD28 expression on CD8+ TILs, it drove much lower levels of TIL expansion than IL-2 or IL-15 with a 50-fold lower T cell yield (Fig. 9B, 9C). In contrast, when IL-21 was combined with IL-15, TIL expansion was comparable to IL-2 or IL-15 alone, but a significant increase in the percentage of CD28+ CD8+ TILs post-REP was found (Fig. 9B, 9C). Thus, a combination of IL-21 and IL-15 seems to preserve CD28 expression in a significant fraction of TILs while still driving high rates of TIL expansion.

FIGURE 9.
FIGURE 9.

Effects of the REP done with IL-15 and IL-21 on TIL yield and CD8+ T cell phenotype. TILs from the indicated patient lines were subjected to the REP using IL-2 (6000 U/ml), IL-15 (100 ng/ml), IL-21 (100 ng/ml), or IL-15 plus IL-21 (both at 100 ng/ml). Media changes with the indicated cytokines were done as for the REP with IL-2. A, TILs harvested after the REP using IL-2 versus IL-15 alone were analyzed for the expression of CD28 and CD27 in the CD8+ T cell subset by FACS. The percentage of CD27+ and CD28+ CD8+ T cells isolated from six different TIL REPs are shown with the average percentage for each parameter shown as a black bar. No statistical differences in CD27 or CD28 expression were found. B, A representative experiment showing the extent of CD28 expression in CD8+ MART-1 tetramer+ TILs after the REP using IL-2 alone, IL-15 alone, IL-21 alone, or a combination of IL-15 and IL-21. Data are representative of three experiments with similar results. C, Five TIL lines were subjected to the REP with the different cytokines indicated. On day 14 of the REP, the percentage of CD28+ cells in the CD8+ subset was determined in comparison with the fold expansion of the CD8+ T cells under the different conditions. In the case of IL-15 alone, only three TIL lines were tested, with ND denoting “not determined” for the other two lines.

We then determined how TILs subjected to the REP using these different cytokine mixtures responded to MART-1 peptide restimulation using TIL lines containing MART-1 tetramer+ CD8+ T cells from HLA-A2.1+ patients. As shown in Fig. 10A, TILs from IL-21 plus IL-15 REP cultures had an improved Ag-specific CD8+ T cell response to restimulation with MART-1 peptide–pulsed DCs over cells from the IL-2 REP and the IL-15 REP. In two out of three TIL lines shown, TILs isolated from IL-21 plus IL-15 REP cultures had markedly higher increase in CD8+ MART-1 tetramer+ T cells 7 d after restimulation. Although IL-21 used alone preserved CD28 expression during the REP, in most cases the isolated post-REP CD8+ T cells did not have an improved Ag-specific restimulation response over cells from the IL-2 REP cultures (Fig. 10A). On further analysis, we found that CD8+ TILs isolated from IL-21 REP cultures had significantly higher levels of PD-1 expression than TILs isolated after the REP with IL-2, IL-15, or IL-15 plus IL-21 (Fig. 10B). Thus, although IL-21 preserves CD28 expression on the CD8+ T cells, the rate of expansion during the REP is poor after TCR stimulation, and PD-1 levels are maintained at a higher level.

FIGURE 10.
FIGURE 10.

A combination of IL-15 and IL-21 in the REP improves the responses of post-REP CD8+ TILs to restimulation with MART-1. TILs from HLA-A2.1+ patients having a significant fraction of MART-1–reactive T cells were subjected to the REP under the cytokine conditions indicated. The isolated post-REP TILs were restimulated for 1 wk with HLA-A2.1-matched DCs pulsed with MART-1 peptide, as described before. A, The fold expansion of CD8+ MART-1 tetramer+ T cells is shown for three different TIL lines. B, The levels of PD-1 expression in CD8+ TILs isolated after rapid expansion under the cytokine conditions indicated were for analyzed by FACS. A representative experiment from one TIL line is shown.

Discussion

The critical goal of ACT using expanded TILs for melanoma and other cancers is generating a large enough population of Ag-reactive T cells that not only mediates immediate antitumor killing after infusion, but also persists to mediate a longer-term control of metastatic tumor deposits. A major problem in current TIL ACT protocols is the relatively rapid disappearance (within 14–21 d) of most of the infused TILs, which yield only transient control of tumor growth. The reason for the transience of infused TILs is a matter of current debate in the field. TILs infused into the body of the patient with melanoma are faced with a number of factors affecting their continued cell division, effector function, and survival. A key challenge for these cells is reactivation with tumor Ags and other self-Ags in addition to IL-2 signaling and contact with additional homeostatic cytokines in the body. These environmental factors, in context to the state of effector T cell differentiation, will ultimately regulate long-term TIL persistence. At present, however, very little is known about how TILs generated for ACT via the classical REP (13) respond to antigenic restimulation and what phenotypic markers are associated with TILs capable of longer-term persistence and expansion after the REP.

