The Finally Rewarding Search for A Cytotoxic Isosteviol Derivative

Acid hydrolysis of stevioside resulted in a 63% yield of isosteviol (1), which served as a starting material for the preparation of numerous amides. These compounds were tested for cytotoxic activity, employing a panel of human tumor cell lines, and almost all amides were found to be non-cytotoxic. Only the combination of isosteviol, a (homo)-piperazinyl spacer and rhodamine B or rhodamine 101 unit proved to be particularly suitable. These spacered rhodamine conjugates exhibited cytotoxic activity in the sub-micromolar concentration range. In this regard, the homopiperazinyl-spacered derivatives were found to be better than those compounds with piperazinyl spacers, and rhodamine 101 conjugates were more cytotoxic than rhodamine B hybrids.


Results
Stevioside (obtained from different local vendors) was hydrolyzed in methanolic hydrochloric acid, and isosteviol (1) was obtained in a 63% isolated yield. Due to some ambiguity in the assignment of its 1 H and 13 C NMR spectra [75][76][77], a complete analysis was undertaken. However, it quickly became apparent that the usual 1D and 2D NMR experiments (e.g., gHSQC and gHMBC) did not allow a complete and accurate assignment of all signals, since even in these spectra, some signals overlapped very strongly. In addition, data from the literature contradicted each other with respect to the assignment of the quaternary carbons. To solve this problem, further NMR experiments were carried out, with which it was possible to directly separate the connectivity in the carbon framework. Two methods are suitable for this purpose: On the one hand, this is the classical INADEQUATE method [78][79][80][81], which requires larger substance amounts but also long measurement times due to the low sensitivity, or, on the other hand, the 1,1-ADEQUATE experiment [82][83][84]. In contrast to the INADEQUATE experiment, which is based on a correlation of the 1 J coupling between two vicinal 13 C nuclei, the 1,1-ADEQUATE experiment uses an INEPT transfer between 1 H and 13 C adjacent nuclei and the subsequent formation of double quantum coherence, thereby circumventing the problem of the low natural abundance of 13 C nuclei.
So far, no INADEQUATE or 1,1-ADEQUATE spectra have been reported in the literature specifically for isosteviol. Due to the very good solubility of isosteviol in chloroform (520 mg in 0.7 mL) and yet sufficiently low viscosity at the measuring temperature used (27 • C), it was possible to perform both experiments and compare the results. While the INADEQUATE experiment required a measuring time of 80 h, 20 h was sufficient for the ADEQUATE experiment. The results are shown in Figures 1 and 2 (in the ADEQUATE experiment, the chemical shift is shown in ppm instead of the double quantum coherence frequency, for better comparability). It was shown that both methods are very well suited for an exact and doubtless assignment. However, it is also evident that the INADEQUATE spectrum is somewhat easier to interpret, since no overlapping proton signals have to experiment, the chemical shift is shown in ppm instead of the double quantum coherence frequency, for better comparability). It was shown that both methods are very well suited for an exact and doubtless assignment. However, it is also evident that the INADEQUATE spectrum is somewhat easier to interpret, since no overlapping proton signals have to be taken into account. Both NMR experiments, however, allow a doubtless and reliable assignment of all signals.   experiment, the chemical shift is shown in ppm instead of the double quantum coherence frequency, for better comparability). It was shown that both methods are very well suited for an exact and doubtless assignment. However, it is also evident that the INADEQUATE spectrum is somewhat easier to interpret, since no overlapping proton signals have to be taken into account. Both NMR experiments, however, allow a doubtless and reliable assignment of all signals.   Compound 1 was transformed in situ with oxalyl chloride into the corresponding acid chloride followed by the addition of the corresponding amine (Scheme 2). The pyridine amides 2-4, the quinoline amides 8 and 9, the amides 12-23, the (homo)-piperazinyl amides 24 and 25 (Scheme 3), the (homo)-morpholinyl amides 26 and 27, and the (homo)thiomorpholinyl amides 28 and 29 were obtained; compound 30 was synthesized from 1 Molecules 2023, 28, 4951 4 of 24 and ethylene diamine in 88% yield. Rhodamine B or rhodamine 101 was activated with oxalyl chloride in the same manner, followed by the reaction with either 26 and 27 to yield 31 and 32 (from rhodamine B) or 33 and 34 (from rhodamine 101), respectively. We refrained from using 30 as a starting material to prepare the corresponding rhodamine conjugates, since it has since become apparent that these conjugates exist preferentially in a non-cationic but neutral spirocyclic form. These electrically neutral molecules proved to be hardly cytotoxic, since they obviously cannot interact with membranes.
