IL-2, IL-3, and IFN-gamma differently affect in vivo frequencies of circulating precursors of cytotoxic T lymphocytes (CTL-P).

Experimental animal and human in vivo studies have previously demonstrated the impact of exogenous administration of various cytokines on frequencies of circulating myeloid and LAK precursor cells. For the first time we investigated whether exogenous cytokines, in the absence of antigenic challenge, may also influence frequencies of circulating antigen-specific cytotoxic T-lymphocyte precursor cells. We further asked whether triggering of autoimmune pathways as has been reported for several cytokines can be confirmed on the cellular level by demonstration of induction of autoreactive CTL-p. Limiting dilution analysis was used to determine alloreactive CTL-p frequencies in 31 patients with nonhematologic diseases before and after short-term systemic treatment with either rIL-2 (4.8 x 10(6) IU/m2 bid), rIL-3 (2.5, 5.0 or 10.0 micrograms/kg qd), rGM-CSF (5 micrograms/kg qd), rIFN-gamma (200 or 400 micrograms qd), or IFN-alpha (3 or 5 x 10(6) IU qod). Simultaneously, autoreactive CTL-p frequencies were determined by split-well analysis in 25 of these patients. We found that rIL-2 significantly expands the circulating precursor pool of alloreactive CTL (p < 0.05). rIL-3 affected CTL-p frequencies in a dose-dependent fashion. Low and intermediate doses of rIL-3 did not exhibit significant effects, whereas 10 micrograms/kg rIL-3 led to expansion of alloreactive CTL-p in the same order of magnitude as did rIL-2. This effect was statistically significant when compared with rGM-CSF (p < 0.02), which apparently had no influence on alloreactive CTL-p frequencies. In contrast to rIL-2 and rIL-3, exogenous rIFN-gamma markedly reduced the circulating precursor pool of CTL. This again was statistically significant compared with rIFN-alpha (p < 0.03), which, like rGM-CSF, did not exhibit any effects on the level of alloreactive CTL-p. Frequencies of autoreactive CTL-p were invariably below the limit of detection in our system (< 1/300,000). In conclusion, these data demonstrate that (a) short-term systemic administration of rIL-2, rIL-3, and rIFN-gamma differently affects the clone size of circulating precursors of alloreactive CTL in man, while rGM-CSF and rIFN-alpha do not exhibit measurable effects, and (b) none of the cytokines administered is capable of uncovering detectable frequencies of autoreactive CTL-p.