The cellular uptake efficiency of
1–4 in NCI-H460 cells after 6-h incubation has been evaluated by determining the intracellular osmium contents using inductively coupled plasma mass spectrometry (ICP-MS). All compounds exhibited concentration-dependent cellular accumulation (Fig. S9a in Supporting information).
1 and
2 showed in a higher cellular uptake, which is consistent with the above antiproliferative evaluation. We have further verified the cytotoxicity of
1 by colony formation assay (Fig. S9b in Supporting information). A 58% colony was formed in the 3.5 µmol/L
1-treated NCI-H460 cells, while no colony formation was observed when the concentration of
1 was increased to 7.0 and 14.0 µmol/L. Among these osmium drugs, we have selected
1 as the representative compound for further biological studies because of its higher stability and cytotoxicity. The GSH level is an indication and of great importance in controlling the redox system in mammalian cells. It acts as a scavenger to reduce ROS such as the superoxide (O
2–) and hydroxyl radical (HO
•) in cells [
27,
28]. As mentioned above,
1 was readily reduced by GSH to give
5. This
1 →
5 transformation should have a great impact on the cellular GSH level and in turn the ROS level. First of all, we determined the cellular GSH level using monochlorobimane as the thiol probe. This thiol probe shows a weak fluorescence itself but it forms a strong fluorescent adduct when bound to GSH [
29]. In the
1-treated cells, a significant cellular GSH suppression was observed. Compared to control group, the GSH levels dropped by 7% and 25% in 14 and 28 µmol/L
1-treated cells, respectively, and buthionine sulphoximine (BSO) was used as the positive control, (
Fig. 4a). Next, we examined the cellular ROS expression using the fluorescent redox probe, 2′, 7′-dichlorodihydrofluorescein diacetate (DCFH-DA) [
30]. DCFH-DA itself is nonfluorescent and it readily converts to the highly fluorescent DCF in the presence of ROS [
31]. Based on the fluorescence intensities determined by flow cytometry (
Fig. 4b), the ROS accumulation enhancement was estimated to be 1.5 and 4.4 fold for the 14 and 28 µmol/L
1-treated cells, respectively; which is further supported by the intensive green fluorescence images (
Fig. 4c). In order to evaluate the effect of ROS, we determined the cell viability using the 24-h
1-incubated NCI-H460 cells that have been pretreated with 5 mmol/L
N-acetylcysteine (NAC) for 2 h (
Fig. 5a). NAC is a well-known ROS scavenger which can capture ROS in cells. As expected, the drop in cell viability is inversely proportional to the concentration of
1-treated cells. On the other hand, the cell viability with the NAC-pretreated cells is higher than the untreated cells in various concentrations of
1-treated cells. These results indicate that
1 induced ROS-mediated cell death. The endoplasmic reticulum has recently emerged as a promising target for anticancer agents [
32]. The enhancement of cellular ROS would induce ER stress and trigger programmed cell death
via apoptosis and autophagy [
27,
33]. ER stress accumulates excessive unfolded or misfolded proteins in the ER lumen that will sequester the ER chaperone binding immunoglobin protein (GRP78) away from the ER stress sensor and activates transcription factor 6 (ATF6) [
32–
34]. As shown in
Fig. 5b, the upregulation of GRP78 and the downregulation of ATF6 (full length) after a 24-h
1-incubation indicate that
1 can impair ER function, initiate the unfolded protein response through ROS accumulation, and activate ATF6.