All commercially available chemicals were purchased from Aldrich and used without further purification.

Human fibroblast cells BJ-hTERT and the human cancer cell lines A549 (non-small human lung cancer, NSCLC), HBL-100, MCF-7 and T-47D (breast cancer), as well as HeLa (cervical cancer), were provided by Dr. Raimundo Freire (Hospital Universitario de Canarias, Tenerife, Canary Islands). The NSCLC cell line SW1573 and the colon cancer cell line WiDr were provided by Prof. G. J. Peters (Cancer Center Amsterdam, Vrije Universiteit, Amsterdam, The Netherlands). The breast cancer cell line MDA-MB-453 was provided by CEAMED S. A. (Tenerife, Canary Islands). Cells were grown in RPMI-1640 medium containing 5% fetal bovine serum (FBS), 2 mM L-glutamine, 100 U/mL of penicillin G and 0.1 mg/mL of streptomycin in a 37 ºC, at a 95% humidified atmosphere of 5% CO₂. Cells were maintained in culture in 60 mm cell culture dishes in growth medium (10 mL) and passaged twice weekly.

Cell line suspensions were counted with a Moxi Z automated cell counter (Orflo Technologies, Ketchum, ID 83340, USA) and diluted to reach the appropriate cell densities for inoculation onto 96-well plates. After 24 hours of equilibration in growth medium, the compounds were added at increasing concentrations (0.01-100 μM). The maximal test concentration was set at 100 μM. Cisplatin was used as a positive control of a conventional cytotoxic agent and DMSO (0.25% v/v) was used as negative control. Drug incubation times were 48 hours, following which monolayer cells were fixed onto the wells using 25 μL ice-cold trichloroacetic acid solution (50% w/v) for 60 min at 4 ^oC. Then, the plates were rinsed with water, following which 25 μL of a sulforhodamine B (SRB) solution (0.4% w/v in 1% acetic acid) was added for 15 min. Unbound SRB was rinsed with 1% acetic acid. Cell-bound SRB was dissolved with 150 μL of a 10 mM Tris solution at pH 10.5. The optical density of each well was determined at 530 and 620 nm using PowerWave XS microplate reader. This dual wavelength was used to reduce optical interference caused by scratches, fingerprints or other matters that equally absorb light at both wavelengths. The anti-proliferative activity, expressed as 50% growth inhibition (GI₅₀), was calculated according to NCI formulas [15].

The 3D structure of the transporter SLC6A14 is not available in Protein Data Bank. Therefore, we used a homology modelling procedure. We selected the Drosophila’s dopamine transporter in complex with cocaine (PDB ID: 4XP4, resolution of 2.8 Å) as a template. We used HyperChem software (Hypercube Inc., FL, USA) to obtain and optimize the molecular structures and OpenBabel application to convert the files into adequate format before exporting them to docking software Glide v8.2 (Schrödinger, MA, USA). We used “Standard Precision” protocol in Glide with OPLS 2005 force field with post-docking minimization.

The Betti bases reported in the current study were prepared according to our general synthetic pathway [16], which is outlined in Fig. 2. In this multicomponent organic synthesis process, 2-naphthol (11) was reacted with primary or secondary cyclic amines and diverse aromatic aldehydes (see Table 1 for residues). This one-pot multicomponent Betti synthesis was performed under green heterogeneous and neat conditions in the presence of Montmorillonite K30 catalyst, at 60 °C and in relatively short reaction times. Using this method, a small and focused library of Betti bases (14a-w) was prepared in good to excellent yield (50-92%). The “greenness” of a chemical reaction can be quantified by calculating both the E-factor and the atom economy, which for our process were 3.92 and 94%, respectively.

Figure 2.

General procedure for the synthesis of Betti bases. Reagents and conditions: (a) Montmorillonite K30, neat, 60 ºC or r.t. See Table 1 for amine and R definitions.

The in vitro anti-proliferative activity of the twenty-three Betti bases 14a-w library was studied using our implementation of the NCI protocol [17]. The compounds were screened against a panel of representative human solid tumor cell lines comprised of A549 (NSCLC), HBL-100 and T-47D (breast cancer), HeLa (cervical cancer), SW1573 (alveolar cell carcinoma), and WiDr (colon). The results of the 50% growth inhibition (GI₅₀) values are shown in Table 1. Taken collectively, the data show that eighteen out of twenty-three compounds were able to inhibit cell growth in all cell lines with GI₅₀ values in the range 4-79 µM. This initial results allow considering Betti bases as privileged structures, with inherent drug-likeness, that can represent an ideal source of a core scaffold for the design and synthesis of combinatorial libraries that could be targeted at various cellular targets including for example receptors, enzymes and other proteins.

From this set of compounds, five compounds displayed GI₅₀ values at the low micromolar range (< 10 µM) against at least one cell line. In terms of cytotoxic potency, we could classify them in the following order 14t > 14j > 14k > 14n-o. From the GI₅₀ values we observed that only two compounds were inactive (GI₅₀ > 100 µM) in all cell lines studied (14a and 14u), while 14n,14c and 14d displayed anti-proliferative effects against five, four and two cell lines, respectively.

