Expansion of tumor-infiltrating lymphocytes and their potential for application as adoptive cell transfer therapy in human breast cancer

Abstract

Adoptive cell transfer (ACT) of ex vivo expanded tumor-infiltrating lymphocytes (TILs) has been successful in treating a considerable proportion of patients with metastatic melanoma. In addition, some patients with several other solid tumors were recently reported to have benefited clinically from such ACT. However, it remains unclear whether ACT using TILs is broadly applicable in breast cancer, the most common cancer in women. In this study, the utility of TILs as an ACT source in breast cancers was explored by deriving TILs from a large number of breast cancer samples and assessing their biological potentials. We successfully expanded TILs ex vivo under a standard TIL culture condition from over 100 breast cancer samples, including all breast cancer subtypes. We also found that the information about the percentage of TIL and presence of tertiary lymphoid structure in the tumor tissues could be useful for estimating the number of obtainable TILs after ex vivo culture. The ex vivo expanded TILs contained a considerable level of central memory phenotype T cells (about 20%), and a large proportion of TIL samples were reactive to autologous tumor cells in vitro. Furthermore, the in vitro tumor-reactive autologous TILs could also function in vivo in a xenograft mouse model implanted with the primary tumor tissue. Collectively, these results strongly indicate that ACT using ex vivo expanded autologous TILs is a feasible option in treating patients with breast cancer.

Keywords: adoptive cell transfer; breast cancer; function of TIL; memory T cell; tumor-infiltrating lymphocyte.

Figure 1. Tumor-infiltrating lymphocytes (TILs) from breast cancer tissues can be successfully expanded after 2 weeks’ ex vivo culture

(A) Percentage (%) of TILs and (B) degree (score) of tertiary lymphoid structure (TLS) in HR⁺/HER2⁻ (n = 95), HR⁺/HER2⁺ (n = 26), HR⁻/HER2⁺ (n = 20), and TNBC (n = 56) breast cancer tissues. The degree of TLS was expressed as the following scores: 0, none; 1, little; 2, moderate; 3, abundant. (C) Number of TILs (per fragment) obtained after 2 weeks’ ex vivo culture from HR⁺/HER2⁻ (n = 83), HR⁺/HER2⁺ (n = 26), HR⁻/HER2⁺ (n = 17), and TNBC (n = 56) breast cancer tissues. (D) Number of TILs (per fragment) obtained after 2 weeks’ ex vivo culture from breast cancer tissues of patients treated with neoadjuvant chemotherapy (NAC) (n = 49) or without NAC (No NAC, n = 133). Kruskal–Wallis test and Mann–Whitney U-test were used for statistical analysis. ^*p < 0.05, ^**p < 0.01, ^***p < 0.001. HR, hormone receptor; TNBC, triple-negative breast cancer.

Figure 2. Correlation between the number of 2 week TILs and TIL percentage or TLS degree in breast cancer tissues

(A) The number of TILs per fragment (y-axis) obtained after 2 weeks’ ex vivo culture from breast cancer tissues was plotted against histologically determined tissue TIL percentage (x-axis): Tissue TILs: 1% (n = 29), 2% (n = 5), 5% (n = 45), 10% (n = 26), 20% (n = 19), 30% (n = 22), 40% (n = 4), 50% (n = 9), 60% (n = 2), 70% (n = 6), 80% (n = 10), and 90% (n = 3). Tumor samples were grouped as follows for statistical analysis: TILs ≤5%, 10%≤ TIL ≤30%, 40%≤ TIL ≤60%, or 70% ≤ TIL. (B) The number of TILs per fragment (y-axis) obtained after the 2 week culture was plotted against tissue TLS degree (x-axis) grouped as follows: Tissue TLS degree, none (n = 42), little (n = 35), moderate (n = 56), and abundant (n = 47). (C) The average number of TILs per fragment (z-axis) obtained after the 2 week culture was plotted in the space defined by both percentage of tissue TILs (x-axis) and TLS degree (y-axis) in the same sample: 1%≤ tissue TILs ≤5% and TLS none (n = 36); 1%≤ tissue TILs ≤5% and TLS little (n = 19); 1%≤ tissue TILs ≤5% and TLS moderate (n = 19); 1%≤ tissue TILs ≤5% and TLS abundant (n = 5); 10%≤ tissue TILs ≤30% and TLS none (n = 4); 10%≤ tissue TILs ≤30% and TLS little (n = 14); 10%≤ tissue TILs ≤30% and TLS moderate (n = 32); 10%≤ tissue TILs ≤30% and TLS abundant (n = 17); 40%≤ tissue TILs ≤60% and TLS none (n = 1); 40%≤ tissue TILs ≤60% and TLS little (n = 2); 40%≤ tissue TILs ≤60% and TLS moderate (n = 3); 40%≤ tissue TILs ≤60% and TLS abundant (n = 9); 70%≤ tissue TILs and TLS none (n = 1); 70%≤ tissue TILs and TLS little (n = 0); 70%≤ tissue TILs and TLS moderate (n = 2); and 70%≤ tissue TILs and TLS abundant (n = 16). Kruskal–Wallis test and Mann–Whitney U-test were used for statistical analysis. ^*p < 0.05, ^**p < 0.01, ^***p < 0.001. (D) Representative examples of H&E sections showing low numbers of TILs (5%) and little TLS (left panels: sample BC16021, 18.7 × 10⁵ TILs/fragment obtained after 2 weeks’ culture), and high numbers of TILs (70%) and no TLS (right panels: sample BC16124, 17.7 × 10⁵ TILs/fragment obtained after culture). Arrows indicate TILs, and the arrowhead indicates TLS.