Compound 1 was transformed in situ with oxalyl chloride into the corresponding acid chloride followed by the addition of the corresponding amine (Scheme 2). The pyridine amides 2-4, the quinoline amides 8 and 9, the amides 12-23, the (homo)-piperazinyl amides 24 and 25 (Scheme 3), the (homo)-morpholinyl amides 26 and 27, and the (homo)thiomorpholinyl amides 28 and 29 were obtained; compound 30 was synthesized from 1 and ethylene diamine in 88% yield. Rhodamine B or rhodamine 101 was activated with oxalyl chloride in the same manner, followed by the reaction with either 26 and 27 to yield 31 and 32 (from rhodamine B) or 33 and 34 (from rhodamine 101), respectively. We refrained from using 30 as a starting material to prepare the corresponding rhodamine conjugates, since it has since become apparent that these conjugates exist preferentially in a non-cationic but neutral spirocyclic form. These electrically neutral molecules proved to be hardly cytotoxic, since they obviously cannot interact with membranes.
For comparison, quaternization was performed with iodomethane, and this provided 5-7 as well as 10 and 11, respectively. Isosteviol (1) and all derivatives were subjected to sulforhodamine B (SRB) assays employing a panel of human tumor cell lines and non-malignant murine fibroblasts NIH 3T3 and HEK293 cells for comparison. The results from these assays are compiled in Table  1.
The SRB assays showed neither the parent compound isosteviol (1) nor almost none of the amides to hold any cytotoxic effect on the human tumor cell lines; they were also non-cytotoxic for the non-malignant cell lines NIH 3T3 and HEK293. The (homo)-piperazinyl amides 24 and 25, however, showed slight cytotoxic effect, with the homopiperazinyl amide 25 performing slightly better than the piperazinyl-spacered compound 24. A significant improvement was made with the (homo)-piperazinyl rhodamine-B-spacered compounds 31 and 32, and the rhodamine 101 hybrids 33 and 34 were even more cytotoxic than the rhodamine B analogs 31 and 32. The selectivity to distinguish between malignant and non-malignant cell lines, however, was low. For comparison, quaternization was performed with iodomethane, and this provided 5-7 as well as 10 and 11, respectively.
Isosteviol (1) and all derivatives were subjected to sulforhodamine B (SRB) assays employing a panel of human tumor cell lines and non-malignant murine fibroblasts NIH 3T3 and HEK293 cells for comparison. The results from these assays are compiled in Table 1.
The SRB assays showed neither the parent compound isosteviol (1) nor almost none of the amides to hold any cytotoxic effect on the human tumor cell lines; they were also non-cytotoxic for the non-malignant cell lines NIH 3T3 and HEK293. The (homo)piperazinyl amides 24 and 25, however, showed slight cytotoxic effect, with the homopiperazinyl amide 25 performing slightly better than the piperazinyl-spacered compound 24. A significant improvement was made with the (homo)-piperazinyl rhodamine-B-spacered compounds 31 and 32, and the rhodamine 101 hybrids 33 and 34 were even more cytotoxic than the rhodamine B analogs 31 and 32. The selectivity to distinguish between malignant and non-malignant cell lines, however, was low.
These results emphasize once again that for high cytotoxic activity of terpene/ rhodamine hybrids, the interplay between the selected terpene, spacer and lipophilic cation is crucial. If the cationic part is not lipophilic enough (as in compounds 6-7, 10 and 11), no cytotoxic activity is obtained. The values obtained for isosteviol derivatives are basically much smaller than those previously obtained for pentacyclic triterpenes. On the other hand, a pronounced cytotoxicity can be achieved also for non-cytotoxic isosteviol if an Molecules 2023, 28, 4951 6 of 24 appropriate rhodamine residue is added to the terpenoid backbone via a suitable spacer. Once again, the homopiperazinyl spacer proves to be superior to the piperazinyl spacer.