When considering the substitution pattern of compounds 14a-w, we could not infer a clear structure-activity relationship (SAR). Furthermore, the data set for each cell line did not provide a SAR trend. Furthermore, a direct comparison between the GI₅₀ data obtained in the breast cancer cells (HBL-100 ad T-47D) drew our attention. As a rule of thumb, HBL-100 cells tended to be more sensitive to drug exposure than T-47D cells. However, fourteen of the Betti bases were less active in HBL-100 cells, whilst seven followed the opposite trend.

In a previous study with tryptophan dipeptides we observed that the compounds were inactive against HBL-100 cells and active towards T-47D [18]. This result was potentially consistent with the fact that T-47D cells overexpress SLC6A14. This protein is a cell membrane transporter that recognizes tryptophan as a substrate with high affinity [19]. In this respect, it was previously shown that the blockade of SLC6A14 induces cell death in SLC6A14-positive tumors and represents the first nutrient transporter to be exploited as a drug target for possible cancer therapeutics [20]. This transporter is markedly up-regulated in some types of carcinomas including breast [21], cervix [22], colon [23] and pancreas [24]. In particular, the blockade of SLC6A14 mainly affects estrogen receptor positive (ER+) breast cancer cells. Therefore, we decided to further explore the activity of a subset of our Betti bases in other breast cancer cell lines.

In HBL-100 cells, the expression of SLC6A14 is undetectable and tryptophan analogues do not affect cell growth [25]. Based on this premise, we selected the pyrrolidine derivatives 14c-d as representative examples of Betti bases that displayed activity in T-47D cells but inactivity in HBL-100 cells. In addition, we also studied the corresponding piperidine siblings 14n-o for comparison purposes. As a model to run the study, we selected cell lines that represent the three main types of breast cancer according to the status of estrogen receptor (ER), progesterone receptor (PR) and the human epidermal growth factor receptor 2 (HER2); cells with none of these receptors are knowns as triple negative breast cancer (TNBC) [26]. The results are shown in Table 2. With the exception of 14d, all compounds were less active against the HER2+ cell line and MCF-7 cells (ER+), when compared to T-47D cells. We speculate that factors other than blockade of SLC6A14 could explain these observed effects. For instance, T-47D cells reportedly harbor a mutant p53, whilst MCF-7 and MDA-MB-453 cells possess wild type p53 [27].

Furthermore, we checked the selectivity towards healthy human cells using as a model the human fibroblast cell line BJ-hTERT. The results indicate that the Betti bases 14c-d and 14n-o exhibited good selectivity, since they did not affect cell growth of BJ-hTERT human fibroblast cells (GI₅₀ > 100 µM).

Next, we decided to run computational assays to predict the possible mode of interaction of the Betti bases with the SLC6A14 transporter and compared these results to α-MT. We did a bibliographic search of the SLC6A14 gene, the biological transporter of tryptophan. The 3D structure of this transporter is not available in Protein Data Bank, the single worldwide archive of structural data of biological macromolecules. Therefore, we used a homology modelling procedure. We used the amino acid sequence of SLC6A14 to search for an indexed sequence as similar as possible to ours. Thus, we selected the Drosophila’s dopamine transporter in complex with cocaine (PDB ID: 4XP4, resolution of 2.8 Å) as a template. This latter transporter also belongs to the solute carrier superfamily and shows 46% amino acid identity and 62% similarity to SLC6A14. Most recently, structural modeling of the human SLC6A14, docking and molecular dynamics studies were undertaken; this study unraveled novel aspects of the human SLC6A14 structure-function relationship, thereby having important implications for cancer treatment through the future design of novel inhibitors of SLC6A14-mediated transport [28].

Herein, both α-MT and Betti bases 14 are racemic mixtures. The tragic example of thalidomide marked a turning point in drug development and revealed that the physicochemical and biochemical properties of racemic mixtures and individual stereoisomers can differ significantly. Therefore, we analyzed independently the binding of both enantiomeric series. The docking scores are given in Table 3. In all cases under study, the calculated docking interaction energy is comparable for all compounds and for both enantiomeric series. This apparent lack of selectivity could be attributed to the fact that SLC6A14 not only takes up the 18 of the 20 proteinogenic amino acids, but also d-amino acids such as D-serine, D-alanine, D-methionine, D-leucine and D-tryptophan [29].

Remarkably, the Betti bases 14c-d displayed comparable docked conformations when superimposed with α-MT, as shown in Fig. 3. These results might support the speculation that Betti bases could behave as tryptophan mimetics, and thus block the transporter SLC6A14 and disrupt cellular uptake of amino acids. From the anti-proliferative data, one cannot rule out the possibility that the compounds inhibit cell growth by mechanisms other than interaction with SLC6A14. However, it is reasonable considering that the Betti bases shown here might serve as a starting point to develop inhibitors of SLC6A14. In support of this notion, the group of Vadivel Ganapathy has recently shown that deletion of the SLC6A14 gene suppresses tumor growth in a mouse model of breast cancer [30].

Figure 3.

Dock superimposed pose of α-MT (pink) with 14c (blue) and 14d (yellow).