4. Discussion
A number of studies have led to the identification of different classes of pharmacological agents able to reactivate HbF synthesis both in vitro and in vivo. For many of these chemicals, however, there are concerns over their toxic and carcinogenic potential in humans. Studies have shown that resveratrol not only exhibits antioxidant activity, but can also stimulate the expression of the γ-globin genes [34]. In the current study, we have shown that resveratrol can increase hemoglobin production in K562 cells. However, increasing concentrations of resveratrol in K562 cells resulted in a dose-dependent reduction in cell survival confirming its anti-proliferative effect, possibly mediated through cell cycle arrest and induction of ER stress as shown by Liu et al. [41]. Due to this anti-proliferative effect, identification of resveratrol derivatives which exhibit potent hemoglobin-inducing properties with lower cytotoxicity is important. In an effort to identify such compounds, nine novel resveratrol derivatives were synthesized and tested in K562 cells. Out of these, only three derivatives (P1, P4 and P11) exhibited hemoglobin inducing activity comparable to the parent compound in K562 cells, with only one (P11) being less cytotoxic. Although previous studies [34] have shown that resveratrol also has an effective HbF inducing activity in primary human erythroid cultures, our results show that none of the agents tested, including resveratrol, have a uniformly potent HbF inducing activity in all the primary human erythroid cultures tested. For each agent tested, some cultures responded quite strongly, in terms of increase in the levels of HbF, while others responded poorly. Therefore, cultures were grouped into responders and non-responders based on their HbF increase following induction with each agent. Resveratrol and derivative P11 showed a statistically significant increase in HbF levels, within the responder group, at 1 and 0.01 μM concentrations, respectively. In contrast, both agents reduced HbF levels below the levels observed in the un-treated control, within the non-responder group. The above findings suggest that the effectiveness of the two agents is limited due to their lack of uniform HbF induction in primary erythroid cultures.