Characterization of the Cytolytic Activity of CD4+ and CD8+ Tumor-infiltrating Lymphocytes in Human Renal Cell Carcinoma1

Previously we showed that IL2 expanded tumor-infiltrating lympho cytes (Ills) from renal cell carcinoma mediated non-major histocompatibility complex-restricted cytotoxicity. Phenotypic analysis showed that cultured TILs were composed mostly of T-lymphocytes with varying numbers of CD4+, CDS*, and CD56* (Leul9*) populations. Here we compared the cytolytic activity of the two predominant TIL subsets, CD3*CD4* and CD3*CD8*, to that of the CD56* populations. Using magnetic beads coated with antibodies to either CD4 or CDS, CD3*CD4*, and CD3*CD8* TILs were isolated in a highly enriched form (>92%) and could be expanded for over 40 days in vitro with 1000 units/ml IL2. In a 4-h "Cr release assay the CD4* and CDS* TILs showed minimal luit activity, whereas unseparated cells exhibited significant levels of non-major histocompatibility complex-restricted cytotoxicity. The lytic activity seen in the 4-h assay with unseparated TILs appeared to be related to the presence of CD56* populations. With one exception none of the purified CD4* or CDS* TILs expressed any significant levels of CD56, while the unseparated TILs contained varying numbers of CD3*CD56* and CD3~CD56* populations. Cell-sorting experiments ver ified that the CD56* populations were responsible for most of the lytic activity in 4 h even though CD3*CD56~ cells represented the predominant cell type. Although CD3*CD56~ TILs were minimally lytic in 4 h, we show here that both CD3*CD4* and CD3*CD8* subsets displayed sub stantial cytotoxicity in long-term assays. In the 18-h '( r release assay 5 of 6 CD4* and 2 of 3 CD8+ TILs were lytic for the autologous tumor. In two cases, restimulation with the autologous tumor induced augmented cytolytic activity of TIL subsets and in one case induced lytic activity in 4 h. The cytotoxic activity of TIL subsets was further examined using a 72-h assay in which TILs were cocultured with a confluent layer of tumor cells. The degree of cytotoxicity was quantitated by measuring the amount of crystal violet dye that was incorporated by tumor cells which remained after the incubation period. CD4* and CDS* TILs typically caused greater than a 50% reduction of tumor cells in 3 days and the level of reduction was increased when IL2 was added to the cultures. All the CD4* and CDS* subset preparations were cytotoxic in the 3-day assay even though some were not lytic for certain targets in the IS-h "Cr release assay. The lytic activity mediated by the CD4* and CDS* TILs in the long-term assay was not restricted to the autologous tumor, although in many cases the TIL subsets displayed greater activity toward autologous tumor when compared to allogeneic targets. These results demonstrate that the TIL subsets, CD3*CD4* and CD3*CD8* which were devoid of CD56* expres sion, had substantial antitumor activity and that their mechanism of tumor cell destruction appears to be distinct from that used by CD56* lymphokine-activated killer cell populations. INTRODUCTION TILs' appear to represent part of the host immune response to the tumor and may contain an enriched population of cells Received 6/30/89; revised 10/27/89. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1This study was supported by a National Cancer Institute contract (CM4767303) and a grant from The Cleveland Foundation. 2To whom requests for reprints should be addressed, at Research Institute. Section of Immunology. Cleveland Clinic Foundation. 9500 Euclid Avenue, Cleveland. OH 44195. ' The abbreviations used are: TIL. tumor-infiltrating lymphocyte: TCR. T-cell receptor: MHC, major histocompatibility complex: CTL, cytotoxic T-lympho cytes; NK. natural killer cells; IL2, interlcukin 2; RCC. renal cell carcinoma: LAK, lymphokine-activated killer cells: rlL2, recombinant interlcukin 2: IFN-->. y-interferon; TNF-i«. tumor necrosis factor-«. with reactivity to the autologous tumor (1-6). The majority of infiltrating cells are T-lymphocytes with the number of CDS* cells being more frequent than CD4* cells (1-6). Studies by Vose et al. (7) indicated that TILs from several human solid tumors contain an increased frequency of cytotoxic precursors against autologous tumor when compared to peripheral blood lymphocytes. In addition, recent studies have shown that TILs were therapeutically more potent than IL2-activated spleen cells in several murine tumor models (8, 