In vitro predictors of therapeutic response in melanoma patients receiving tumor-infiltrating lymphocytes and interleukin-2.

PURPOSE: To correlate in vitro characteristics of tumor-infiltrating lymphocytes (TIL) with clinical response to TIL immunotherapy in patients with metastatic melanoma. PATIENTS AND METHODS: Forty-one melanoma patients undergoing 43 separate treatment courses with TIL and interleukin-2 (IL-2) from December 1990 through November 1992 were studied prospectively. Multiple patient and treatment characteristics were evaluated for response correlates. In addition, TIL were assayed within 7 days of infusion for characteristics such as doubling time, cell-surface phenotype, autologous tumor lysis in 4-hour chromium-51 release assays, and cytokine secretion following autologous tumor stimulation. RESULTS: Nine patients experienced complete or partial tumor regressions. Clinical parameters such as age, sex, sites of disease, performance status, and prior therapies were similar in responders and nonresponders. Treatment variables such as the cumulative IL-2 dose and concomitant administration of cyclophosphamide or interferon (IFN)-alpha were not predictive of response, although responders received 33% more TIL. However, statistically significant differences in favor of clinical response were noted for extranodal source of TIL (v lymph node), shorter culture duration (mean, 38 v 47 days), shorter TIL doubling time (2.6 v 3.7 days), greater autologous tumor lysis by TIL (30% v 15%; effector-to-target [E:T], 40:1), and secretion of granulocyte-macrophage colony-stimulating factor (GM-CSF) by TIL following autologous tumor stimulation (six of nine responders v eight of 32 nonresponders). CONCLUSION: The associations of TIL lysis of autologous tumor and younger TIL age with clinical response observed in this study are supportive of previous reports, and these findings will be useful in designing future clinical trials. The new observation correlating GM-CSF secretion by TIL with clinical response is interesting and needs further substantiation.

Hairy cells possess more interferon receptors than other lymphoid cell types.

To investigate the possible direct effects of alpha-interferon (IFN-alpha) in hairy cell leukemia, IFN-alpha receptor expression by hairy cells (11 cases) was measured by a radiolabeling technique and compared with that of MOLT-4, chronic lymphocytic leukemia (15 cases), and various other leukemic and normal cell types. Purified peripheral blood and splenic hairy cells showed higher levels of receptor expression (approximately 1,000 +/- 200 binding sites/cell; 11 cases tested) than other normal and leukemic cell types. B cells from normal blood and tonsils showed low levels of receptors (approximately 120 +/- 100 binding sites/cell), while a range of B cell leukemias displayed intermediate levels of expression (100-500 sites/cell). In the 15 cases of chronic lymphocytic leukemia tested, 530 +/- 330 binding sites/cell) were demonstrated, the high standard deviation reflecting the fact that one third of cases had receptor levels comparable with those in hairy cell leukemia. Normal and hairy cell leukemia T cells, red cells, and platelets had no demonstrable IFN-receptors. These findings may be relevant to the efficacy of IFN in hairy cell leukemia.

The mechanism of action of interferon-alpha (IFN-alpha) in hairy-cell leukaemia; Hu-IFN-alpha 2 receptor expression by hairy cells and other normal and leukaemic cell types.

To investigate the possible direct effects of interferon-alpha (IFN-alpha) in hairy-cell leukaemia, IFN-alpha receptor expression by hairy cells (HCs) (11 cases) was measured by a radiolabelling technique and compared with that of MOLT-4, chronic lymphocytic leukaemia (CLL; 14 cases) and various other leukaemic and normal cell types. Purified peripheral blood (PB) and splenic HCs showed higher levels of receptor expression (approx. 1000 +/- 200 binding sites/cell; 11 cases tested) than other normal and leukaemic cells types. Purified normal PB and tonsil B cells showed low levels of receptors (approx. 120 +/- 100 binding sites/cell), while a range of B-cell leukaemias displayed intermediate levels of expression (approx. 100-500 sites/cell). In the 15 cases of CLL tested, 530 +/- 330 binding sites/cell were demonstrated, the high standard deviation reflecting the fact that approximately one third of cases had receptor levels comparable with those in HCL. Normal and HCL T cells, red cells and platelets had no demonstrable IFN receptors. It is suggested that these findings may be relevant to the efficacy of IFN in hairy-cell leukaemia.

