Endoplasmic reticulum stress as a correlate of cytotoxicity in human tumor cells exposed to diindolylmethane in vitro.

The dietary phytochemical indole-3-carbinol (I3C) protects against cervical cancer in animal model studies and in human clinical trials. I3C and its physiologic condensation product diindolylmethane (DIM) also induce apoptosis of tumor cells in vitro and in vivo, suggesting that these phytochemicals might be useful as therapeutic agents as well as for cancer prevention. Deoxyribonucleic acid microarray studies on transformed keratinocytes and tumor cell lines exposed to pharmacologic concentrations of DIM in vitro are consistent with a cellular response to nutritional deprivation or disruptions in protein homeostasis such as endoplasmic reticulum (ER) stress. In this report we investigate whether specific stress response pathways are activated in tumor cells exposed to DIM and whether the ER stress response might contribute to DIM's cytotoxicity. Induction of the stress response genes GADD153, GADD34 and GADD45A, XBP-1, GRP78, GRP94, and asparagine synthase was documented by Western blot and real-time reverse transcriptase-polymerase chain reaction in C33A cervical cancer cells, and induction of a subset of these was also observed in cancer cell lines from breast (MCF-7) and prostate (DU145). The results are consistent with activation of more than 1 stress response pathway in C33A cells exposed to 75 microM DIM. Phosphorylation elF2alpha was rapidly and transiently increased, followed by elevated levels of ATF4 protein. Activation of IRE1alpha was indicated by a rapid increase in the stress-specific spliced form of XBP-1 messenger ribonucleic acid and a rapid and persistent phosphorylation of JNK1 and JNK2. Transcriptional activation dependent on an ATF6-XBP-1 binding site was detected by transient expression in MCF-7, C33A, and a transformed epithelial cell line (HaCaT); induction of the GADD153 (CHOP) promoter was also confirmed by transient expression. Cleavage of caspase 12 was observed in both DIM-treated and untreated C33A cells but did not correlate with cytotoxicity, whereas caspase 7 was cleaved at later times, coinciding with the onset of apoptosis. The results support the hypothesis that cytotoxic concentrations of DIM can activate cellular stress response pathways in vitro, including the ER stress response. Conversely, DIM was especially cytotoxic to stressed cells. Thapsigargin and tunicamycin, agents that induce ER stress, sensitized cells to the cytotoxic effects of DIM to differing degrees; nutrient limitation had a similar, but even more pronounced, effect. Because DIM toxicity in vitro is enhanced in cells undergoing nutritional deprivation and ER stress, it is possible that stressed cells in vivo, such as those within developing solid tumors, also have increased sensitivity to killing by DIM.