Synthesis and Antiproliferative Activity of Thiazolyl-bis-pyrrolo[2,3-b]pyridines and Indolyl-thiazolyl-pyrrolo[2,3-c]pyridines, Nortopsentin Analogues

Two new series of nortopsentin analogues, in which the imidazole ring of the natural product was replaced by thiazole and indole units were both substituted by 7-azaindole moieties or one indole unit was replaced by a 6-azaindole portion, were efficiently synthesized. Compounds belonging to both series inhibited the growth of HCT-116 colorectal cancer cells at low micromolar concentrations, whereas they did not affect the viability of normal-like intestinal cells. A compound of the former series induced apoptosis, evident as externalization of plasma membrane phosphatidylserine (PS), and changes of mitochondrial trans-membrane potential, while blocking the cell cycle in G2/M phase. In contrast, a derivative of the latter series elicited distinct responses in accordance with the dose. Thus, low concentrations (GI30) induced morphological changes characteristic of autophagic death with massive formation of cytoplasmic acid vacuoles without apparent loss of nuclear material, and with arrest of cell cycle at the G1 phase, whereas higher concentrations (GI70) induced apoptosis with arrest of cell cycle at the G1 phase.

The spacer structural profile can vary from a linear chain to carbocycles or heterocycles differently sized. Thus, coscinamides A-C, isolated from deep marine sponge Coscinoderma sp. bearing a linear chain as a spacer, showed HIV inhibitory activity (Chart 1) [17]. Asterriquinone, isolated from Aspergillus fungi, bears a quinone symmetrical structure and showed in vivo activity against Ehrlich carcinoma, ascites hepatoma AH13, and mouse P388 leukemia [18]. Dragmacidins have been isolated from a large number of deep water sponges such as Dragmacidon, Halicortex, Spongosorites, Hexadella and the tunicate Didemnum candidum, and present different spacers and diverse related activities. Dragmacidin and dragmacidins A-C, which exhibit the saturated six-membered heterocyclic link piperazine, showed modest cytotoxic activity [19][20][21]. Instead, a more complex member of this family, dragmacidin D, bearing a pyrazinone moiety as central core, exhibited several biological properties such as inhibition of serine-threonine protein phosphatases, antiviral, antimicrobial, and anticancer activities [22,23].
Nortopsentins A-C also bearing the imidazole ring as a spacer, were isolated from Spongosorites ruetzleri, and showed in vitro cytotoxicity against P388 cells (GI50, 4.5-20.7 μM). Replacement of the indole nitrogen with a methyl group led to derivatives that showed a significant improvement in cytotoxicity against P388 cells (GI50, 0.8-2.1 μM) [26,27].
Due to a great limitation in the use of the reservoir of marine organism that allow the isolation of very small amount of the biologically active substances from the natural material, several total synthesis of nortopsentins were proposed [28][29][30][31].
Moreover, due to the considerable activities shown, indolyl alkaloids have become an attractive field in medicinal chemistry and several dragmacidin analogues bearing six membered rings such as pyridine, pyrazine, pyrazinone and pyrimidine as spacer between the two indole units were synthesized. These analogues showed good antiproliferative activity against a wide range of human tumor cell lines [32][33][34][35]. Many papers reported the synthesis and the evaluation of the antiproliferative activity of nortopsentin analogues bearing five membered heterocycles which replaced the imidazole ring of the natural product such as bis-indolyl-thiophenes [36], -pyrazoles [37], -furans [38], -isoxazoles [38], -pyrroles [39], and -1,2,4-thiadiazoles [40]. Most of these analogues exhibited good antiproliferative activity against wide range of human tumor cell lines often reaching GI50 values at submicromolar level.  Moreover, the structural manipulation of the natural nortopsentins, beside the heterocyclic spacer, was extended to one or both indole units and led to 3-[(2-indolyl)-5-phenyl]pyridines and phenylthiazolyl-7-azaindole derivatives. Both these series of compounds showed antiproliferative activity against a wide range of human tumor cell lines in the micromolar-submicromolar range and were able to inhibit the activity of the cyclin-dependent kinase 1 (CDK1) with GI50 values lower than 1 μM [41,42].
All compounds of this series were tested against the NCI full panel of human cancer cell lines and STO and MesoII cells, derived from human diffuse malignant peritoneal mesothelioma (DMPM). The most active compounds, that also act as CDK1 inhibitors, consistently reduced DMPM cell proliferation and induced a caspase-dependent apoptotic response, with a concomitant reduction of the expression of the active Thr 34 -phosphorylated form of the anti-apoptotic protein survivin. Moreover, the combined treatment of DMPM cells with the most active derivative and paclitaxel produced a synergistic cytotoxic effect, which was parallel by an enhanced apoptotic response. In the mouse model, i.p. administration of active derivatives was effective, resulting in a significant tumor volume inhibition of DMPM xenografts (range, 58%-75%) at well-tolerated doses, and two complete responses were observed in each treatment group [44].
In this paper, continuing our studies on indole and related azaindole systems [45][46][47][48][49], we report the synthesis of derivatives of type 1, nortopsentin analogues in which both indole units are replaced by a 7-azaindole moiety, and of derivatives of type 2-4 in which one indole unit was replaced by a 6-azaindole moiety. We also report the antiproliferative activity of these new nortopsentin analogues and studies directed to elucidate their mode of action.