In the current study, we addressed these questions in a well-defined in vitro assay system using expanded MART-1–specific TILs from a number of different melanoma patient tumors. Although other shared melanoma Ags and unique patient-specific Ags are also important in the antitumor response (23), MART-1 is the most widely expressed and most immunogenic melanoma Ag (10, 24). MART-1–specific CD8+ T cells are readily detected in most melanoma TILs and can be reliably tracked using a number of methods such as tetramer staining, allowing us to track the phenotype and fate of these T cells in multiple patient samples (10, 24, 25). MART-1–reactive T cells can also mount effective antitumor responses during ACT in patients (9, 11, 24). Paradoxically, we found that post-REP CD8+MART-1–specific TILs were hyporesponsive to restimulation, even when stimulated in an optimal setting using mature peptide-pulsed DCs. Most TILs failed to enter the cell cycle, with many of the cells undergoing delayed apoptosis 5 d after restimulation with Ag. TILs rested in lower doses of IL-2 (300 U/ml) for a period of 3 to 4 d after the REP also exhibited this hyporesponsiveness. In contrast, TILs isolated from the initial phase of expansion from melanoma tumor fragments (before the REP) were able to enter the cell cycle and proliferated markedly better in response to MART-1 peptide restimulation. Thus, large-scale expansion in the REP induces a set of functional and phenotypic changes in TILs, limiting their continued survival and responsiveness rechallenge with Ag, a situation that may explain the lack of persistence observed in patients. We sought to further understand the nature of this acquired hyporesponsiveness by tracking changes in a number of key effector-memory markers at both the pre-REP and post-REP stages, as well as examining a number of other parameters such as telomere erosion and gene expression.

One of the key issues during T cell expansion in vitro and in vivo is the induction of differentiation during multiple cell divisions associated with appearance of end-stage effector cells and decreased proliferative potential. TILs isolated from melanomas are preactivated T cells with mainly an effector-memory phenotype. We and others have found that freshly isolated CD8+ TILs have mostly a CD27+CD28+CD57−GB+Perf−/low phenotype, indicating an early effector-memory stage of differentiation (26, 27; Y. Li and L. Radvanyi, unpublished observations) before ex vivo expansion for ACT. CD27 and CD28 are critical costimulatory molecules for T cell activation and survival, and their expression levels are used to track the effector-memory and terminal-effector stages of CD8+ T cell differentiation (17, 28–30). Long-term cell division of viral-specific CD8+ T cells due to periodic or chronic viral restimulation eventually downmodulates CD27 and CD28 associated with terminal CTL differentiation, high cytotoxic activity, IFN-γ secretion, and loss of proliferative potential. CD57 increases on fully differentiated CTLs and is another marker that has been associated with hypoproliferative T cells. A classic case is CMV-reactive CD8+CD27−CD28−CD57+GB+Perforin+ cells that are hypoproliferative, yet highly cytotoxic, and increase with age (26). In this context, a rapid and acute expansion of T cells (like the REP of TILs used for ACT) may mimic these long-term changes and generate T cells reminiscent of such senescent antiviral T cells. In our experiments, we stained for CD27, CD28, CD57, and other T cell phenotypic markers known to change during CD8+ T cell differentiation. We found that the most consistent change between pre-REP and post-REP TILs was a profound loss of cell-surface CD28 expression with a slight increase in the percentage of CD57+ cells. In studying over a dozen different TIL lines from different patients, however, we did not observe a statistically significant change in CD27 expression, especially in the MART-1–specific CD8+ subset. Our results support previous studies (e.g., 12, 31, 32) on melanoma TIL expansion where losses of CD28 and CD27 expression were found after extensive expansion of T cells in the REP used for ACT. However, these studies also noted a loss of CD27 expression after the REP that we did not observe in this study. The reason for this difference is unclear, but it is important to note that the other study did not specifically look at a melanoma Ag-specific subpopulation and only tracked changes in the bulk TIL population after the REP. Moreover, subtle differences in protocols and feeders, as well as the timing of IL-2 addition to the REP cultures, may account for this difference. High IL-2 signaling (>3000 U/ml) can downmodulate cell surface CD27 expression through ligation by CD70 on CD8+ TILs, whereas lower IL-2 doses led to CD27 being detected again on the cell surface (33). Ligation of CD27 by CD70 expressed on the CD8+ T cells themselves was found to be the mechanism involved in the CD27 downmodulation (33). Although we did not track CD70 expression on our TILs during the REP, it is possible that the CD8+ TILs used in our study did not express these high levels of CD70 or that the cell density in our cultures did not promote enough contact between CD70+ and CD27+ T cells to mediate this downmodulation. Despite these differences in CD27 modulation between our study and previous studies, the key finding in our study was that CD27 expression status did not affect the proliferation or survival potential of post-REP TILs when restimulated with melanoma Ag. Similar results were obtained in our experiments with pre-REP TILs sorted for CD27+ and CD27− subpopulations before restimulation. Rather, what correlated with the ability of TILs to continue dividing in response to TCR restimulation was the state of CD28 expression.

