*2.5. Discussion*

Many efforts have been designed to identify and develop new small molecule HIF-1α inhibitors. Natural products have played an important role in drug discovery and development by providing novel chemical entities. In the present study, we found that salternamide A (SA) is a novel HIF-1α inhibitor derived from natural sources. SA is one of the salternamides that we recently isolated from a halophilic *Streptomyces* sp. of saltern-derived actinomycetes [16]. In terms of chemical characteristics, SA is novel and is the first chlorinated member of the manumycin class. Initially, SA exhibited the most potent cytotoxicity among the isolated salternamides against a panel of cancer cell lines [16]. However,

the precise mechanism of action of the anti-proliferative activity of SA is not fully understood. Based on the unique chemical structure and potent growth inhibition of cancer cells, we attempted to elucidate its mechanism of action in cancer cells. From the initial analysis of chemical structures of salternamides with the amide-side chain, we assumed that HIF-1α should be a candidate target molecule. As a result, SA suppressed the accumulation of the HIF-1α protein under hypoxic conditions in human colorectal HCT116 cancer cells.

VHL is the substrate recognition component of an E3 ubiquitin ligase complex that targets HIF-1α and HIF-2α for ubiquitin-mediated degradation under normoxic conditions. The interaction of pVHL (the protein transcript of the VHL gene) with HIF-1α is one of several O2-dependent events thought to regulate HIF activity [24]. Inactivation of the VHL gene is correlated to the development of highly vascularized tumors. pVHL directs the polyubiquitylation of HIF-1α for ubiquitin-mediated degradation in an oxygen-dependent manner. Therefore, loss of VHL is sufficient not only to stabilize HIF-1α subunits under normoxia, but also to fully activate HIF-mediated responses [24]. Maranchie *et al*. [25], however, reported that loss of the tumor suppressor VHL did not enhance the growth of subcutaneous tumors derived from primary cells. In the present study, the level of VHL was decreased by SA at 10 μM, suggesting that the inhibition of HIF-1α accumulation by SA under hypoxic conditions might not be correlated to the increasing of the degradation of HIF-1α. However, the detailed mechanisms of this action need to be further clarified. Further study revealed that the suppressive effect of SA on the accumulation of HIF-1α might be partly due to the inhibition of the translation of HIF-1α, without affecting the expression level of HIF-1α mRNA or the degradation of the HIF-1α protein.

The expression of HIF-1α is closely regulated by both its protein synthesis and degradation. The translation of HIF-1α is promoted by the axis of mTOR signaling through the activation of p70S6K, 4E-BP1, and eIF4E [26]. The PI3K/Akt/mTOR and p42/p44 MAPK pathways also activate translational regulatory proteins, including 4E-BP1, eIF4E, and p60S6K [4]. SA modulates the mTOR signaling pathways and the axis of PI3K/Akt/MAPK-mediated downstream signaling, thus leading to the suppression HIF-1α translation under hypoxic conditions in cancer cells.

Recent studies have also reported that STAT3 is activated in response to hypoxia in various cancers [13,14], and STAT3 signaling is required for HIF-1α protein accumulation in the activation of endogenous HIF-1α target genes [27,28]. In the present study, SA suppressed the activation of STAT3, which led to the downregulation of HIF-1α protein expression under hypoxic conditions. These findings also suggest that the regulation of STAT3 by SA is partly involved in the suppression of HIF-1α accumulation under hypoxic conditions in cancer cells.

Taken together, although further detailed study is needed to elucidate how the translation of HIF-1α is regulated, SA is considered a novel inhibitor of HIF-1α protein synthesis under hypoxic conditions in cancer cells.

In addition, the anti-proliferative activity of SA was associated with G2/M cell cycle arrest and subsequent apoptotic cell death in HCT116 cancer cells. G2/M phase arrest was highly correlated with the downregulation of CDC25C and CDC2 checkpoint protein expression [27]. The progression from G2 to M phase is regulated by a number of cyclin family members, such as cyclin B1 and CDC2. Cyclin B1, together with cyclin A, promotes the G2/M transition [29,30]. Meanwhile, CDC2 is crucial for the G1/S and G2/M phase transitions of the eukaryotic cell cycle. The phosphatase activity of CDC25C is also implicated in the regulation of the progression of G2/M phase [31]. Chk1 (Checkpoint Kinase 1) and Chk2 (Checkpoint Kinase 2) are multi-functional protein kinases that coordinate the response to specific types of DNA damage [32]. As a result, we found that SA-mediated G2/M phase cell cycle arrest was associated with the regulation of CDC25C and CDC2 and its related checkpoint kinases Chk1/2.

Apoptosis is one of the main types of programmed cell death, and many anti-cancer agents induce apoptosis against cancer cells to achieve therapeutic efficacy [33,34]. We also found that longer and higher concentrations of SA exposure are able to induce apoptotic cell death in the HCT116 cells. Apoptotic cell death by SA was confirmed by significant increases in the proteolytic cleavage of caspase-8, -3 and PARP, and the downregulation of the antiapoptotic Bcl-2 and Bcl-xL proteins.

Autophagy plays a complex role in cancer development, with a tumor-progressive or a tumor-suppressive effect [35]. Apoptosis and autophagy may be interconnected, either antagonistically or cooperatively, in response to various anti-cancer therapeutics in different cancer cells [36]. Herein, we observed that the formation of LC3-II, a marker of autophagy, was also induced by SA, suggesting that autophagy-mediated cell death is also involved in the anti-proliferative activity of SA in HCT116 cancer cells.
