Adoptive immunotherapy using tumor-reactive T lymphocytes has emerged as a powerful approach for the treatment of bulky, refractory cancer , however the ability to generate large numbers of TILs for therapy is a challenge that has significant regulatory hurdles, and requires technically sophisticated cell processing and extended in vitro lymphocyte culturing periods. Long-term culture of tumor-derived T cells in high-dose interleukin-2 (IL-2) allows for the generation of high numbers of TILs (>1 × 1011) but with preferential expansion of CD4+ lymphocytes [2–4]. Initial IL-2-based TIL expansion followed by a "rapid expansion method" (REM) [5–9] is a more time and labor efficient method, requiring an excess of irradiated allogeneic peripheral blood mononuclear cells (PBMC) as feeder cells, anti-CD3 antibody and high doses of IL-2, that can result in a 1,000-fold expansion of TILs over a 14-day period . While routinely used, the REM has introduced technical, regulatory, and logistic challenges that have prevented larger and randomized clinical trials as a prelude to widespread application. First, large numbers of allogeneic feeders (200-fold excess), often from multiple donors, are required for clinical expansions. Second, allogeneic feeder cells harvested by large-volume leukapheresis from healthy donors exhibit donor to donor variability in their viability after cryopreservation and capacity to support TIL expansion, and thus test expansions are often required. Finally, this process necessitates additional extensive and costly laboratory testing of each individual donor cell product to confirm sterility.
Artificial antigen presenting cells (aAPCs) expressing ligands for the T cell receptor and costimulatory molecules can activate and expand T cells for transfer, while improving their potency and function. The first generation of aAPC consisted of anti-CD3 and anti-CD28 monoclonal antibodies (mAbs) covalently bound to magnetic beads (CD3/CD28 beads) which crosslink CD3 and CD28 on T cells, enabling efficient polyclonal expansion of circulating T cells (50 to 1000-fold) over 10-14 days of ex vivo culture with preferential expansion of naïve and memory CD4+ T cells , however their efficiency in TIL expansion has not been examined. Second generation cell-based aAPCs can substitute for natural APCs, mediate efficient expansion of antigen-specific T cells from peripheral blood [11–16] and stably express multiple gene inserts, including CD64 (the high-affinity Fc receptor), CD32 (the low-affinity Fc receptor), and CD137L (4-1BBL), among others [13, 15]. Compared to beads, cell-based aAPCs bearing the costimulatory ligand CD137L can more efficiently induce the proliferation of antigen-experienced CD8+ CD28- T cells from peripheral blood and improve their in vivo persistence and antitumor activity upon adoptive transfer to tumor-bearing mice [15, 17]. In these studies, enhanced proliferation of antigen-experienced CD8+ CD28- T cells mediated by aAPCs is dependent on CD137 ligation [15, 17].
Unlike peripheral blood lymphocytes (PBL), most tumor antigen-specific CD8+ TILs derived from solid tumors express low levels of CD28 [18, 19]. Together, the above studies suggest that approaches utilizing CD137 ligation may support ex vivo TIL expansion. In a trial of adoptive TIL transfer with REM generated cells, the persistence of TILs in vivo after infusion represented a major limitation to successful therapy . In vivo persistence and clinical response were both associated with expression of the costimulatory molecules CD28 and CD27 by TILs, as well as their telomere length [18, 21–24]. The REM requires extended duration TIL culture which results in telomere length shortening and reduced expression of CD28 and CD27 [18, 25], thus there remains a need for the development of improved, standardized methods and materials for generating TILs rapidly for adoptive transfer with greater potency and engraftment capability.
Here we investigate the use of engineered K562 cell-based aAPCs as an "off-the-shelf" platform for ex vivo TIL expansion. K562 aAPCs that express CD137L offer the potential to expand antigen-experienced TILs and represent a potential new cell-based platform for the standardization of ex vivo TIL expansion. Ovarian cancer and melanoma biospecimens were used to test the notion that aAPC can stimulate TIL expansion in different tumor histotypes [26, 27], based on the knowledge that TILs from these cancers can recognize autologous tumor as well as known tumor antigens in vitro [28–32], and exhibit tumor-specific reactivity ex vivo [33, 34] and in vivo [5, 7, 35]. We found that aAPCs efficiently expand IL-2 cultured TILs from solid tumor specimens of ovarian cancer similar to the REM, resulting in a favorable CD4/8 T cell ratio, and low FOXP3+ CD4 T cell composition. aAPC-based TIL expansion depends on the provision of exogenous IL-2 cytokine support in culture and is largely CD28-independent. Under these conditions, tumor antigen-specific TILs with demonstrated anti-tumor reactivity can be expanded. Further, aAPC can induce the rapid and efficient expansion of TILs directly from freshly digested tumor samples, reducing overall culture time, and output TILs are highly skewed in CD8+ lymphocyte composition, possess high levels of CD28 and CD27 expression after activation and are amenable to secondary aAPC-based expansion. The aAPC platform as described here thus establishes a standardized methodology for the rapid, clinical-grade expansion of TILs for therapy.