A key observation was that the few CD28+ TILs remaining after the REP were capable of continued proliferation after restimulation, whereas CD27+ and CD27− TILs reactivated with MART-1 peptide–pulsed APCs (mature DCs and CD40L-activated B cells) responded similarly regardless of whether costimulation was provided by CD70 on the APC (in our case with CD40L-activated B cells). Thus, sustained cell-surface CD28 expression and not CD27 expression delineates TILs capable of continued expansion and persistence. This may explain why so few TIL clonotypes survive long term in ACT patients and why long-term persistent clonotypes found in patients with durable remission after ACT are associated most consistently with high CD28 expression (7, 12). Interestingly, we also found that, despite being a marker for TILs capable of persisting after the REP, CD28 itself did not seem to play a signaling or costimulatory role in driving cell division after restimulation with the peptide, despite the high levels of CD80 and CD86 on the APCs used. This was surprising to us considering recent reports demonstrating a role for CD28 costimulation beyond the naive T cell stage during reactivation of effector memory cells (18). CD28 therefore seems to be phenotypic marker rather than a functional marker on TILs capable of long-term persistence.

An alternative explanation for our inability to detect a signaling/costimulatory role for CD28 in inducing cell division of post-REP MART-1–specific CD8+CD28+ T cells is that a dominant negative signaling pathway may have inhibited the effect of any CD28 costimulation. FACS staining of post-REP TILs from patients with melanoma unexpectedly revealed a relatively high level of cell-surface PD-1 expression on CD8+CD28+ post-REP TILs, whereas significantly less PD-1 expression was found in CD8+CD28− post-REP TILs (Fig. 8C; Y. Li and L. Radvanyi, unpublished observations). We are currently addressing the possibility that PD-1 may play a dominant negative costimulatory role, masking any costimulatory effect of the CD28 pathway during the restimulation of CD8+CD28+ post-REP TILs with MART-1 peptide–pulsed DCs. Increased expression of cell-surface PD-1 on CD8+CD28+ post-REP TILs was supported by our microarray study on sorted post-REP CD8+CD28+ versus CD8+CD28− T cells that found 2-fold higher PD-1 gene expression in the CD8+CD28+ sorted post-REP TILs from all three patients analyzed.

Another issue we addressed is the potential loss of telomere length during T cell division. CD28− T cells have been found to have decreased telomere lengths, and this has been associated with lack of proliferative ability and persistence in vivo (34, 35). In our experiments, we did observe a significant amount of telomere erosion in CD8+ T cells during the REP, which may explain the loss of proliferation capacity. The few remaining CD28+ TILs may have maintained sufficient telomere length to promote continued cell division. We did not separately measure telomere length in the post-REP CD28+ and CD28− TILs using the flow-FISH assay because of the low numbers of CD28+ remaining in the cultures. However, after analyzing multiple TIL lines, the average telomere length in the CD8+ population following the REP was still 3.7 kb, a length still sufficient to support more cell division (36–38). In addition, we have tried a stronger T cell stimulus by activating post-REP TILs using plate-bound anti-CD3 Abs and IL-2. Although this induced a high amount of activation-induced cell death, we found that the remaining live CD8+ and CD8+ MART-1 tetramer+ TILs were capable of cell division, as measured by CFSE dilution (Y. Li and L. Radvanyi, unpublished observations). Although this is a relatively crude assay and does not mimic the more subtle stimulation with Ag presented on HLA, it indicates that at least some of these cells can still be driven into cell division.

The better survival and proliferative capacity of the few remaining MART-1–specific CD8+CD28+ post-REP TILs prompted us to go on and analyze whether any differences in gene expression other than the markers studied existed between these CD28+ cells and the majority of CD8+CD28− T cells. The microarray results from three independent post-REP TIL lines showed a high degree of concordance and revealed significant differences between CD8+CD28+ and CD8+CD28− TILs in the expression of genes in key T cell functional pathways, such as cell cycle regulation, TCR signaling, apoptosis, p53, and KIR. As mentioned above, PD-1 was one of the genes whose expression was increased (confirmed by FACS analysis). Two other genes, CD127/IL-7Rα and CCR7, associated with central memory T cells, were also significantly overexpressed in CD28+ post-REP TILs. We are presently using FACS to determine which fractions of post-REP CD8+CD28+ TILs express the cell-surface protein products of these two genes as well as other similar markers, such as CD62L. It could be that a small subset of MART-1 tetramer+ CD8+CD28+ post-REP TILs has central-memory–like characteristics with an even higher proliferation and survival potential, especially when IL-7 is added. This needs to be determined in future experiments. Two other genes that exhibited increased expression in CD8+CD28+ post-REP TILs were IL-17RB (IL-25R) and IL-23. Both of these genes play a role in the activation and expansion of Th17 cells (39), but the role of IL-17 cells in CD8+ TILs is unknown at present. It is important to note, however, that CD8+ IL-17–producing T cells have been found in breast and colon cancers (40).

A striking difference between CD8+CD28− and CD8+CD28+ sorted post-REP TILs in our microarray study was the high degree of KIR overexpression in the CD28− subset. KIR2DL family members 1 to 5 were all considerably overexpressed in CD28− cells, at levels ranging from 6-fold (KIR2DL5A) to 19-fold (KIR2DL1). Interestingly, the activating NK receptor, NKG2D, did not exhibit higher expression on the CD28− TILs. Other NK markers, such as CD56, known to be expressed on highly differentiated CD8+ CTLs, were also overexpressed. These results correlate with the increased cytolytic activity noted in post-REP TILs and the reduced proliferative potential. The relatively high expression of these KIRs in the post-REP CD28− subset further underscores the great drive toward terminal CD8+ T cell differentiation taking place during the REP. KIRs have been shown to limit the cytolytic activity of differentiated CTLs through ligation of self-HLA molecules such as HLA-C and HLA-A (20). This may have an effect on the tumor-killing function of infused post-REP TILs in patients with melanoma and may also inhibit T cell activation during antigenic restimulation. In the future, it will be important to track the expression of these and other inhibitory NK receptors (e.g., CD94/NKG2A) in patients after TIL infusion in vivo to determine whether any correlation exists between clinical responses and TIL persistence and the quality and quantity of KIRs. If so, blocking KIRs in vivo may enhance ACT success.