9). A number of studies have begun to define the immune re sponse to autologous tumor by examining the functional prop erties of TIL subsets and most of the emphasis has been on the cytolytic activity of TILs (10-16). The subsets of effector cells that mediate lysis as well as their specificity have been examined for several human solid tumors, and it is clear that there are differences in the effector population depending on the tumor type (10-16). Analysis of melanoma-derived TILs from bulk cultures and at the clonal level revealed that a portion of the cytolytic cells are specific (4, 5, 10). These cells recognize melanoma antigen via the TCR complex and are MHC-restricted (10). The presence of CTL specific for autologous tumor in TILs from other types of solid tumors has not been readily docu mented (12-16). Effector cells in IL2-expanded TILs from ovarian carcinomas, squamous cell carcinomas of head and neck, and liver tumors displayed nonspecific killing (12, 13, 15, 16). An analysis of the TIL subsets responsible for the lytic activity from ovarian (15, 16) and squamous cell carcinoma of head and neck (12,13) revealed that most cytotoxicity was mediated by NK cells (CD3~CD56*) and MHC-unrestricted CTL (CD3*CD56+) (17-19). Although neither of these popu lations represent a significant portion of the infiltrating lym phocytes, these cells do proliferate to IL2 and can represent a major population of expanded TILs (12,13,15, 16). Incontrasi, CD3*CD56~ cells, which represent the majority of TILs from most of these tumors, did not display significant lytic activity. Work from our laboratory (20) and those of others (10, 21) have shown that TILs expanded in IL2 from renal cell carci nomas were nonspecific in their lytic activity and that both NK cells and T-cells play a role. Here we report on the cytolytic and proliferative potential of isolated CD4* and CDS* TILs from renal cell carcinoma fol lowing in vitro expansion with IL2. CD3*CD4* and CD3*CD8* cells that were purified by positive selection with antibodycoated beads could be expanded in vitro with IL2. We have found that the CD3+CD56~ populations which express either CDS or CD4 display minimal lytic activity in a 4-h 5lCr release assay. Most of the lytic activity observed with cultured TILs was mediated by CD56* cells. However, we show here that both the CD3*CD4* and CD3*CD8* TILs have significant cytotoxic activity that can be detected in long-term assays. Most of the purified CD4* and CDS* TIL subsets were cytotoxic in an 18h 5lCr release assay and all were lytic in a 3-day tumor reduction bioassay. 2363 on April 13, 2017. © 1990 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from CYTOLYTIC ACTIVITY OF TIL SUBSETS MATERIALS AND METHODS Human Renal Cell Carcinoma TIL Isolation and Expansion in Vitro. The patient population with renal cell carcinoma consisted of 12 males and 2 females ranging in age from 31 to 68 years. Renal cell carcinomas were obtained from 13 of 14 patients who had undergone nephrectomy or partial nephrectomy. TILs from I of 14 patients were obtained from an adrenal metastatic lesion. Six of the 14 patients studied here were part of a phase I clinical trial to evaluate TIL therapy in patients with metastatic RCC. The kidneys obtained from complete nephrectomies were perfused with 500 ml of Hanks' balanced salt solution (Whittaker Bioproducts, \Valkersville, MD) to remove extraneous blood and processed as pre viously described (20). Normal tissue, fat, and capsular material were removed, and the tumor tissue was then minced, weighed, and rinsed once with RPMI 1640 (Whittaker Bioproducts). For every 10 g tissue. 20 ml collagenase type III solution (1 mg/ml in RPMI 1640, 109 units/ mg; Cooper Biomédical, Malvern, PA) and 1 ml of DNase type IV solution (2 mg/ml in phosphate-buffered saline; Sigma Chemical Co.. St. Louis, MO) were added to the tissue. Tissue was incubated at room temperature for 4 h. and the resulting cell suspension was washed twice in RPMI 1640 and counted. TILs were placed in Aim V media (Gibco Laboratories, Grand Island. NY) in tissue culture flasks at a total cell concentration of 2.5 x IO5/ ml supplemented with 1000 units/ml human recombinant IL2 (rIL2; Hoffman-La Roche, Nutley, NJ). Cultures were incubated at 37°Cin 5% CO:and 95% humidity. TIL cultures were fed on day 10 with fresh media and rIL2 and then fed as needed to keep the cell concentration at 2.5 x lO'/mlIsolation of CD4* and CD8+ TILs. IOT8-coated