Biology
All the synthesized thiazoles 1a-l, 2a-j, 3a-f and 4a-f were evaluated by the National Cancer Institute (Bethesda MD) for cytotoxicity against the NCI-60 cell line panel using standard protocols [51]. Initially, all derivatives were tested at a single dose (10 −5 M) on the full panel of approximately 60 human tumor cell lines derived from 9 human cancer cell types, that have been grouped in disease sub-panels including leukemia, non-small-cell lung, colon, central nervous system (CNS), melanoma, ovarian, renal, prostate and breast cancers (data not shown). Compounds 1k and 4c were further selected for full evaluation at five concentration levels (10 −4 -10 −8 M).
The antitumor activity of compounds 1k and 4c was given by three parameters for each cell line: GI50 (GI50 is the molar concentration of the compound that inhibits 50% net cell growth), TGI (TGI is the molar concentration of the compound leading to total inhibition of net cell growth), and LC50 (LC50 is the molar concentration of the compound that induces 50% net cell death). The average values of mean graph midpoint (MG_MID) were calculated for each of these parameters. An evaluation of the data reported in Table 2 pointed out that compounds 1k and 4c exhibited antiproliferative activity against most of the human cell lines at GI50 values from micromolar to submicromolar (0.81-27.7 and 0.93-4.70 µM respectively). The indolyl-thiazolyl-pyrrolo[2,3-c]pyridine derivative 4c resulted more active than thiazolyl-bispyrrolo[2,3-b]pyridines derivative 1k in terms either of GI50 (mean value 2.59 and 5.26 µM respectively) and percentage of sensitive cell lines out of the total number of cell lines investigated (100% and 89%, respectively).
Moreover for compound 4c positive TGI and LC50 values were observed with respect to a good number of cell lines (98% and 81%, respectively; data not shown).
Derivative 4c was shown to be selective with respect to the renal cancer subpanel having all the subpanel cell lines GI50 in the range 0.93-4. 16  Derivative 1k showed selectivity with respect to MDA-MB-468 (GI50 0.81 µM ) of breast cancer, EKVK (GI50 1.27 µM ) of non small cell lung, and CAKI-1 (GI50 1.56 µM ) of renal cancer subpanel.
Selective toxicity of 1k and 4c nortopsentin analogues towards tumor cells was investigated. To this aim human HTC-116 colorectal carcinoma cells, against which both compounds exhibited a comparable antiproliferative effects (GI50 2.91 and 2.35 μM, Table 2), and intestinal normal-like differentiated Caco-2 cells were exposed to the compounds for 24 h and viability compared. While both derivatives, in the range 5 to 100 μM, dose-dependently inhibited the intestinal HCT-116 cell proliferation, they did not affect the differentiated Caco-2 cell viability, suggesting tumor cells as the main target of their cytocidal action ( Figure 1). GI50 values of 1k or 4c calculated after 24 h treatment of HCT-116 cells, were 33.55 ± 2.31 μM and 13.15 ± 0.95 μM, respectively. Taking in account the GI50 values measured by NCI on the same cell line after 72 h treatment, the cytotoxic effects of the nortopsentin analogues appeared time-dependent.

