Carmustine (BCNU) lung toxicity: A two-step process involving free radical formation and inflammation

1133

Carmustine (BCNU) is used for the treatment of brain tumors and is included in the preparative regimen of stem cell transplantation (SCT) performed in lymphomas. The incidence of BCNU pulmonary toxicity is significant (60%) and it manifests 6-8 weeks after SCT, following bone marrow recovery. We investigated the effect of BCNU and actinomycin-D (ACT), which is not associated with clinical lung toxicity, on pulmonary epithelial cells using A549 cell line. Daudi cells (B-cell lymphoma) were also studied as part of an in vitro model comparable to clinical BCNU lung toxicity in SCT. Cytotoxicity was measured by flow cytometric detection of annexin V-positive cells, DiOC6(3) staining and trypan blue exclusion in repeated experiments. The generation of free radicals was measured by flow cytometric DCF staining, and glutathione levels by a spectrophotometric assay. Activation and death-related surface markers on A549 cells were measured by flow cytometry. Allogeneic lymphokine activated killer (LAK) cell-mediated cytotoxicity against treated and untreated A549 cells was measured by flow cytometric cytotoxicity assay. The effect of butylated hydroxyanisole (BHA), a free radical scavenger, and dexamethasone (DXM) on BCNU and ACT treatment was also studied. Both drugs were cytotoxic to A 549 cells in vitro. BCNU treatment was associated with 2.5 fold increase in free radical generation and markedly increased CD40, CD54, CD95 and TLR-4 expression rendering them better targets for immune response. BCNU cytotoxicity was suppressed by BHA (17%), but more significantly by DXM (65%) correlating with decreased DCF fluorescence and CD40, CD54, CD95 and TLR-4 expression. Also, LAK cell-mediated BCNU-treated A 549 cell kill was significantly lower when cells were co-incubated with DXM (50%), but was mildly affected by BHA (28%). DXM co-incubation enhanced BCNU cytotoxicity, but BHA decreased it in Daudi cells. ACT cytotoxicity compared to BCNU was associated with higher free radical generation and less prominent CD40, CD54, CD95 and TLR-4 expression. BHA inhibition on ACT cytotoxicity was more pronounced than DXM. ACT and DXM co-incubation resulted in increased LAK cell activity against A 549 cells. Glutathione activity was significantly suppressed by ACT treatment, but not with BCNU. There was a prominent increase in the oxidized to total glutathione ratio with ACT treatment. Therefore, in vitro ACT injury appeared to be mediated predominantly by free radical generation. These results indicate that BCNU pulmonary toxicity is induced by a two-step process: epithelial cell damage involving free radical generation followed by immune reaction to surviving cells resulting in an inflammatory process. Addition of dexamethasone to BCNU-including preparative SCT regimens may decrease the incidence and severity of pulmonary toxicity by affecting both steps, without limiting the anti-lymphoma efficacy.