beads and IOT4coated beads (Amac, Inc., Westbrook. MA) were used to positively select for CDS* and CD4* cells, respectively. TILs cultured with rIL2 (Hoffman-La Roche) were incubated at a 4:1 beadxell ratio in a 50-ml flask for 30 min at 4°Cwith gentle agitation every 5 min. The immediate negative cells (nonadherent) were separated using a BioMag separator (Advanced Magnetics, Cambridge, MA). The cells with beads attached were washed twice with RPMI 1640 (Whittaker Bioproducts), pipetted vigorously, and put in culture with Aim V and 1000 units/ml rIL2. After 24 h, the CDS* cells were separated from the IOT8 beads using the BioMag separator. The beads were washed twice with RPMI 1640 and placed back in culture with rIL2 for 7-14 days, at which time the separation procedure was repeated to yield more CDS* cells. Using this procedure TILs removed from the IOT8 beads after 24 h were greater than 92% CD3*CD8* cells. TILs removed from the IOT4 beads after 24 h were not pure CD3*CD4* cells; therefore, TILs were cultured with the IOT4 beads and rIL2 for 5-7 days prior to separation. This nu idÃ-licai Kiii allowed the separation of CD4* TILs that were greater than 92% pure. The degree of purity of CD4* and CDS* TILs was examined 3-7 days after the cells were taken off the beads. This recovery period was necessary to allow the reexpression of cell surface CD4 and CDS molecules following exposure to antibody-coated beads.4 Cytolytic Activity. Detection of lytic activity was performed using a 4and 18-h MCr release assay (20). Various concentrations of cultured cells were added to U-bottom 96-well plates to achieve effectontarget cell ratios of 25:1. down to 0.78:1 (6 dilutions). Prior to cytotoxicity assays target cells were labeled with 250 pCi Na"CrO4(New England Nuclear, Boston, MA), washed two times, and then resuspended in media at a concentration of 5 x IO4 viable cells/ml. Thereafter, target cells (5 x lO'/lOO ¿il)were be added to 96-well plates. Following a 4or 18-h incubation period, supernatant fluid from each well was har vested with the Skatron harvesting system and the amount of released "Cr determined in a Beckman gamma 4000 counter. Maximum MCr release was measured by lysing triplicate wells of 5 x 10' cells with \Tc sodium lauryl sulfate. Spontaneous release was measured in 5 x 10' target cells to which medium alone was added. In all the 4and 18-h 5lCr release assays the spontaneous release was never greater than 35%. For each effector:target cell ratio, percentage of specific lysis was calculated using the formula % specific lysis = (Experimental cpm) (spontaneous cpm) (Maximal cpm) (spontaneous cpm) x 100 4J. Finkc. F*. Rayman, J. Alexander. E. Pontes. R. Bukowski. unpublished data. Cytolytic activity was expressed as lytic units/lO* mononuclear cells tested as determined from linear regression analysis of dose-response curves in which the natural log of the number of effectors is plotted against percentage of specific lysis. One lytic unit was defined as the number of effector cells required to produce 15% specific lysis of IO6 target cells in a 4-h (and 18-h) period (10). Isolation of Tumor Targets. Tumor targets used in stimulation ex periments and the cytotoxicity assays were obtained from short-term cultures of autologous and allogeneic RCC. After the 4-h incubation period in collagenase and DNase 1-2 x IO6 tumor cells were placed in Primaria grade tissue culture flasks (Becton Dickinson. Oxnard, CA). The tumor cells were grown in complete RPMI (10% fetal calf serum, 200 niM i.-glutamine, 25 ITIM 4-(2-hydroxyethyl)-l-piperazineethanesulfonic acid, 100 ITIM sodium pyruvate, 10 mivi nonessential amino acids, and 5 x IO"5 M 2-mercaptoethanol) fora minimum of 2-3 weeks. Tumor cells were removed from the flask using trypsin and washed thoroughly. That these short-term cultures represent renal cell carci nomas was demonstrated by morphological examination and by immunostaining with antibodies to RCC, Uro 2, and AE1/3. When sufficient cells were available, growth in nude mice was demonstrated for the short-term-cultured tumors. Previously we showed that IL2 expanded TILs from RCC were cytotoxic for uncultured autologous and allogeneic renal cell tumors (20). Here we show that the same degree of lysis can be detected from TILs when short-term cultures of RCC are used as targets. The shortterm cultures of RCC generally make better tumor targets because they are more homogeneous in contrast to fresh