Cell Cycle Alterations
We next determined alterations in the cell cycle caused by derivatives 1k and 4c in colorectal cancer cells. Flow cytometry analysis of nuclear DNA content after 24 h treatment of HCT-116 cells is reported in Figure 2A. Drug concentrations were chosen on the basis that they represent values above and below the respective GI50 values. Compared to control cells, 1k caused a dose-dependent accumulation of cells in G2/M phase, paralleled by a reduction in the percentage of cells in the G1 phase and by a significant increase of cells in the sub-G1 phase, which is representative of cells with fragmented DNA. On the other hand, 4c induced a dose-dependent accumulation of cells in G1 phase accompanied by a decrement in the percentage of cells in G2/M phases. Moreover, accumulation of sub-G1 population was significantly higher than control (p < 0.05) only at high concentration of the drug. These results indicated that the two nortopsentin analogues caused arrest of the HCT-116 cancer cell growth involving different check points of the cell cycle.

Cell Death
To determine whether HCT-116 cells undergo apoptosis upon treatment with the nortopsentin analogues, cells were treated with 1k or 4c for 24 h, stained with both propidium iodide (PI) and Annexin V-fluorescein isothiocyanate (FITC), and analyzed by flow cytometry.
Neither compound caused cell necrosis ( Figure 2B). Rather, while the percentage of cells in late apoptosis increased at the increase of the 1k doses, apoptotic effects of 4c were evident only at high concentrations (GI70), when cells in early apoptosis appeared significantly increased with respect to control (p < 0.05).
Mitochondria play a critical role in regulating the apoptotic machinery. We then examined mitochondrial membrane potential (Δψm) loss using DiOC6, a fluorescent mitochondria-specific and voltage-dependent dye. As indicated by the decrement in DiOC6 green-associated fluorescence, treatment of HCT-116 cells with 1k, for 24 h, induced a remarkable dose-dependent dissipation of Δψm ( Figure 3).
On the other hand, incubation with 4c did not cause mitochondrial dysfunction at GI30, whereas induced a significant increase in Δψm loss at higher concentrations. Overall our findings indicated that, although the nortopsentin derivatives inhibited the HCT-116 tumor cell growth, they may elicit different molecular pathways of programmed cell death. Morphology of HTC-116 cells treated for 24 h with nortopsentin analogues 1k and 4c was assessed by microscopy analysis after Giemsa staining. Both compounds caused an evident dose-dependent reduction of the cell population with respect to control ( Figure 4A). However, whereas 1k caused highly condensate cells, as a sign of their apoptotic fate, more complex alterations were observed after treatment with 4c. A low concentration of drug (GI30) caused evident expansion of whole cytoplasm, with a massive accumulation of multiple-membrane bounded vacuoles without apparent loss of nuclear material, a morphology characteristic of autophagic cell death. On the other hand, when HCT-116 cells were treated with higher concentration of 4c (GI70), vacuolization was less evident and cells with condensed morphology prevailed, indicating evolution of the cell fate towards apoptosis. Then, we performed FACS analysis of acridine-orange stained acid vacuoles (AVO) using the red to green ratio as an indicator of acid vacuolar organelle accumulation [52]. As shown in Figure 4B, treatment of the cells with 4c (GI30), for 24 h, led to a 10-fold increase of the bright red to green fluorescence intensity ratio of acrydine orange, indicating an elevated induction of cell vacuolization associated to autophagy. Instead, the treatment of cells with the highest drug concentration did not significantly modify the percent of cells with AVO with respect to control, supporting the concept of a different cell death. A comparison of the antiproliferative activity of the natural products nortopsentins A-C with 1k and 4c, the most active derivatives of the two new series of analogues, cannot be performed since nortopsentins were tested against only one cell line (murine leukemia P388), whereas, compounds 1k and 4c showed to be active against the 89% and 100% respectively, of the tested cell lines belonging to the NCI full panel of human cancer cell lines (ca. 60) with GI50 values in the low micromolar range.
A comparison of the antiproliferative activity of 1k and 4c with the series of nortopsentin analogues previously reported reveals that these compounds showed to be active against a very large range of human cell lines at the same level of both the analogues in which the central heterocycle was replaced by thiophene, pyrazole, furan, isoxazole, and pyrrole [33][34][35][36][37][38][39]; and those series in which also the indole units were substituted by azaindole systems [42][43][44].
The feature that makes these series of analogues interesting was revealed by the investigations into the cell cycle alterations and cell death modes which suggest that these derivatives inhibiting the HCT-116 tumor cell growth may elicit different molecular pathways of programmed cell death. In fact, 1k induced apoptosis, and 4c exhibited different responses depending on the dose: low concentrations induced morphological changes typical of autophagic death with extensive formation of cytoplasmic acid vacuoles without apparent loss of nuclear material, and with arrest of cell cycle at the G1 phase, instead, higher concentrations induced apoptosis with arrest of cell cycle at the G1 phase.

