Rhodamine 101 Conjugates of Triterpenoic Amides Are of Comparable Cytotoxicity as Their Rhodamine B Analogs

Pentacyclic triterpenoic acids (betulinic, oleanolic, ursolic, and platanic acid) were selected and subjected to acetylation followed by the formation of amides derived from either piperazine or homopiperazine. These amides were coupled with either rhodamine B or rhodamine 101. All of these compounds were screened for their cytotoxic activity in SRB assays. As a result, the cytotoxicity of the parent acids was low but increased slightly upon their acetylation while a significant increase in cytotoxicity was observed for piperazinyl and homopiperazinyl amides. A tremendous improvement in cytotoxicity was observed; however, for the rhodamine B and rhodamine 101 conjugates, and compound 27, an ursolic acid derived homopiperazinyl amide holding a rhodamine 101 residue showed an EC50 = 0.05 µM for A2780 ovarian cancer cells while being less cytotoxic for non-malignant fibroblasts. To date, the rhodamine 101 derivatives presented here are the first examples of triterpene derivatives holding a rhodamine residue different from rhodamine B.


Introduction
Despite significant progress, cancer therapy still falls short of the expectations placed in it many years ago [1,2]. The prognosis for a complete cure is very good for some types of cancer, but still poor for many others, especially when regular screening is taken into account. The high cost of therapy [3][4][5] is often offset especially for cancers that are difficult to treat by only a slight increase in life expectancy and, at the same time, a significantly reduced quality of life. Thus, the survival rate [6] for testicular cancer is approximately 98% while for pancreatic cancer it is about 1%. The main reason for the reduced quality of life and, thus, a reduced compliance by the affected patients is more or less often insufficient selectivity of the chemotherapeutic agents used. As a consequence, there has been no lack of attempts to improve the efficacy but also to reduce the side effects caused by antitumor drugs (such as weight loss, hair loss, etc.). Furthermore, many different strategies have been tested for a successful drug targeting of tumors-whereby the real problem is usually not the solid primary tumor but the metastases that have already formed and spread throughout the body. These attempts [7] included the use of micelles [8], antibodies [9], liposomes [10], polymers but also of drug-loaded nanoparticles [7].
This cytotoxicity (but also to some extent their pronounced tumor/non-tumor cell selectivity) seems to depend on many parameters. On the one hand, this concerns a dependence on the type of terpene ( Figure 1) used: compounds derived from dehydroabietylamine [23] were-by and large-less cytotoxic than those with a pentacyclic triterpenoid backbone [17]. Amides at position C-28 were mostly more active than the analogous esters [22], whereas a direct attachment of a rhodamine B moiety to the triterpenoid backbone resulted in compounds of significantly lowered selectivity [17]. Therefore, the use of a suitable spacer is of crucial importance. Furthermore, triterpene/rhodamine B hybrids holding an ethylenediamine spacer [20] were significantly less active than those with a piperazine spacer; in some cases, the use of a homopiperazine spacer [17] proved successful. However, the presence of a distal cationic center alone is not sufficient for achieving good cytotoxic activity [17,[24][25][26]. Only special delocalized lipophilic cations are useful for a successful mitochondria-targeted chemotherapy. Thereby, quaternary ammonium salts [24] but also malachite green-derived compounds [27] proved to be significantly less cytotoxic than their rhodamine B analogs [17]. Furthermore, the presence of a rhodamine residue is of crucial importance, which is why we decided to extend our studies to rhodamine B and other rhodamines, and to investigate especially the synthesis and cytotoxic activity of (homo)-piperazinyl-spaced triterpenes holding a rhodamine 101 residue in more detail, and to compare their cytotoxic activity with those carrying a rhodamine B unit.