A key question emerging from our studies is whether using other methods of performing the REP can yield post-REP T cells with a “younger” phenotype associated with maintenance of CD28 expression and other effector-memory markers that are capable of better persistence in vivo during ACT. In other words, can we have the best of both worlds by generating high numbers of tumor-reactive cytotoxic T cells while maintaining a memory phenotype favorable for continued cell division and long-term survival in vivo? One possibility is changing the culture technology used in the REP (using bioreactors or other large-scale systems), but this is unlikely to affect the resulting phenotype of the TILs after rapid expansion due to the fact that, ultimately, cell division is the major factor driving further T cell differentiation and the phenotypic changes observed. In fact, a recent study with melanoma TILs rapidly expanded with anti-CD3 and IL-2 in a new closed continuous perfusion bioreactor system versus the traditional REP, used by us in this study as well, found no significant difference in the resulting TIL phenotype in terms of the expression of CD28, CD27, and other major phenotypic markers of T cell differentiation (41).

The other obvious alternative is to test the effects of other cytokines besides IL-2 in the REP, such as IL-15 or IL-21. IL-2 used in traditional TIL expansion protocols has been shown to more rapidly drive effector cell differentiation than IL-15 or IL-21 (21, 22). IL-15 may be a good alternative to IL-2, as it has been shown to sustain CD8+ T cells with a younger effector-memory phenotype and facilitates maintenance of a central-memory (CD62L+ and CCR7+) T cell phenotype (42–45). IL-21 has been shown to have an opposing role to IL-2 in driving CTL differentiation (22, 46) and has been found to prevent the loss of CD28 expression after multiple rounds of stimulation of peripheral blood CD8+ T cells with MART-1 peptide (22). We tested the effects of IL-15 and IL-21 versus IL-2 in the REP, but in each case found that it did not improve the situation. IL-15 drove equivalent levels of CD8+ T cell expansion, and the resulting phenotype of the cells was similar to that obtained with IL-2. IL-21 was remarkably good at maintaining a younger phenotype with high CD28 expression, but was poor at expanding the TIL, resulting in low T cell yields post-REP that may not be optimal for clinical application where high numbers (billions) of Ag-specific TILs are needed for infusion. In addition, in most cases, TILs isolated from REP cultures grown with IL-21 did not exhibit an improved response to restimulation with Ag (MART-1) despite the higher fraction of CD28+ CD8+ TILs recovered. A possible explanation for this is that in addition to the increased levels of CD28 and a “younger” phenotype, CD8+ TILs from IL-21 REP cultures also had significantly higher PD-1 expression that may have inhibited the activation and further expansion of the cells (the DCs used as APCs in these experiments all expressed B7H1 and B7DC). Although IL-15 or IL-21 alone did not seem to be viable alternatives to IL-2 in the REP, surprisingly we found that in combination they supported both a significant preservation of CD28+ expression on the post-REP TILs together with high levels of TIL expansion. Thus, IL-15 and IL-21 may have synergistic effects acting through different receptors, with IL-15 strongly driving cell division and having some effects at inhibiting further differentiation and CD28 loss and IL-21 more strongly inhibiting differentiation and preserving CD28 expression while itself being a poor inducer of cell division. It will be interesting to determine how IL-15 and IL-21 signaling is integrated in the CD8+ TILs and whether any unique gene expression signatures exist in TILs expanded with both of these cytokines together. Although we did not test the effects of IL-2 and IL-21 together in the REP, we predict that this combination will have similar effects because both IL-2 and IL-15 signal through the IL-2Rβ chain. However, this needs to be formally addressed.

In conclusion, we have found a profound loss of melanoma Ag–reactive proliferative and survival potential in TILs following the classic REP protocol used for years for melanoma ACT. Loss of CD28 expression in post-REP TILs was the most consistent effector-memory marker loss during the REP, whereas CD27 levels did not change significantly. This observation had functional relevance in that the few remaining post-REP CD28+ Ag-specific CD8+ TILs were the only cells capable of continued cell division following restimulation with APCs. Thus, our data offer an explanation for why CD28+ T cells dominate the repopulating T cell pool in lymphodepleted patients during ACT and why CD28+ T cells dominate the long-term persistent T cell population in patients with more durable responses. Our microarray data on CD28+ and CD28− post-REP TILs also revealed potential targets for further therapeutic intervention (e.g., PD-1 blockade and KIR blockade) in vivo after TIL infusion, or during the REP, that may further enhance the success of current ACT protocols. Finally, our data point out some of the limitations with current ACT protocols using TILs and the need to develop methods propagating TILs that maintain a “younger” phenotype associated with higher survival and expansion capabilities in vivo after infusion into lymphodepleted patients. This may be achieved by using pre-REP instead of post-REP TILs as the ACT infusion product. In addition, our results suggest that IL-15 in synergy with IL-21 may be superior to IL-2 in driving TIL expansion during the classical REP used to generate ACT products and may maintain a more Ag-responsive T cell phenotype associated with the preservation of CD28 expression. More detailed examination of the changes in gene expression profiles during the REP is another approach that can identify key genes that can be modulated to enhance post-REP TIL survival and improve Ag-induced cell division.

