Co-stimulation through 4-1BB/CD137 improves the expansion and function of CD8(+) melanoma tumor-infiltrating lymphocytes for adoptive T-cell therapy

Abstract

Adoptive T-cell therapy (ACT) using tumor-infiltrating lymphocytes (TIL) can induce tumor regression in up to 50% or more of patients with unresectable metastatic melanoma. However, current methods to expand melanoma TIL, especially the "rapid expansion protocol" (REP) were not designed to enhance the generation of optimal effector-memory CD8(+) T cells for infusion. One approach to this problem is to manipulate specific co-stimulatory signaling pathways to enhance CD8(+) effector-memory T-cell expansion. In this study, we determined the effects of activating the TNF-R family member 4-1BB/CD137, specifically induced in activated CD8(+) T cells, on the yield, phenotype, and functional activity of expanded CD8(+) T cells during the REP. We found that CD8(+) TIL up-regulate 4-1BB expression early during the REP after initial TCR stimulation, but neither the PBMC feeder cells in the REP or the activated TIL expressed 4-1BB ligand. However, addition of an exogenous agonistic anti-4-1BB IgG4 (BMS 663513) to the REP significantly enhanced the frequency and total yield of CD8(+) T cells as well as their maintenance of CD28 and increased their anti-tumor CTL activity. Gene expression analysis found an increase in bcl-2 and survivin expression induced by 4-1BB that was associated with an enhanced survival capability of CD8(+) post-REP TIL when re-cultured in the absence or presence of cytokines. Our findings suggest that adding an agonistic anti-4-1BB antibody during the time of TIL REP initiation produces a CD8(+) T cell population capable of improved effector function and survival. This may greatly improve TIL persistence and anti-tumor activity in vivo after adoptive transfer into patients.

Figure 1. Addition of anti-4-1BB antibody to the REP increased in the percentage of CD8⁺ TIL.

Anti-4-1BB antibody was added to the TIL on day 0 of the REP at the indicated concentrations. All other conditions were the same in each culture. After 14 days the cells were harvested and stained for CD4 and CD8 expression and viable cell counts were done using a hemocytometer following Trypan Blue staining. 4-1BB co-stimulation during the REP increased the frequency of CD8⁺ T cells in a dose-dependent fashion in a representative TIL lines #2354 and #2199 (A). The total yield of CD8⁺ T cells after the REP with different doses of anti-4-1BB is shown in two independent TIL lines (#2354 and #2199) (B). A dose-dependent increase in CD8⁺ T-cell yield was noted, with 500ng/ml of anti-4-1BB antibody being optimal. The results of triplicate cell counts ± standard deviation are shown.

Figure 2. Addition of 4-1BB antibody in the REP increases CD8⁺ T-cell frequency and yield in a large cohort of patient samples (n = 34).

The effects of addition of an optimal dose of anti-4-1BB (500 ng/ml), as determined previously, were tested in rapid expansions of TIL from 34 separate patient tumors. Addition of anti-4-1BB significantly increased the frequency of CD8⁺ T cells in the final TIL product in the patient population (A). CD8⁺ T-cell yield (B), and the fold expansion of CD8⁺ cells (C), was significantly increased when anti-4-1BB was added to the REP over the patient population. However, 4-1BB stimulation did not alter the total T-cell yield (B) or the total TIL fold expansion (C). Statistical analysis was done using the Wilcoxon signed rank test using biological relevance occurring when p<0.05.

Figure 3. Addition of anti-4-1BB antibody maintained CD28 expression on CD8⁺ T cells during the REP in a large cohort of patient samples (n = 34).

TIL from a representative patient was subjected to the REP with or without the addition of anti-4-1BB antibody and the cells stained for CD8 and CD28 (A) or CD8 and CD27 (B) and analyzed by flow cytometry both before the REP (pre-REP) and after the control (IL-2) REP or REP with anti-4-1BB (IL-2+4-1BB). Viable cells were gated and the frequency of CD28⁺ (A) or CD27⁺ (C) in the CD8⁺ subset analyzed. The change in CD28 or CD27 expression in the CD8⁺ T-cell subset during the REP with or without added anti-4-1BB was determined in 34 independent TIL lines (B and D). The left hand panels in B and D show the overall median difference in the frequency of CD28⁺ or CD27⁺ cells in the CD8⁺ subset over the entire patient population. The right hand panels in B and D show the fold expansions in the CD28⁺ or CD27⁺ populations. Statistical analysis was done using Wicoxon signed rank sum test using biological relevance occurring when p<0.05.

Figure 4. Addition of anti-4-1BB antibody increased Granzyme B (GB) and Perforin (Perf) expression in CD8⁺ T cells during the REP in multiple patient samples.

TIL from a representative patient was subjected to the REP with or without the addition of anti-4-1BB antibody and the cells stained for CD8 and Perf (A) or CD8 and GB (B) and analyzed by flow cytometry both before the REP (pre-REP) and after the control (IL-2) REP or REP with anti-4-1BB (IL-2+4-1BB). Viable cells were gated and the frequency of Perf⁺ (A) or GB⁺ (C) in the CD8⁺ subset analyzed. The change in Perf or GB expression in the CD8⁺ T-cell subset during the REP with or without added anti-4-1BB was determined in larger cohort of patients TIL samples (B and D). 4-1BB co-stimulation during the REP significantly increased Perf expression in CD8⁺ T cells in 34 separate TIL line tested (B) and increased GB expression in 11 separate TIL lines tested (D). Statistical analysis was done using Wicoxon signed rank sum test using biological relevance occurring when p<0.05.

