Alteration in interactions between tumor-infiltrating lymphocytes and tumor cells in human melanomas after chemotherapy or immunotherapy

Alteration in interactions between tumor-infiltrating lymphocytes (TILs) and tumor cells after chemotherapy or immunotherapy was studied in metastatic melanoma patients. Tumors were harvested from surgical specimens 17 days after the end of chemotherapy with cisplatin, vinblastine, and dacarbazine (CVD). Tumors of nonlymph-node metastases from two responders yielded neither TILs nor tumor cells, whereas those from all four nonresponders had both TILs [(1.1-13.8) × 10⁶ cells/g tumor] and tumor cells [(2.8-30.8) × 10⁶ cells/g tumor). Tumors of lymph node metastases from nine patients yielded substantial numbers both of TILs and tumor cells, regardless of different clinical responses, except with one complete responder, whose tumor did not contain tumor cells. The mean increase of TILs from these tumors (n = 14) 3-4 weeks after incubation with 200 U/ml recombinant interleukin-2 (rIL-2) was 2.5-fold, whereas there was a 56-fold increase in TILs from untreated tumors (n = 3). CD3⁺ T cells predominated in TILs before and after expansion with IL-2. IL-2-activated TILs from five of six tumors tested displayed higher cytotoxicity against autologous tumor cells than against cells from any of three allogeneic tumors. Mean tumor cell numbers (10⁶ cells/trial) obtained by serial needle biopsies for the same tumor in five patients decreased from 1.2 before therapy to 0.25 at day 4 of therapy (interferon α alone), and to 0.02 at day 8 (interferon α and IL-2). This decrease did not correlate with clinical responses. Yields (× 10⁶ cells/g tumor) of TILs and tumor cells in subcutaneous melanomas obtained by excisional biopsies in one nonresponder under IL-2 therapy were respectively 0.2 and 1.1 before therapy (day 0), 0.1 and <0.01 during (day 7), 0.2 and <0.01 at the end of therapy (day 21), and 0.5 and 0.5 at the time of tumor progression (day 66). Yields of TILs and tumor cells in the other nonresponder were respectively 3 and 26 before (day 0), 16 and 3 during (day 7), and 0.4 and <0.01 at the end of IL-2 therapy (day 17), and 2.5 and 6 at the time of progression (day 62). TILs in these two patients before therapy proliferated well in culture with IL-2 (570-and 720-fold, respectively), and showed higher cytotoxicity against autologous tumor cells, whereas none of those from the five tumors biopsied during or at the end of IL-2 therapy proliferated. TILs at the time of progression showed modest proliferation (54- and 76-fold, respectively) and showed major-histocompatibility-complexnonrestricted cytotoxicity. In summary, a decrease in the number of live tumor cells did not always correlate with clinical response in either therapy. CVD chemotherapy may simply impair IL-2-induced proliferation of TILs. IL-2 therapy may induce transient unresponsiveness of TILs to IL-2.