Acknowledgments

We thank Natalia Martin-Orozco for help in preparation of the figures and tables. We thank Priscilla Miller for help in obtaining tissue samples and Victor Prieto, Hafeez Diwan, and Alex Lazar of the Melanoma Tissue Bank at M. D. Anderson Cancer Center. The following surgeons at M. D. Anderson Cancer Center generously helped in obtaining tissues for our ACT clinical trial and for in vitro studies: Jeffrey Lee, Jeffrey Gershenwald, Merrick Ross, Janice Cormier, Anthony Lucci, and Paul Mansfield. The following technicians from M. D. Anderson Cancer Center helped process tumor material to obtain TILs for in vitro studies: Kathryn Bushnell, Rahmatu Mansaray, Orenthial Fulbright, Marissa Gonzalez, Chris Toth, and Renjith Ramachandran.

Disclosures The authors have no financial conflicts of interest.

Footnotes

  • This work was supported by National Institutes of Health/National Cancer Institute Grants 1RO1 CA111999-01A2 and P50 CA093459 and by a grant from the Miriam and Sheldon Adelson Medical Research Foundation.

  • The sequences presented in this article have been submitted to GEO database under accession number GSE16517.

  • The online version of this article contains supplemental material.

  • Abbreviations used in this paper:

    7-AAD
    7-aminoactinomycin D
    ACT
    adoptive T cell therapy
    DC
    dendritic cell
    FISH
    fluorescence in situ hybridization
    ITIP
    mixture of IL-1β, TNF-α, IL-6, and PGE2
    KIR
    killer inhibitory receptor
    REP
    rapid expansion protocol
    TIL
    tumor-infiltrating lymphocyte
    TIL-CM
    TIL culture medium.

  • Received April 6, 2009.
  • Accepted October 26, 2009.

