To study the effects of localized secretion of cytokines on tumor progression, the gene for human interleukin 2 (IL-2) was introduced via retroviral vectors into CMS-5 cells, a weakly immunogenic mouse fibrosarcoma cell line of BALB/c origin. Secretion of low levels of IL-2 from the tumor cells abrogated their tumorigenicity and induced a long-lasting protective immune response against a challenge with a tumorigenic dose of parental CMS-5 cells. Co-injection of IL-2-producing CMS-5 cells with unmodified tumor cells inhibited tumor formation even when highly tumorigenic doses of CMS-5 cells were used. Cytolytic activity in mice injected with parental CMS-5 cells was transient and was greatly diminished 3 wk after injection, as commonly observed in tumor-bearing animals. However, in mice injected with IL-2-producing cells, tumor-specific cytolytic activity persisted at high levels for the duration of the observation period (at least 75 d). High levels of tumor-specific cytolytic activity could also be detected in parental CMS-5 tumor-bearing animals 18 d after inoculation with tumor cells, if IL-2-producing CMS-5 cells but not unmodified parental tumor cells were used as targets. These studies highlight the potential advantages of localized secretion of cytokines mediated via gene transfer to induce potent anti-tumor immune responses.

Sllll~mal~Target cell expression of major histocompatibility complex (MHC) class I molecules correlates with resistance to lysis by natural killer (NK) cells. Prior functional studies of the murine NK cell surface molecule, Ly-49, suggested its role in downregulating NK cell cytotoxicity by specifically interacting with target cell H-2D d molecules. . Although the molecular basis for NK cell specificity is poorly understood, target cell susceptibility to spontaneous cytotoxicity is inversely proportional to their level of expression of certain MHC class I molecules (3-5). Two explanations (6) have been proposed. MHC class I molecules may "mask" or interfere with recognition of putative target cell ligands by activation receptors on NK cells. Alternatively, target cell MHC class I molecules may directly engage NK cell receptors leading to inhibitory events in NK ceils. Our previous functional studies (7) of the murine Ly-49 molecule have supported the latter hypothesis.Ly-49 is a type II integral membrane protein (8, 9), with an external C-type lectin domain (9), expressed on a distinct subset of murine NK cells (10) as a disulfide-linked homodimer of44-kD subunits (9). Functional studies with IL-2-activated NK cells demonstrated that Ly-49 + NK cells generally manifested a lytic capacity equivalent to Ly-49-NK cells (7). However, Ly-49 + NK cells were unable to lyse targets of H-2 k or H-2 d haplotype despite efficient lysis of these targets by Ly-49-NK cells, demonstrating that Ly-49 expression significantly influences NK cell target specificity. Transfection of a susceptible target cell line with cDNA encoding H-2D a rendered it resistant to lysis by Ly-49 + NK cells. Ly-49 + NK cells were unable to lyse the D a transfectant by other stimuli, including Ab-dependent cellular cytotoxicity (7, 11). Gene transferred resistance was abrogated by mAbs against either Ly-49 or the otl/o~2 domain of D a. These data were compatible with our hypothesis that Ly-49 is an inhibitory receptor that specifically recognizes D a molecules on targets.Our interpretation of t.hese functional studies predicts that Ly-49 directly interacts with H-2D d. Alternatively, the results may be explained by a NK cell surface molecule that is coexpressed with Ly-49 and that is indirectly influenced by the anti-Ly-49 mAb. Moreover, reversal of functional inhibition by mAbs could be due to other effects such as direct anti-Ly-49 triggering or anti-H-2D d dissociation of another effector cell ligand (12). A final possibility involves the "masking model" in which H-2D a masks another target molecule specifically recognized by . To distinguish between these possibilities, we sought to unequivocally establish the ligand for Ly-49. We have now overexpressed Ly-49 on Chinese hamster ovary (CHO) cells by transfection and DNA amplification. CHO cells bind target cells only when the cells express high levels of Ly-49 and H-2D a, respectively, by transfection. This interaction is specifically blocked by appropriate mAbs. Thus, this study demonstrates that...

No abstract