tumor cell suspensions that can contain many nontumorous cells. Also, for the cultured RCC the viability is greater and spontaneous release is lower than what is observed with fresh tumor cells.4 Daudi cells and the renal cell carcinoma line RC2 (22) were main tained in complete RPMI and used as targets for cytotoxicity studies. All cell lines were periodically tested and found to be free of Mycoplasma infection. Three-Color Immunocytometry. Immunocytometric analysis of cul tured TILs was performed as previously described (23). Fluorescein isothiocyanate, phycoerythrin. and biotin-conjugated monoclonal an tibodies were used to phenotypically identify and quantitatc lymphocytic subsets. Antibody titers were adjusted according to the fluores cence sensitivity of the flow cytometer (FACScan/BD) used. Isotypic controls for each particular subclass of immunoglobulin and system used were utilized to allow for the most accurate delineation of positive and autofluorescent populations and to control for nonspecific binding by a particular subclass of immunoglobulin. Analyses on the FACScan were performed utilizing an argon ion laser (Cyonics) with 15 mW of 488 nm excitation. Optics in the fluorescence path included a focusing objective, splitters, dichroic mir rors, a 530-nm bandpass filter, a 585-nm bandpass filter, and a 650nm longpass/cut-on filter. Triggering was set on the forward scatter channel and the threshold adjusted to exclude debris. Live gating of the forward and orthogonal scatter channels, as determined by fluorescence (CD45* CD 14") backgating, was used to this same end and to selectively acquire events for lymphocytes. Results were reported as a percentage of total mononuclear cells in suspension corrected for nonspecific binding by isotypic controls as determined by measurement of autofluorescent background. Cell Sorting. Populations were isolated using a FACStar* flow cytom eter (Becton Dickinson). Cells (10 x IO*1)were immunostained with fluorescein isothiocyanate-anti-CD5 antibody (Becton Dickinson) and phycoerythrin-anti-CD56 antibody (Becton Dickinson) at 4°Cfor 30 min with vortexing every 10 min. followed by two washes with calcium and magnesium-free Hanks' solution. The washed pellets were resus pended in sterile culture media for sorting. Populations were physically isolated according to the phenotype desired with modifications of previously described methods (24). Cells were sorted at a maximum of 800 cells/s to allow for optimal purity and recoveryFurther purity was assured through the use of nonrectangular gates as provided by an Acquisition Sort Processor (Becton Dickinson) used in conjunction with a MicroVAX II/GPX Vaxstation. Cells were sorted directly into 2364 on April 13, 2017. © 1990 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from CYTOLYTIC ACTIVITY OF TIL. SUBSETS culture media to ensure the highest postsorting viability obtainable. Long-Term Cytotoxicity Assay. The antitumor potential of TILs in vitro was assessed using a modification of a 3-day cytotoxicity assay (25). Autologous or allogeneic tumor cells were seeded into 96-well flat bottom plates at a density of 2 x IO4 cells/well. Twenty-four h later when tumor cells were attached and confluent the media were removed and replaced with fresh media containing varying numbers of TILs in triplicate such that the effectontarget cell ratios ranged from 6:1 to 1.5:1. In all the experiments TILs were cultured with viable tumor cells in the presence and absence of IL2 (1000 units/ml). After 3 days of cultures at 37°Cin 5% CO3and 95% humidity nonadherent cells were removed by washing the wells 3 times with phosphate-buffered saline. The plates were then fixed with 0.1 ml 5% formalin for 20 min. Thereafter, attached cells were stained with a 0.05% solution of crystal violet for 20 min and then the plates were washed twice with tap water. The remaining dye which was incorporated into the attached cells was eluted from the wells with 95% methanol and read at 570 nM in an enzyme-linked immunosorbent assay reader (MR 580; Dynatech). The amount of crystal violet eluted from the attached cells was proportional to the number of cells remaining in the wells. The maximal level of dye taken up by a confluent layer of tumor cells was determined by culturing tumor cells alone. The degree of tumor cell destruction mediated by TILs was quantitated as of cytotoxicity = 1 Absorbance of tumor cells treated with TIL Absorbance control x 100.