Figure 5. TIL isolated after rapid expansion with anti-4-1BB antibody displayed an increased ability to secrete IFN-γ, TNF-α, and IL-2 after TCR re-stimulation.

Melanoma TIL rapidly expanded with or without the addition of the anti-4-1BB antibody were re-stimulated with anti-CD3 antibody in 96-well plates. Culture supernatants were collected after 24 hours and assayed using anti-cytokine beads for IFN-γ, TNF-α, and IL-2 using a Luminex-100 system. Results from 3 different TIL lines comparing the control (IL-2) group with the IL-2+4-1BB group are shown for IFN-γ (A) TNF-α (B), and IL-2 (C). In each case the net production of the cytokines was calculated by subtracting the control wells (no anti-CD3) from the wells that had anti-CD3. The averages and standard deviation of triplicate wells are shown in each case. A paired student’s t-test was used to determine statistical significance between groups with p<0.05 indicating statistical significance.

Figure 6. Addition of anti-4-1BB antibody to the REP led to increased post-REP TIL tumor antigen-specific CTL activity.

Melanoma TIL from HLA-A0201⁺ patients with a significant population of CD8⁺MART-1 tetramer⁺ cells were rapidly expanded with or without anti-4-1BB as before. The post-REP TIL were sorted by FACS for CD8⁺ T cells and assayed for tumor antigen-specific CTL activity using a flow-cytometry-based assay that measures caspase-3 cleavage in target cells. The results of three different patient TIL lines are shown. The top panels (A) show the CTL activity of TIL #2292 using the melanoma cell line 624 (HLA-A0201⁺) and the control HLA-A-unmatched line 938 as targets (left side), or MART-1 peptide-pulsed T2 target cells as targets (right side). The bottom panels (B) show the CTL activity of two other HLA-A0201⁺ TIL lines (#2276 and #2122) against 624 or 938 cells with similar results. In all cases (A and B) the levels of non-specific killing were markedly lower.

Figure 7. Increased expression of anti-apoptotic molecules and improved survival in post-REP TIL that received 4-1BB co-stimulation during the REP.

RNA from post-REP TIL that were rapidly expanded with or without anti-4-1BB were subjected to real-time PCR (qRT-PCR) analysis for bcl-2, bcl-xL, bim, and survivin (A). The results of two representative patient TIL lines are shown (#2392 and #2396). Each PCR reaction was ran in triplicate with a CV of <5% for all assays. The results with the IL-2+4-1BB REP were normalized against the levels of gene expression in the control (IL-2) REP, as indicated with the 1-fold expression for the control REP for each gene. The increase in bcl-2 expression in the IL-2+4-1BB group was confirmed in three different TIL samples by intracellular staining for bcl-2 using flow cytometry (B). The post-REP TIL were also analyzed for their ability to survive and further expand when re-cultured with or without IL-2 (100 U/ml) for 5 days, as described in the Materials and Methods (C and D). Viable cell counts together with flow cytometry analysis for CD8 expression was performed to calculate the fold change in CD8⁺ T cells after the re-culturing period (C). TIL from 4-1BB REP (IL-2+4-1BB) or control REP (IL-2) cultures from 4 different patients were also subjected to staining with Annexin V and 7-AAD after re-plating without or with 200 IU/ml IL-2 for 5 days and the frequency of Annexin V⁺7-AAD⁻ (apoptotic cells) were determined (D).

Figure 8. CD8⁺ MART-1-reactive TIL provided 4-1BB co-stimulation during the REP have a greater proliferative response to re-stimulation with MART-1 peptide.

TIL from HLA-A0201⁺ patients were rapidly expanded with or without added anti-4-1BB antibody. The cells were labeled with eFluor670 dye and re-stimulated with HLA-A0201⁺ MART-1 peptide-pulsed DC, as previously described . The post-REP TIL were labeled with the cell division dye eFluor670 (Invitrogen) and re-stimulated with HLA-A0201-matched mature DC pulsed with MART-1 peptide for 7 days. The cells were harvested and stained for CD8 and MART-1 tetramer and analyzed by flow cytometry for the frequency of divided (eFluor670^low) in the gated CD8⁺ MART-1 tetramer⁺ subset. One representative experiment out of two is shown. As shown in A, CD8⁺ TIL isolated from 4-1BB costimulated REP cultures (#2183) had an enhanced response to MART-1 peptide re-stimulation, as shown by the increased number of cell divisions measured by eFluor670 dilution in the CD8⁺ MART-1 tetramer⁺ gated cells. TIL isolated from REP cultures that received anti-4-1BB also exhibited an increase in the fold expansion of CD8⁺ MART-1 tetramer⁺ cells following re-stimulation of the post-REP TIL with MART-1 peptide pulsed DC as shown in experiments with two separate TIL lines (#2183 and #2559) (B).