General
All melting point were taken on a Büchi-Tottoly capillary apparatus and are uncorrected. IR spectra were determined in bromoform with a Shimadzu FT/IR 8400S spectrophotometer. 1 H and 13 C NMR spectra were measured at 200 and 50.0 MHz, respectively, in DMSO-d6 or CDCl3 solution, using a Bruker Avance II series 200 MHz spectrometer. Compounds 2d-f,h,j, and 3e,f were characterized only by 1 H NMR spectra, because of their poor solubility the 13 C spectra were not performed. Column chromatography was performed with Merk silica gel 230-400 mesh ASTM or with Büchi Sepacor chromatography module (prepacked cartridge system). Elemental analyses (C, H, N) were within ±0.4% of theoretical values and were performed with a VARIO EL III elemental analyzer. Purity of all the tested compounds was greater than 98%, determined by HPLC (Agilent 1100 Series).

Synthesis of 1-
To a cold solution of 5a (1.0 g, 7.0 mmol) in anhydrous toluene (20 mL), t-BuOK (1.1 g, 9.5 mmol) and TDA-1 (1 or 2 drops) were added at 0 °C. The reaction mixture was stirred at room temperature for 6 h and then MeI (7.0 mmol, 0.44 mL) was added at 0 °C. TLC analysis (DCM/ethyl acetate 9/1) revealed that methylation was complete after 1 h. The solvent was evaporated under reduced pressure. The residue was treated with water, extracted with DCM (3 × 20 mL), dried (Na2SO4), evaporated and purified by column chromatography using DCM/ethyl acetate (9/1) as eluent. White solid; yield 85%; A solution of appropriate carbonitriles 5a,b (6.3 mmol) in concentrate sulphuric acid (3.32 mL) was stirred at room temperature for 15-60 min. The solution was slowly poured into ice and basified with concentrated NaOH. The residue was extracted with ethyl acetate (3 × 20 mL), dried (Na2SO4) and concentrated under reduced pressure to afford the carboxamides 6a,b. To a cold solution of appropriate pyrrolo-pyridines 8a,c,e (2.5 mmol) in anhydrous toluene (25 mL), t-BuOK (0.38 g, 3.4 mmol) and TDA-1 (1 or 2 drops) were added at 0 °C. The reaction mixture was stirred at room temperature for 3 h and then MeI (2.5 mmol, 0.2 mL) was added at 0 °C. TLC analysis (ethyl acetate) revealed that methylation was complete after 1 h. The solvent was evaporated under reduced pressure. The residue was treated with water, extracted with DCM (3 × 20 mL), dried (Na2SO4), evaporated and purified by column chromatography using DCM/ethyl acetate (9/1) as eluent to give derivatives 8b,d,f [42].

