The mechanisms of CTL-mediated tumor regression in vivo remain to be fully understood. If CTL do mediate tumor regression in vivo by direct cytotoxicity, this may occur via two major effector mechanisms involving the secretion of perforin/granzymes and/or engagement of Fas by Fas ligand (FasL) expressed by the activated CTL. Although the perforin pathway has been considered the dominant player, it is unclear whether Fas-mediated cytotoxicity is additionally required for optimal tumor rejection. Previously, we produced H-2Ld-restricted CTL reactive against the CMS4 sarcoma, which expresses a naturally occurring rejection Ag recognized by these CTL and harbors a cytokine (IFN-γ plus TNF)-inducible, Fas-responsive phenotype. The adoptive transfer of these CTL to syngeneic BALB/c mice with minimal (day 3 established) or extensive (day 10 established) experimental pulmonary metastases resulted in strong antitumor responses. Here we investigated whether a FasL-dependent CTL effector mechanism was important for optimal tumor regression in this adoptive immunotherapy model. The approach taken was to compare the therapeutic efficacy of wild-type to FasL-deficient (gld) CTL clones by adoptive transfer. In comparison with wild-type CTL, gld-CTL efficiently mediated tumor cytolysis and produced comparable amounts of IFN-γ, after tumor-specific stimulation, as in vitro assessments of Ag recognition. Moreover, gld-CTL mediated comparably potent antitumor effects in a minimal disease setting, but were significantly less effective under conditions of an extensive tumor burden. Overall, under conditions of extensive lung metastases, these data revealed for the first time an important role for a FasL-dependent CTL effector mechanism in optimal tumor regression.

-EJB 95 l670/1 I S 1 74T human colon carcinoma cells that contain approximately equal amounts of CAMP-dependent protein kinase (PKA) isozymes, PKA-I and PKA-11, were infecled with retroviral vectors coding for regulatory (R) and catalytic (C) subunits of human PKA. In cells overexpressing RIIcr, Rllij, and RTT, P (a RII,, mutant at the autophosphorylation site), PKA-JT levels increased whilc PKA-I levels decreased. PKA-I was almost completely eliminated in cells overexpressing RII,{ or RIT,,-P. In contrast, overexpression of either RI,, or C, had little or no effect on PKA isozyme levels. Although all infectants expressed high levels of PKA subunit inRNAs in accordance with gene introduction, the R subunit protein expression was reflected in PKA isozyme levels rather than in subunit mRNA levels. Only RII, infectants demonstrated marked growth inhibition in monolayer C U~~L I I .~, reduced thymidine incorporiition into DNA, and inability to grow in semisolid medium or in serum-free medium. Conversely, it11 other infectants displayed growth propcrties similar to uninfected parental cells. The growth-retardation properties of RIT, infectants were reflected i n their altered phenotypic appearances. Our findings that the mutant RIl,rP could not mimic the growth-inhibitory effect of RII, suggest thc functional importance of the autophosphurylation site in RII,,. Our results suggest a role for KIT,, in the suppression of neoplastic cell growth, and thus abnormal expression of R subunit isoforms of PKA may bc involved in neoplastic transformation.