Figure 3. Ex vivo TIL culture enriches memory phenotype CD8⁺ T cells

Analysis of T cells and their subsets (A, B), and memory phenotype T cells (C, D) in TILs. Single cells directly derived from fresh breast cancer tissues (fresh, n = 33), TILs obtained after 2 weeks’ ex vivo culture of breast cancer tissue fragments (2 wks, n = 132), and post-REP TILs (REP, n = 45) were analyzed by flow cytometry. (A) Representative FACS data showing percentage of T cells (CD3⁺) among live cells, percentage of CD4⁺ T cells (CD3⁺CD8⁻) and of CD8⁺ (CD3⁺CD8⁺) T cells among T cells, and (C) percentages of memory phenotype CD4⁺ T cells (CD3⁺CD8⁻CD45RO⁺) and CD8⁺ T cells (CD3⁺CD8⁺CD45RO⁺) among CD4⁺ T cells (CD3⁺CD8⁻) and CD8⁺ T cells (CD3⁺CD8⁺), respectively. (B, D) FACS results expressed as the mean ± SD summarized as histograms. Kruskal–Wallis test and Mann–Whitney U-test were employed for statistical analysis. ^***p < 0.001.

Figure 4. Ex vivo TIL culture does not compromise the proportion of the central memory phenotype T cell subset

Memory CD4⁺ (A, B) and CD8⁺ (C, D) T cell subpopulations among single cells directly derived from fresh breast cancer tissue (fresh, n = 33), those among TILs obtained after the 2 week ex vivo culture of breast cancer tissue fragments (2 wks, n = 132), and those among post-REP TILs (REP, n = 45) were analyzed by flow cytometry. Representative FACS data showing percentage of memory stem T cells (T_SCM, CD45RO⁻CCR7⁺CD62L⁺CD95⁺), central memory phenotype T cells (T_CM, CD45RO⁺CCR7⁺), or effector memory T cells (T_EM, CD45RO⁺CCR7⁻) among CD4⁺ (A) and CD8⁺ (C) T cells. (B, D) FACS results expressed as the mean ± SD summarized as histograms. Kruskal–Wallis test and Mann–Whitney U-test were employed for the statistical analysis. ^***p < 0.001.

Figure 5. Post-REP TILs are functional and retain reactivity towards autologous primary tumor cells cultured in vitro

Two week cultured TILs were further expanded for 2 weeks under a standard REP condition and tested for functional activity in vitro. (A) PMA/ION response: 1 × 10⁵ post-REP TILs (from each of 15 randomly selected samples; sample IDs are shown) were seeded in 96-well plates and treated with or without PMA (32.4 nM)/ION (1 μg/ml). After 24 h, cell culture supernatants were collected and the IFNγ level was measured by ELISA (n = 3 or 4 per group). Open bars indicate no treatment (control), and filled bars indicate PMA/ION treatment (PMA/ION). (B) The reactivity of TILs to autologous primary cancer cells. Post-REP TILs from six selected samples were co-cultured with (filled bar) or without (open bar) the autologous primary tumor cells (1 × 10⁵) at an effector:target cell ratio of 1:1, 2:1, or 4:1. After 24 h, cell culture supernatants were collected and the IFNγ level was measured by ELISA (n = 3 or 4 per group). A non-paired Student’s t-test was performed. ^*p < 0.05, ^**p < 0.01, ^***p < 0.001.

Figure 6. In vitro tumor-reactive post-REP TILs are functional in a PDX mouse model in vivo

PDX mice were created by implanting tumor tissues derived from each of two patient samples (from a lymph node with metastatic breast cancer, sample BC16110 (A), and from breast, sample BC16158 (B)). When tumors in the PDX mice were palpable, 1 × 10⁷ post-REP TILs (TIL group, n = 4 for BC16110 and n = 5 for BC16158) or PBS (control group, n = 4 for BC16110 and n = 5 for BC16158) was administered into the tail vein, and IL-2 (200,000 IU/mice) and IL-15 (1 μg/mice) were administered simultaneously into the peritoneal cavity of the mice in both groups. Then, IL-2 (200,000 IU/mice) was administered for the following 2 days (total of three times) and IL-15 (1 μg/mice) was administered every other day for 8 days (total of five times). After administration of TILs (day 0), tumor size was measured using calipers. Tumor volumes were calculated using the formula ½ × length × (width)² and expressed as the mean ± SD. A non-paired Student’s t-test was performed. ^*p < 0.05, ^**p < 0.01.