**5. Structure-activity relationship**

Structure-activity relationship (SAR) of SHetA2 and its derivatives has been valuable to identify important structural modifications that have contributed to its anticancer potency and selectivity, and to guide the design of more potent and less toxic SHetA2 analogs. **Figure 2** illustrates the structural evolution of SHetA2 from ATRA. The goal was to increase the selectivity of retinoids towards RAR/RXR subtypes, so as to reduce the associated toxicities while retaining its anticancer activities. It has been shown that RAR-specific ligands can rescue *Raldh2−/−* embryos as effectively as ATRA, whereas RXR ligand showed no effect [36]. One strategy was to conformationally restrict the double-bonds of RA to allow a better fit into a specific receptor subtype by incorporating aromatic rings. For example, SR3986 (**Figure 2**) was developed with an aromatic ring from ATRA. Subsequently, TTNPB (**5**), the lead arotinoid, was modified with a diaryl group to increase its rigidity, and it was found to be selective for RAR receptors and 10 times more potent as compared to ATRA [38]. Unfortunately, this compound also exhibited a 10,000-fold increase in toxicity, which limited its clinical usage [29].

In order to reduce toxicity, a benzylic carbon in the tetrahydronaphthalene was replaced with a heteroatom (O, N, and S). The purpose was to prevent benzylic (metabolic) oxidation which could result in toxic metabolites. This single modification resulted in Hets (**Figure 1**, **6**–**12**) with similar biological activities to RA but significantly reduced toxicities [29, 31]. An example was the diaryl heteroarotinoid (**Figure 1**, **6**), a RAR-selective retinoid derivative. It differed from TTNPB by an oxygen heteroatom, but exhibited a significant decrease in toxicity, increasing the maximum tolerated dose (MTD) by 3000-fold as compared to TTNPB (**Figure 1**, **5**) [39]. Other monoaryl heteroarotinods (**Figure 1**, **7a**, **7b**) were also evaluated, and revealed a 3-fold decrease in toxicity along with a decreased ability to activate the RAR receptors when compared to ATRA (**Figure 5**) [40, 30].

**Figure 5.** Maximum tolerated dose (MTD) for retinoids and their derivatives.

#### **5.1. Thiochroman rings**

**Figure 3.** Synthesis of aminothiochroman.

74 Anti-cancer Drugs - Nature, Synthesis and Cell

**Figure 4.** Improved synthesis of aminothiochroman and SHetA2.

Structure-activity relationship (SAR) of SHetA2 and its derivatives has been valuable to identify important structural modifications that have contributed to its anticancer potency and selectivity, and to guide the design of more potent and less toxic SHetA2 analogs. **Figure 2** illustrates the structural evolution of SHetA2 from ATRA. The goal was to increase the selectivity of retinoids towards RAR/RXR subtypes, so as to reduce the associated toxicities while retaining its anticancer activities. It has been shown that RAR-specific ligands can rescue *Raldh2−/−* embryos as effectively as ATRA, whereas RXR ligand showed no effect [36]. One strategy was to conformationally restrict the double-bonds of RA to allow a better fit into a specific receptor subtype by incorporating aromatic rings. For example, SR3986 (**Figure 2**) was developed with an aromatic ring from ATRA. Subsequently, TTNPB (**5**), the lead arotinoid, was modified with a diaryl group to increase its rigidity, and it was found to be selective for RAR receptors and 10 times more potent as compared to ATRA [38]. Unfortunately, this compound also exhibited a 10,000-fold increase in toxicity, which limited its clinical usage [29]. In order to reduce toxicity, a benzylic carbon in the tetrahydronaphthalene was replaced with a heteroatom (O, N, and S). The purpose was to prevent benzylic (metabolic) oxidation which could result in toxic metabolites. This single modification resulted in Hets (**Figure 1**, **6**–**12**) with similar biological activities to RA but significantly reduced toxicities [29, 31]. An example was

**5. Structure-activity relationship**

SAR studies have identified structures containing six-membered ring (**Figure 1**, **7a**, **7b**) tend to confer increased RARβ selectivity over five-membered ring systems (**Figure 1**, **12a**, **12b**), while sulfur heteroatom confers a greater RARγ selectivity over oxygen atom [19]. The thiochroman ring system is flexible and has been shown to induce apoptosis to a greater extent than a rigid planar quinoline unit [41]. These findings highlight the important role played by

the thiochroman ring in enhancing the activity of Flex-Hets. Therefore, the thiochroman ring forms one of the fundamental moieties of SHetA2 and its analogs (14–20).

## **5.2. Two-atom linkers**

To further increase the selectivity for each RAR and RXR receptors, various linkers were placed between the two aryl groups of the Hets by modifying their structure rigidity. Two-atom linker compounds such as amide (**Figure 1**, **8**) and ester (**Figure 1**, **9**) were reported [42, 43]. Compound (**8**) was found to be a receptor panagonist, while compound (**9**) was RXR selective. Both showed significant growth inhibitory activities against head and neck cancer using a tumor xenograft mouse model [42]; however, only compound (**8**) induced apoptosis in ovarian cancer cells. This indicated that RXR activation is sufficient to inhibit tumor growth, while activation of both RAR and RXR are required for the maximum activity, at the expense of toxicity. Other two ester-linked compounds (**Figure 1**, **10**, **11**) were found to activate both RARs and RXRs [44]. On the other hand, these ester-linked Hets appeared to only induce growth inhibition but not apoptosis.