References

  1. ↵
    1. June C. H.
    . 2007. Adoptive T cell therapy for cancer in the clinic. J. Clin. Invest. 117: 1466–1476.
    OpenUrlCrossRefPubMed
  2. ↵
    1. June C. H.
    . 2007. Principles of adoptive T cell cancer therapy. J. Clin. Invest. 117: 1204–1212.
    OpenUrlCrossRefPubMed
  3. ↵
    1. Rosenberg S. A.,
    2. N. P. Restifo,
    3. J. C. Yang,
    4. R. A. Morgan,
    5. M. E. Dudley
    . 2008. Adoptive cell transfer: a clinical path to effective cancer immunotherapy. Nat. Rev. Cancer 8: 299–308.
    OpenUrlCrossRefPubMed
    1. Dudley M. E.,
    2. S. A. Rosenberg
    . 2003. Adoptive-cell-transfer therapy for the treatment of patients with cancer. Nat. Rev. Cancer 3: 666–675.
    OpenUrlCrossRefPubMed
    1. Rosenberg S.A.
    . 2005. Cancer immunotherapy comes of age. Nat. Clin. Pract. Oncol. 2: 115.
    OpenUrlCrossRefPubMed
  6. ↵
    1. Rosenberg S. A.,
    2. J. C. Yang,
    3. N. P. Restifo
    . 2004. Cancer immunotherapy: moving beyond current vaccines. Nat. Med. 10: 909–915.
    OpenUrlCrossRefPubMed
  7. ↵
    1. Robbins P. F.,
    2. M. E. Dudley,
    3. J. Wunderlich,
    4. M. El-Gamil,
    5. Y. F. Li,
    6. J. Zhou,
    7. J. Huang,
    8. D. J. Powell Jr.,
    9. S. A. Rosenberg
    . 2004. Cutting edge: persistence of transferred lymphocyte clonotypes correlates with cancer regression in patients receiving cell transfer therapy. J. Immunol. 173: 7125–7130.
    OpenUrlAbstract/FREE Full Text
  8. ↵
    1. Zhou J.,
    2. X. Shen,
    3. J. Huang,
    4. R. J. Hodes,
    5. S. A. Rosenberg,
    6. P. F. Robbins
    . 2005. Telomere length of transferred lymphocytes correlates with in vivo persistence and tumor regression in melanoma patients receiving cell transfer therapy. J. Immunol. 175: 7046–7052.
    OpenUrlAbstract/FREE Full Text
  9. ↵
    1. Benlalam H.,
    2. V. Vignard,
    3. A. Khammari,
    4. A. Bonnin,
    5. Y. Godet,
    6. M. C. Pandolfino,
    7. F. Jotereau,
    8. B. Dreno,
    9. N. Labarrière
    . 2007. Infusion of Melan-A/Mart-1 specific tumor-infiltrating lymphocytes enhanced relapse-free survival of melanoma patients. Cancer Immunol. Immunother. 56: 515–526.
    OpenUrlCrossRefPubMed
  10. ↵
    1. Derré L.,
    2. M. Ferber,
    3. C. Touvrey,
    4. E. Devevre,
    5. V. Zoete,
    6. A. Leimgruber,
    7. P. Romero,
    8. O. Michielin,
    9. F. Lévy,
    10. D. E. Speiser
    . 2007. A novel population of human melanoma-specific CD8 T cells recognizes Melan-AMART-1 immunodominant nonapeptide but not the corresponding decapeptide. J. Immunol. 179: 7635–7645.
    OpenUrlAbstract/FREE Full Text
  11. ↵
    1. Morgan R. A.,
    2. M. E. Dudley,
    3. J. R. Wunderlich,
    4. M. S. Hughes,
    5. J. C. Yang,
    6. R. M. Sherry,
    7. R. E. Royal,
    8. S. L. Topalian,
    9. U. S. Kammula,
    10. N. P. Restifo,
    11. et al
    . 2006. Cancer regression in patients after transfer of genetically engineered lymphocytes. Science 314: 126–129.
    OpenUrlAbstract/FREE Full Text
  12. ↵
    1. Powell D. J., Jr.,
    2. M. E. Dudley,
    3. P. F. Robbins,
    4. S. A. Rosenberg
    . 2005. Transition of late-stage effector T cells to CD27+ CD28+ tumor-reactive effector memory T cells in humans after adoptive cell transfer therapy. Blood 105: 241–250.
    OpenUrlAbstract/FREE Full Text
  13. ↵
    1. Riddell S. R.,
    2. P. D. Greenberg
    . 1990. The use of anti-CD3 and anti-CD28 monoclonal antibodies to clone and expand human antigen-specific T cells. J. Immunol. Methods 128: 189–201.
    OpenUrlCrossRefPubMed
  14. ↵
    1. Berger T. G.,
    2. B. Feuerstein,
    3. E. Strasser,
    4. U. Hirsch,
    5. D. Schreiner,
    6. G. Schuler,
    7. B. Schuler-Thurner
    . 2002. Large-scale generation of mature monocyte-derived dendritic cells for clinical application in cell factories. J. Immunol. Methods 268: 131–140.
    OpenUrlCrossRefPubMed
  15. ↵
    1. He L.,
    2. J. Hakimi,
    3. D. Salha,
    4. I. Miron,
    5. P. Dunn,
    6. L. Radvanyi
    . 2005. A sensitive flow cytometry-based cytotoxic T-lymphocyte assay through detection of cleaved caspase 3 in target cells. J. Immunol. Methods 304: 43–59.
    OpenUrlCrossRefPubMed
  16. ↵
    1. Hamann D.,
    2. P. A. Baars,
    3. M. H. Rep,
    4. B. Hooibrink,
    5. S. R. Kerkhof-Garde,
    6. M. R. Klein,
    7. R. A. van Lier
    . 1997. Phenotypic and functional separation of memory and effector human CD8+ T cells. J. Exp. Med. 186: 1407–1418.
    OpenUrlAbstract/FREE Full Text
  17. ↵
    1. Sallusto F.,
    2. J. Geginat,
    3. A. Lanzavecchia
    . 2004. Central memory and effector memory T cell subsets: function, generation, and maintenance. Annu. Rev. Immunol. 22: 745–763.