Heterogeneity of CLL: high CD23 antigen and alpha IFN receptor expression are features of favourable disease and of cell activation.

The relationship between alpha-interferon receptor (alpha IFNR) numbers, B-cell-antigen expression and clinical stage was determined in 35 cases of typical chronic lymphocytic leukaemia (CLL). alpha IFNR numbers were shown to be low in patients with advanced disease and high in those with a more favourable prognosis. The B-cell activation antigen CD 23 was similarly related to stage, being high in more favourable disease. Also, alpha IFNR expression was directly related to CD 23 positivity, but alpha IFN binding was not inhibited by CD 23 monoclonal antibody. There was no correlation between CD 19, 20, 22 and 24 antigen expression and either alpha IFNR numbers or clinical stage. Since CD23 antigen expression is a feature of B-cell activation, we suggest that high CD23 and alpha IFNR positivity are manifestations of an activated cell phenotype and that cell activation in CLL is a feature of favourable disease.

Interleukin-7 selectively enhances natural kill cytotoxicity mediated by the CD56bright natural killer subpopulation.

Both the CD56bright and CD56dim NK cell subpopulation mediate non-major histocompatibility complex-restricted cytolysis of NK-sensitive tumor cell lines, and IL-2-dependent augmentation of cytolysis and proliferation of CD56bright and CD56dim NK cells was recently reported. We investigated the effects of IL-7 and IL-6 on the killing mediated by these cells to determine whether other cytokines besides IL-2 regulate their activity. IL-7 increased the cytotoxicity in only the CD56bright NK cell population. The effect of IL-7 varied from donor to donor but was comparable to that of IL-2. Furthermore, IL-7 was found to induce lymphokine-activated killer (LAK) cell generation primarily in the CD56bright cells. CD56bright NK cells also proliferated in response to IL-7, but only weakly in comparison with IL-2. In contrast to the results with CD56bright NK cells, IL-7 had little effect on the CD56dim subset. However, IL-2 enhanced NK cytotoxicity, induced LAK activity, and caused proliferation of these cells. An anti-IL-2 antibody did not inhibit the IL-7-induced increase in CD56bright cytotoxicity, suggesting that IL-7 acted independently of IL-2. However, the IL-7 effect on CD56bright NK cell cytotoxicity was partially inhibited by anti-CD2, anti-CD11a, and anti-CD18 antibodies and almost completely abrogated by a combination of anti-CD2 and anti-CD11a. These data suggest that cell adhesion molecules (CAM) play a role in the regulation of IL-7-induced CD56bright NK cell cytolysis. In contrast to IL-7-mediated effects, IL-6 alone had no effect on CD56+ NK cell cytotoxicity.(ABSTRACT TRUNCATED AT 250 WORDS)

CD4+ T lymphocytes infiltrating human breast cancer recognise autologous tumor in an MHC-class-II restricted fashion.

Tumor-infiltrating lymphocytes (TIL) were derived from primary breast tumors, metastatic lymph nodes and malignant pleural effusions from 34 patients with breast cancer. TIL were cultured for approximately 30 days and studied for phenotype, cytotoxicity, and the ability to secrete cytokines in response to autologous tumor stimulation. Tumor specimens were obtained from two different sites in 7 patients, resulting in 41 samples from which 38 TIL cultures were established. In addition to screening 38 bulk TIL cultures, TIL from 21 patients were separated into CD4+ and CD8+ subsets and extensively studied. Three CD4+ TIL were found specifically to secrete granulocyte macrophage-colony-stimulating factor and tumor necrosis factor alpha when stimulated by autologous tumor and not by a large panel of stimulators (24-34) consisting of autologous normal cells, allogeneic breast or melanoma tumors and EBV-B cells. This cytokine release was found to be MHC-class-II-restricted, as it was inhibited by the anti-HLA-DR antibody L243. These 3 patients' EBV-B cells, when pulsed with tumor lysates, were unable to act as antigen-presenting cells and induce cytokine secretion by their respective CD4+ TIL. These findings demonstrate that MHC-class-II-restricted CD4+ T cells recognising tumor-associated antigens can be detected in some breast cancer patients.