2-Bromo-1-(1H-pyrrolo[2,3-b]pyridin-3-yl)pthanone (9a)
White solid; yield: 92%; mp: 280-282 °C; Analytical and spectroscopic data were previously reported [42].  To a solution of appropriate pyrrolo[2,3-c]pyridines 21, 22 (2.5 mmol) in 10 mL of anhydrous DCM, anhydrous AlCl3 (1.2 g, 8.8 mmol) was slowly added. The reaction mixture was heated under reflux and BrCOCH2Br (2.5 mmol, 0.2 mL) in 2 mL of anhydrous DCM was added dropwise. The resulting solution was allowed to stir under reflux for 40 min. After cooling, water and ice were slowly added and the obtained precipitate (for derivative 23) was filtered off or the oil residue (for derivative 24) was extracted with DCM (3 × 20 mL) and purified by column chromatography using DCM/ethyl acetate 9/1 as eluent. (100 U/mL), streptomycin (100 μg/mL) and gentamicin (5 μg/mL). Cells were maintained in log phase by seeding twice a week at a density of 3 × 10 8 cells/L in humidified 5% CO2 atmosphere, at 37 °C. In all experiments, HCT-116 cells were made quiescent through overnight incubation before treatment with tested compounds or vehicle alone (control cells), while Caco-2 cells were treated 15 days after confluence, at which time the cells are differentiated in normal intestinal-like cells [53]. No differences were found between cells treated with DMSO 0.1% and untreated cells in terms of cell number and viability.
Cytotoxic activity of the nortopsentin derivatives 1k and 4c was determined by the colorimetric assay based on the reduction of 3-(4,5-dimethyl-2-thiazolyl)bromide-2,5-diphenyl-2H-tetrazolium (MTT) to purple formazan by mitochondrial dehydrogenases. Briefly, HCT-116 and Caco-2 lines cells were seeded at 2 × 10 4 cells/well in 96-well plates containing 200 μL RPMI. When appropriated, monolayer cultures were treated for 24 h with various concentrations (5-100 μM) of the drugs. Then cells were washed with fresh medium and 50 μL FBS-free medium containing 5 mg/mL MTT added. Cells were incubated 2 h at 37 °C, then medium was discarded by centrifugation, formazan blue formed in the cells dissolved in DMSO, and absorbance measured at 570 nm in a microplate reader (Bio-RAD, Hercules, CA, USA). Formazan of control cells was taken as 100% viability. GI50 was calculated by the curve of percent viability vs. concentration. Each experiment was repeated at least three times in triplicate.

Cell Cycle Analysis
Cell cycle stage was analyzed by flow cytometry. HCT-116 cells (5.0 × 10 4 cells/cm 2 ) were seeded in triplicate in 24-wells culture plates. After an overnight incubation, the cells were washed with fresh medium and incubated with compounds 1k and 4c in RPMI for 24 h. Then cells were harvested by trypsinization. Aliquots of 1 × 10 6 cells were washed with PBS and incubated in the dark in a PBS solution containing 20 μg/mL propidium iodide (PI) and 200 μg/mL RNase, for 30 min, at room temperature. Then samples of at least 1.0 × 10 4 cells were subjected to fluorescence-activated cell sorting (FACS) analysis by Epics XL™ flow cytometer using Expo32 software (Beckman Coulter, Fullerton, CA, USA).

Measurement of Phosphatidylserine (PS) Exposure
The apoptosis-induced PS externalization to the cell surface was measured by flow cytometry by double staining with Annexin V-Fluorescein isothiocyanate (Annexin V-FITC)/propidium iodide (PI). Annexin V binding to phosphatidylserine is used to identify the earliest stage of apoptosis. PI, which does not enter cells with intact membranes, is used to distinguish between early apoptotic cells (Annexin V-FITC positive and PI negative), late apoptotic cells (Annexin V-FITC/PI-double positive) or necrotic cells (annexin V-FITC negative and PI positive). After 24 h treatment, HCT-116 cells were harvested by trypsinization and adjusted at 1.0 × 10 6 cells/mL with combining buffer according to the manufacturer instructions (eBioscience, San Diego, CA, USA). One hundred μL of cell suspensions were added to a new tube, and incubated with Annexin V-FITC and PI solution at room temperature in the dark for 15 min. Then samples of at least 1.0 × 10 4 cells were subjected to FACS analysis using appropriate 2-bidimensional gating method.