    OpenUrlCrossRefPubMed
  18. ↵
    1. Borowski A. B.,
    2. A. C. Boesteanu,
    3. Y. M. Mueller,
    4. C. Carafides,
    5. D. J. Topham,
    6. J. D. Altman,
    7. S. R. Jennings,
    8. P. D. Katsikis
    . 2007. Memory CD8+ T cells require CD28 costimulation. J. Immunol. 179: 6494–6503.
    OpenUrlAbstract/FREE Full Text
  19. ↵
    1. Fang M.,
    2. L. J. Sigal
    . 2006. Direct CD28 costimulation is required for CD8+ T cell-mediated resistance to an acute viral disease in a natural host. J. Immunol. 177: 8027–8036.
    OpenUrlAbstract/FREE Full Text
  20. ↵
    1. Vivier E.,
    2. N. Anfossi
    . 2004. Inhibitory NK-cell receptors on T cells: witness of the past, actors of the future. Nat. Rev. Immunol. 4: 190–198.
    OpenUrlCrossRefPubMed
  21. ↵
    1. Rochman Y.,
    2. R. Spolski,
    3. W. J. Leonard
    . 2009. New insights into the regulation of T cells by gamma(c) family cytokines. Nat. Rev. Immunol. 9: 480–490.
    OpenUrlCrossRefPubMed
  22. ↵
    1. Li Y.,
    2. M. Bleakley,
    3. C. Yee
    . 2005. IL-21 influences the frequency, phenotype, and affinity of the antigen-specific CD8 T cell response. J. Immunol. 175: 2261–2269.
    OpenUrlAbstract/FREE Full Text
  23. ↵
    1. Michaeli Y.,
    2. G. Denkberg,
    3. K. Sinik,
    4. L. Lantzy,
    5. C. Chih-Sheng,
    6. C. Beauverd,
    7. T. Ziv,
    8. P. Romero,
    9. Y. Reiter
    . 2009. Expression hierarchy of T cell epitopes from melanoma differentiation antigens: unexpected high level presentation of tyrosinase-HLA-A2 Complexes revealed by peptide-specific, MHC-restricted, TCR-like antibodies. J. Immunol. 182: 6328–6341.
    OpenUrlAbstract/FREE Full Text
  24. ↵
    1. Jäger E.,
    2. M. Maeurer,
    3. H. Höhn,
    4. J. Karbach,
    5. D. Jäger,
    6. Z. Zidianakis,
    7. A. Bakhshandeh-Bath,
    8. J. Orth,
    9. C. Neukirch,
    10. A. Necker,
    11. et al
    . 2000. Clonal expansion of Melan A-specific cytotoxic T lymphocytes in a melanoma patient responding to continued immunization with melanoma-associated peptides. Int. J. Cancer 86: 538–547.
    OpenUrlCrossRefPubMed
  25. ↵
    1. Speiser D. E.,
    2. P. Baumgaertner,
    3. C. Barbey,
    4. V. Rubio-Godoy,
    5. A. Moulin,
    6. P. Corthesy,
    7. E. Devevre,
    8. P. Y. Dietrich,
    9. D. Rimoldi,
    10. D. Liénard,
    11. et al
    . 2006. A novel approach to characterize clonality and differentiation of human melanoma-specific T cell responses: spontaneous priming and efficient boosting by vaccination. J. Immunol. 177: 1338–1348.
    OpenUrlAbstract/FREE Full Text
  26. ↵
    1. Sacre K.,
    2. G. Carcelain,
    3. N. Cassoux,
    4. A. M. Fillet,
    5. D. Costagliola,
    6. D. Vittecoq,
    7. D. Salmon,
    8. Z. Amoura,
    9. C. Katlama,
    10. B. Autran
    . 2005. Repertoire, diversity, and differentiation of specific CD8 T cells are associated with immune protection against human cytomegalovirus disease. J. Exp. Med. 201: 1999–2010.
    OpenUrlAbstract/FREE Full Text
  27. ↵
    1. Takata H.,
    2. M. Takiguchi
    . 2006. Three memory subsets of human CD8+ T cells differently expressing three cytolytic effector molecules. J. Immunol. 177: 4330–4340.
    OpenUrlAbstract/FREE Full Text
  28. ↵
    1. Kaech S. M.,
    2. E. J. Wherry
    . 2007. Heterogeneity and cell-fate decisions in effector and memory CD8+ T cell differentiation during viral infection. Immunity 27: 393–405.
    OpenUrlCrossRefPubMed
    1. Appay V.,
    2. P. R. Dunbar,
    3. M. Callan,
    4. P. Klenerman,
    5. G. M. Gillespie,
    6. L. Papagno,
    7. G. S. Ogg,
    8. A. King,
    9. F. Lechner,
    10. C. A. Spina,
    11. et al
    . 2002. Memory CD8+ T cells vary in differentiation phenotype in different persistent virus infections. Nat. Med. 8: 379–385.
    OpenUrlCrossRefPubMed
  30. ↵
    1. Appay V.,
    2. S. L. Rowland-Jones
    . 2004. Lessons from the study of T-cell differentiation in persistent human virus infection. Semin. Immunol. 16: 205–212.
    OpenUrlCrossRefPubMed
  31. ↵
    1. Tran K. Q.,
    2. J. Zhou,
    3. K. H. Durflinger,
    4. M. M. Langhan,
    5. T. E. Shelton,
    6. J. R. Wunderlich,
    7. P. F. Robbins,
    8. S. A. Rosenberg,
    9. M. E. Dudley
    . 2008. Minimally cultured tumor-infiltrating lymphocytes display optimal characteristics for adoptive cell therapy. J. Immunother. 31: 742–751.
    OpenUrlCrossRefPubMed
  32. ↵
    1. Huang J.,
    2. H. T. Khong,
    3. M. E. Dudley,
    4. M. El-Gamil,
    5. Y. F. Li,
    6. S. A. Rosenberg,
    7. P. F. Robbins
    . 2005. Survival, persistence, and progressive differentiation of adoptively transferred tumor-reactive T cells associated with tumor regression. J. Immunother. 28: 258–267.