Association between clonogenic cell growth and clinical risk group in B-cell chronic lymphocytic leukemia.

Chronic lymphocytic leukemia of B-cell origin (B-CLL) is a disease with a variable clinical course, despite the fact that the neoplastic cells in this disorder are homogeneous with respect to morphology, immunophenotype, and cell cycle stage. To further investigate the heterogeneity observed in the clinical behavior of B-CLL, we determined the phenotype and growth requirements of clonogenic cells from 28 patients with B-CLL from low-, intermediate-, and high-risk groups as defined by the Rai staging system. Using methyl-cellulose as a semi-solid media with feeder cells and/or growth factors, colonies were observed with one or more of the culture conditions tested in 25 of 28 CLLs. Phenotypic analysis of colonies demonstrated that the clonogenic cells uniformly expressed la, CD19, CD20, CD5, and the identical light chain as the original CLL cell cultured. However, heterogeneity was observed in clonogenic B-CLL cell growth among the three different CLL risk groups. Clonogenic cells from patients with low-risk CLL required either irradiated unstimulated T cells, with or without conditioned media (CM) or irradiated activated T cells alone for colony formation. Both the number of colonies (227 +/- 15) as well as the number of cells per colony (220 +/- 82) were large, with a mean cloning efficiency of 0.39%. In contrast, clonogenic cells from patients with intermediate- and high-risk CLL required the combination of both irradiated activated T cells and CM. As compared with the low-risk CLLs, both the number and size of the colonies formed by the intermediate- (74 +/- 17, 70 +/- 39) and high- (83 +/- 28, 40 +/- 14) risk groups were significantly lower (P less than .0001). Similarly, the mean cloning efficiency was significantly reduced to 0.15% and 0.14%, respectively. None of the recombinant cytokines (interleukin 1 [IL-1] to IL-7, tumor necrosis factor, alpha and gamma-interferon, B-cell growth factor, and granulocyte macrophage colony-stimulating factor) alone or in combination with each other could entirely replace the stimulatory effect of the activated T cells. These data suggest that clinical progression of B-CLL is associated with a loss of clonogenic potential in the circulating pool of neoplastic cells, which require as yet undefined factors provided by activated T cells and CM.

Tumor-infiltrating lymphocytes from human ovarian cancer patients recognize autologous tumor in an MHC class II-restricted fashion.

PURPOSE: To search for tumor-specific in vitro reactivity by lymphocytes derived from patients with ovarian carcinoma. METHODS: Tumor-infiltrating lymphocytes (TIL) were derived from primary or metastatic solid tumors and tumor-associated lymphocytes (TAL) were derived from ascites from 13 patients with ovarian cancer. TIL or TAL were cultured for approximately 30 to 60 days and studied for phenotype, cytotoxicity, and cytokine secretion in response to autologous tumor stimulation. RESULTS: Twenty-nine bulk TIL or TAL cultures were successfully established from 10 patients using various culture conditions. Thirteen cultures were predominantly CD4+ and 16 were mainly CD8+. In contrast to reports by others, none of the cultures tested were specifically lyric for autologous tumor. Five predominantly CD4+ bulk TIL (from four patients) preferentially secreted tumor necrosis factor-alpha and granulocyte macrophage-colony stimulating factor when stimulated with autologous tumor and not when stimulated by autologous Epstein Barr virus-B cells, fibroblasts, peripheral blood mononuclear cells, or allogeneic HLA matched or mismatched stimulators. This cytokine secretion was found to be MHC class-II restricted in three patients because it was inhibited by the anti-MHC class-II antibody IVA12 and the HLA-DR specific antibody L243. CONCLUSION: We believe these data are the first to suggest that tumor reactive CD4+ lymphocytes exist in some ovarian cancer patients. This finding may be useful in the development of novel immunotherapies for these patients.