Measurement of Mitochondrial Transmembrane Potential
Changes of mitochondrial transmembrane potential (Δψm) were assessed by flow cytofluorometry, using the cationic lipophilic dye 3,3′-dihexyloxacarbocyanine iodide(3) (DiOC6) (Molecular Probes, Inc., Life Technologies Italia, Monza, Italy) which accumulates in the mitochondrial matrix. Changes in mitochondrial membrane potential are indicated by a reduction in the DiOC6-induced fluorescence intensity. After 24 h treatment, HCT116 cells were incubated with DiOC6 at a 40 nmol/L final concentration, for 15 min at 37 °C. After centrifugation, cells were washed with PBS and suspended in 500 μL PBS. Fluorescent intensities were analysed in at least 1 × 10 4 cells for each sample.

Morphology
To analyse cell morphology, cells were fixed with methanol for 10 min and then stained with Giemsa (10% in PBS) for 15 min followed by washing with distilled water. Cell images were captured using Zeiss, AxioCam, AxioSkop microscope, (West Germany) with 40× lenses.

Quantification of Acidic Vesicular Organelles (AVO) by Acridine Orange (AO) Staining
AO is a fluorescent molecule used either to identify apoptotic cell death or autophagy. It can interact with DNA emitting green fluorescence or accumulate in acidic organelles where it is protonated forming aggregates that emit bright red fluorescence [52]. Briefly, cells were stained with a AO solution (5 μg/mL) for 15 min. Then they were washed, re-suspended in PBS and subjected to FACS analysis. The green (510-530 nm, FL-1) and red (650 nm, FL-3) fluorescence of AO with blue (488 nm) excitation was determined over 10,000 events.

Statistics
Results are given as means and their standard deviations. Three independent observations were carried out for each experiment, replicated three times. Statistical comparisons were made using a one-way ANOVA, followed by Bonferroni's test. p < 0.05 was considered statistically significant.

Conclusions
Two new series of nortopsentin analogues were conveniently prepared. In these series the imidazole ring of the natural product was replaced by a thiazole, and, in one case, both indole units were substituted by 7-azaindole moieties; in the other series, one indole moiety was replaced by a 6-azaindole system.
The newly synthesized nortopsentin analogues 1k and 4c showed cytotoxicity towards a broad spectrum of human cancer cell lines included in the NCI panel. Importantly, the compounds did not appreciably impair vitality of intestinal normal like cells. Investigating mechanisms underlying the antiproliferative effect in HCT-116 colon cancer cells, we showed that 1k caused a dose-dependent increase of the apoptotic cell population, activating the mitochondria-mediated pathway and inducing cell cycle arrest at the G2/M phase. Comparably, previous work showed that other similar analogues caused apoptosis and arrest of cell cycle with inhibition of cyclin-dependent kinase 1 activity kinase [42][43][44]. On the other hand 4c at a concentration lower than its GI50, induced antiproliferative effects with massive accumulation of autophagic vacuoles without apparent signs of apoptosis. The observed arrest of cell cycle at G1 phase, supports the autophagic fate of the cells [54]. Evolution of the cells toward apoptotic death following treatment with higher drug concentrations, suggests that 4c may orchestrate a potential axis of autophagy and apoptosis in HCT-116 cells which can facilitate cellular destruction. The autophagy-signaling cascade induced by 4c in HCT-116 cells is currently under investigation. Modulating authophagy appears of great interest in cancer. Indeed, current evidence suggests that authophagic cell death can be induced as an alternative to apoptosis with therapeutic purpose in cancer cells that are resistant to apoptosis. In this context, 4c may candidate as a lead compound for nortopsentin derivatives with autophagic activity.