    OpenUrlCrossRefPubMed
  33. ↵
    1. Huang J.,
    2. K. W. Kerstann,
    3. M. Ahmadzadeh,
    4. Y. F. Li,
    5. M. El-Gamil,
    6. S. A. Rosenberg,
    7. P. F. Robbins
    . 2006. Modulation by IL-2 of CD70 and CD27 expression on CD8+ T cells: importance for the therapeutic effectiveness of cell transfer immunotherapy. J. Immunol. 176: 7726–7735.
    OpenUrlAbstract/FREE Full Text
  34. ↵
    1. Monteiro J.,
    2. F. Batliwalla,
    3. H. Ostrer,
    4. P. K. Gregersen
    . 1996. Shortened telomeres in clonally expanded CD28-CD8+ T cells imply a replicative history that is distinct from their CD28+CD8+ counterparts. J. Immunol. 156: 3587–3590.
    OpenUrlAbstract
  35. ↵
    1. Palmer L. D.,
    2. N. Weng,
    3. B. L. Levine,
    4. C. H. June,
    5. H. C. Lane,
    6. R. J. Hodes
    . 1997. Telomere length, telomerase activity, and replicative potential in HIV infection: analysis of CD4+ and CD8+ T cells from HIV-discordant monozygotic twins. J. Exp. Med. 185: 1381–1386.
    OpenUrlAbstract/FREE Full Text
  36. ↵
    1. Ahmed A.,
    2. T. Tollefsbol
    . 2001. Telomeres and telomerase: basic science implications for aging. J. Am. Geriatr. Soc. 49: 1105–1109.
    OpenUrlCrossRefPubMed
    1. Brümmendorf T. H.,
    2. S. Balabanov
    . 2006. Telomere length dynamics in normal hematopoiesis and in disease states characterized by increased stem cell turnover. Leukemia 20: 1706–1716.
    OpenUrlCrossRefPubMed
  38. ↵
    1. Röth A.,
    2. G. M. Baerlocher,
    3. M. Schertzer,
    4. E. Chavez,
    5. U. Dührsen,
    6. P. M. Lansdorp
    . 2005. Telomere loss, senescence, and genetic instability in CD4+ T lymphocytes overexpressing hTERT. Blood 106: 43–50.
    OpenUrlAbstract/FREE Full Text
  39. ↵
    1. Dong C.
    . 2008. Regulation and pro-inflammatory function of interleukin-17 family cytokines. Immunol. Rev. 226: 80–86.
    OpenUrlCrossRefPubMed
  40. ↵
    1. Nam J. S.,
    2. M. Terabe,
    3. M. J. Kang,
    4. H. Chae,
    5. N. Voong,
    6. Y. A. Yang,
    7. A. Laurence,
    8. A. Michalowska,
    9. M. Mamura,
    10. S. Lonning,
    11. et al
    . 2008. Transforming growth factor β subverts the immune system into directly promoting tumor growth through interleukin-17. Cancer Res. 68: 3915–3923.
    OpenUrlAbstract/FREE Full Text
  41. ↵
    1. Klapper J. A.,
    2. A. A. Thomasian,
    3. D. M. Smith,
    4. G. C. Gorgas,
    5. J. R. Wunderlich,
    6. F. O. Smith,
    7. B. S. Hampson,
    8. S. A. Rosenberg,
    9. M. E. Dudley
    . 2009. Single-pass, closed-system rapid expansion of lymphocyte cultures for adoptive cell therapy. J. Immunol. Methods 345: 90–99.
    OpenUrlCrossRefPubMed
  42. ↵
    1. Bodnár A.,
    2. E. Nizsalóczki,
    3. G. Mocsár,
    4. N. Szalóki,
    5. T. A. Waldmann,
    6. S. Damjanovich,
    7. G. Vámosi
    . 2008. A biophysical approach to IL-2 and IL-15 receptor function: localization, conformation and interactions. Immunol. Lett. 116: 117–125.
    OpenUrlCrossRefPubMed
    1. Daudt L.,
    2. R. Maccario,
    3. F. Locatelli,
    4. I. Turin,
    5. L. Silla,
    6. E. Montini,
    7. E. Percivalle,
    8. R. Giugliani,
    9. M. A. Avanzini,
    10. A. Moretta,
    11. D. Montagna
    . 2008. Interleukin-15 favors the expansion of central memory CD8+ T cells in ex vivo generated, antileukemia human cytotoxic T lymphocyte lines. J. Immunother. 31: 385–393.
    OpenUrlCrossRefPubMed
    1. Klebanoff C. A.,
    2. L. Gattinoni,
    3. P. Torabi-Parizi,
    4. K. Kerstann,
    5. A. R. Cardones,
    6. S. E. Finkelstein,
    7. D. C. Palmer,
    8. P. A. Antony,
    9. S. T. Hwang,
    10. S. A. Rosenberg,
    11. et al
    . 2005. Central memory self/tumor-reactive CD8+ T cells confer superior antitumor immunity compared with effector memory T cells. Proc. Natl. Acad. Sci. USA 102: 9571–9576.
    OpenUrlAbstract/FREE Full Text
  45. ↵
    1. Mueller K.,
    2. O. Schweier,
    3. H. Pircher
    . 2008. Efficacy of IL-2- versus IL-15-stimulated CD8 T cells in adoptive immunotherapy. Eur. J. Immunol. 38: 2874–2885.
    OpenUrlCrossRefPubMed
  46. ↵
    1. Hinrichs C. S.,
    2. R. Spolski,
    3. C. M. Paulos,
    4. L. Gattinoni,
    5. K. W. Kerstann,
    6. D. C. Palmer,
    7. C. A. Klebanoff,
    8. S. A. Rosenberg,
    9. W. J. Leonard,
    10. N. P. Restifo
    . 2008. IL-2 and IL-21 confer opposing differentiation programs to CD8+ T cells for adoptive immunotherapy. Blood 111: 5326–5333.
    OpenUrlAbstract/FREE Full Text
Back to top

In this issue

Print
Download PDF
Article Alerts
Email Article
Citation Tools

Jump to section