Experiment
NMR spectra were recorded using the Varian spectrometers (Darmstadt, Germany) DD2 and VNMRS (400 and 500 MHz, respectively). MS spectra were taken on a Advion expressionL CMS mass spectrometer (Ithaca, USA; positive ion polarity mode, solvent: methanol, solvent flow: 0.2 mL/min, spray voltage: 5.17 kV, source voltage: 77 V, APCI corona discharge: 4.2 µA, capillary temperature: 250 • C, capillary voltage: 180 V, sheath gas: N2). Thin-layer chromatography was performed on pre-coated silica gel plates supplied by Macherey-Nagel (Düren, Germany). IR spectra were recorded on a Spectrum 1000 FT-IRspectrometer from Perkin Elmer (Rodgau, Germany). The UV/Vis-spectra were recorded on a Lambda 14 spectrometer from Perkin Elmer (Rodgau, Germany); optical rotations were measured using a JASCO-P2000 instrument (JASCO Germany GmbH, Pfungstadt, Germany). The melting points were determined using the Leica hot stage microscope Galen III (Leica Biosystems, Nussloch, Germany) and are uncorrected. The solvents were dried according to usual procedures. Microanalyses were performed with an Elementar Vario EL (CHNS) instrument (Elementar Analysensysteme GmbH, Elementar-Straße 1, D-63505 Langenselbold, Germany). All dry solvents were distilled over respective drying agents, except for DMF, which was distilled and stored under argon and molecular sieve. Reactions using air-or moisture-sensitive reagents were carried out under argon atmosphere in dried glassware. Triethylamine was stored over potassium hydroxide. Biological assays were performed as previously reported, employing cell lines obtained from the Department of Oncology [Martin-Luther-University Halle Wittenberg; they were bought from ATCC: malignant: A 375, HT29, MCF7 and A2780; non-malignant: NIH 3T3]. Rhodamine B and stevioside were obtained from local vendors and used as received.
For the SRB assay: cells were seeded into 96-well plates on day zero at appropriate cell densities to prevent confluence of the cells during the period of the experiment. After 24 h, the cells were treated with different concentrations (1, 3, 7, 12, 20 and 30 µM), but the final concentration of DMSO/DMF never exceeded 0.5%, which was non-toxic to the cells. After 72 h of treatment, the supernatant media from the 96-well plates were discarded, and then the cells were fixed with 10% trichloroacetic acid and allowed to rest at 4 • C. After 24 h of fixation, the cells were washed in a strip washer and then dyed with SRB solution (200 µL, 10 mM) for 20 min. Then the plates were washed four times with 1% acetic acid to remove the excess dye and allowed to air-dry overnight. Tris base solution (200 µL, 10 mM) was added to each well. The absorbance was measured with a 96-well plate reader from Tecan Spectra.

General Procedure for the Synthesis of Amides (GPA)
A solution of 1 (1 equiv.) in dry DCM (10 mL) was treated with oxalyl chloride (4 equiv.) and DMF (catal.) for 1 h. The volatiles were evaporated under reduced pressure. To a solution of the residue in dry DCM (10 mL) the corresponding amine (3 equiv.) was added, and the mixture was stirred at room temperature for 1 h. The usual aqueous workup, followed by chromatography, gave amides.

General Procedure for the Quaternization (GPB)
To a solution of 2-4, 8, or 9 in dry DCM (3 mL), iodomethane (3 mL, 0.05 mmol) was added, and the mixture was stirred at room temperature for 2 h. The volatiles were evaporated under reduced pressure, and the residue was subjected to chromatography to afford 5-6, 10, and 11.

General Procedure for the Synthesis of Rhodamine Conjugates (GPC)
The respective rhodamine was dissolved in dry dichloromethane (10 mL) and mixed with oxalyl chloride (4 eq.) and catalytic amounts of DMF. Following the conditions of GPA as described above, the residue was dissolved in dry DCM (10 mL) and compounds 22 or 23 (3 eq.) were added. Stirring at room temperature was continued for 1 h. The usual aq. work-up followed by chromatography furnished the conjugates 31-34.