Results
Acetylation (Scheme 1) of betulinic (BA, Figure 1), oleanolic (OA), ursolic (UA), and platanic (PA) acid gave well known acetates 1-4; their carboxyl group was activated with oxalyl chloride followed by the addition of either piperazine or homopiperazine to furnish amides 5-8 and 9-12, respectively. Activation of rhodamine B or rhodamine 101 with oxalyl chloride and reaction with amides 5-12 furnished piperazine/rhodamine B derived conjugates 13-16 and 17-20 as well as rhodamine 101 derived hybrids 21-24 and 25-28, respectively. All of these conjugates were violet in color, hence, indicating the presence of an intact cationic rhodamine moiety. This is regarded as a prerequisite for obtaining high cytotoxicity due to interaction with the mitochondrial membrane(s).
The cytotoxicity of the compounds was determined in sulforhodamine B (SRB) assays employing several human tumor cell lines (A375, HT29, MCF-7, A2780, FaDu) as well as two non-malignant cell lines (NIH 3T3, HEK293). The results from these assays are summarized in Tables 1-4.    Table 1 shows the results from the SRB assays for the parent compounds and their acetates. Except for BA and UA, all other triterpenoids held EC 50 values > 30 µM (cut-off of the assay) for the cancer cell lines but also for the non-malignant fibroblasts NIH 3T3. Acetates 1-4 showed slightly improved cytotoxicity (except PA derived 4); by-and-large, EC 50 values between 7.2 µM (1 for FaDu cells) and 21.3 (1 for A375 cells) were observed. Highest cytotoxicity was found for 1 and FaDu cells, for 2 with respect to A2780 and for 3 also with A2780 cells, respectively. Interestingly, PA derived acetate 4 was not cytotoxic at all within the limits of the assay.
Significant improvement was observed for the piperazinyl amides (Table 2), and EC 50 values between 1.00 (5 for HT29) and 3.86 (for PA derived 8 and HT29 cells) were determined. Except for the latter, all EC 50 values were smaller than 3 µM. While 5 was cytotoxic with an EC 50 = 1.5 µM for A375 cells, its homopiperazinyl analog 9 was significantly less active (EC 50 = 18.7 µM). However, by-and-large, the cytotoxicity of the homopiperazinyl derivatives 9-12 was of the same order as that of the piperazinyl analogs 5-8.
Thereby, all of the compounds showed high cytotoxicity for all human tumor cell lines; EC 50 values ranged from EC 50 = 0.02 µM (compound 18 and A2780 cells) to EC 50 = 0.76 µM (compound 17 and A375 cells). Thus, the former compound was as cytotoxic as standard doxorubicin.
Previously, we have shown the high cytotoxicity of several rhodamine B conjugates. Hence, it became of interest to investigate whether this high cytotoxicity is limited to rhodamine B conjugates or can also be found in conjugates holding a rhodamine 101 scaffold. As a result (Table 4), conjugates holding either a piperazinyl or homopiperazinyl spacer were only slightly less cytotoxic than those holding a rhodamine B moiety.
Interestingly enough, in this series of compounds, UA derived 27 (carrying a homopiperazine spacer and a rhodamine 101 residue) held the highest cytotoxicity, and the EC 50 values for this compound were as low as EC 50 = 0.05 µM (A2780 cells). Cytotoxicity for non-malignant fibroblasts NIH 3T3 were approximately five times lower for both rhodamine scaffolds. Extra staining experiments of A375 cells (acridine orange (AO), Hoechst 33342, rhodamine 123 ( Figure 2)) showed 26 to act as a mitocan.
A summary of the hitherto known structural prerequisites to obtain pentacyclic triterpenoids of high cytotoxicity is depicted in Figure 3.

Conclusions
Four representative pentacyclic triterpenoic acids (BA, OA, UA, and PA) were selected for a systematic evaluation of cytotoxic derivatives. As a result-and as exemplified for A2780 cancer cells-the cytotoxicity of the parent acids is low but increased slightly upon their acetylation. A significant increase in cytotoxicity was observed when acetates 1-4 were transformed into their piperazinyl amides 5-8. For the latter, compounds EC 50 values between EC 50 = 2.6 to 1.7 µM have been determined. The same trend was observed for the homopiperazinyl derivatives 9-12. Interestingly, betulinic acid derived 9 (EC 50 = 12.0 µM) was significantly less cytotoxic than its piperazinyl derivative 5 (EC 50 = 1.9 µM). A tremendous improvement in cytotoxicity was observed, however, for the rhodamine conjugates, and EC 50 values between EC 50 = 0.05-0.032 µM were observed for the piperazinyl rhodamine B conjugates. The corresponding piperazinyl rhodamine 101 conjugates were of comparable bioactivity (EC 50 = 0.09-0.17 µM). A similar trend was observed for the homopiperazinyl rhodamine conjugates, and EC 50 = 0.02-0.45 µM (for rhodamine B derived 17-20) and EC 50 = 0.05-0.19 µM (for rhodamine 101 derived 25-28) were determined. Thus, it can be concluded that an optimal combination of pentacyclic triterpene, a suitable spacer and a lipophilic cationic residue must be found to achieve good cytotoxic activity. It was shown that both piperazinyl and homopiperazinyl spacers are equally suitable to serve as anchors for the binding of either rhodamine B or rhodamine 101. Furthermore, the conjugates derived from either rhodamine B or rhodamine 101 are of comparable cytotoxicity.

Acridine Orange (AO) Staining
On the first day, A375 cells were counted and seeded 1 × 10 5 in a Petri dish (diameter 4 cm) with coverslips (22 mm × 22 mm) in 2 mL medium. After 24 h, the medium was removed, and treatment was performed with 2 mL of new medium (control) and 2 mL each of 2 times the EC 50 concentration of compounds 27. After 24 h, the medium was removed from the samples, the coverslip was rinsed with 1 mL of PBS (with Ca 2+ and Mg 2+ ), placed on a slide containing 20 µL of AO solution (2.5 µg/mL in PBS), and measured directly on the fluorescence microscope.

Hoechst 33,3342 and Rhodamine 123 Staining
On day 1, A375 cells were counted and seeded 1 × 10 5 in a Petri dish (diameter 4 cm) with coverslips (22 mm × 22 mm) in 2 mL medium. After 24 h, the medium was removed, and treatment was performed with 1 mL of new medium containing 1 µL of compound 27 (0.08 mM solution). After another 24 h, additional treatment was performed with 1 µL rhodamine (1 mg/mL in EtOH) and 2 µL Hoechst 33342 (100 µg/mL in DMSO) for (at least) 30 min. The medium was then removed, rinsed once with PBS (with Ca 2+ and Mg 2+ ), placed on a slide containing 20 µL PBS (with Ca 2+ and Mg 2+ ), and measured directly on the fluorescence microscope.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.