Selective inhibition of cancer cell lines is a major advantage for anticancer agents, as non specific cytotoxicity in primary cells can cause undesirable side effects preventing FDA drug approval [10,30]. The two D. orbita fractions were tested for their cytotoxicity against three reproductive cancer cells lines and primary-derived human reproductive granulosa cells (HGC) using the crystal violet and MTT assay. Tyrindoleninone (fraction one) was more active against the reproductive cancer lines than the HGC and inhibited the metabolic function of the cells (MTT assay) before cell death occurred (Crystal violet assay; Table 1). Tyrindoleninone selectively reduced the viability of all three reproductive cancer cell lines KGN (IC₅₀ = 39 µM MTT and 156 µM Crystal violet), JAr (IC₅₀ = 39 µM MTT and 234 µM Crystal violet) and OVCAR-3 (IC₅₀ = 156 µM MTT and 313 µM Crystal violet) at much lower concentrations than the corresponding primary cells (IC₅₀ = 3516 µM MTT and >1953 µM Crystal violet) after just 4 h exposure (Table 1). After 24, 48 and 72 h treatment, tyrindoleninone was significantly more cytotoxic towards all three reproductive cancer cell lines, with IC₅₀ values approximately 10-fold lower than for the HGC (Table 1). Benkendorff et al. [30] showed a 50% reduction in the MCF-7 breast, CaCo₂ colon and U937 lymphoma carcinoma cell viability after 4 h exposure to 391 µM tyrindoleninone in an impure extract compared to untransformed mononuclear cells. Another study predicted a 50% reduction in lymphoma cell viability after 1 h exposure to 4 µM tyrindoleninone [31]. Furthermore, Vine et al. [31] identified that tyrindoleninone was much less cytotoxic to untransformed human mononuclear cells with IC₅₀ = 195 µM after 1 h exposure. The different inhibitory concentrations in the various studies is likely to be due to the relative purity of tyrindoleninone in the extract tested, as well as innate differences in the susceptibility of different cell lines. Furthermore, Vine et al. [31] noted that a contaminating plasticizer in their study may have increased the solubility and availability of tyrindoleninone to the cells. Cell types vary with regards to their metabolic rate and cell culture conditions (such as cell density and passage numbers) and this can also influence the outcome of an assay [42].

6-Bromoisatin, while not as potent as tyrindoleninone, still significantly decreased cell numbers in all three reproductive cancer cell lines, whilst having only a minimal effect on HGC (Table 1). The JAr cell line was the most sensitive after shorter incubation periods, with cell numbers halved at 223 µM (Crystal violet and MTT assay) and 178 µM and 446 µM (Crystal violet and MTT assay) of 6-bromoisatin after 4 and 24 h exposures respectively (IC₅₀; Table 1). In comparison, the HGC numbers were only reduced by >2232 µM 6-bromoisatin after 24 h exposure. After 48 and 72 h exposure to 6-bromoisatin the concentration that inhibited 50% of KGN and OVCAR-3 cell lines was much lower (22 µM 48 and 72 h KGN and 89 and 22 µM OVCAR-3) as determined by the MTT assay. Vine et al. [31] have demonstrated that a range of substituted isatins including 5 and 7-dibromoisatin are 10× more active against the lymphoma cell lines, U937 and Jurkat than 6-bromoisatin. Nevertheless, the results from this study and previous research on the anticancer properties of these marine compounds in particular tyrindoleninone, all support the specificity of these compounds towards rapidly dividing cancer cell lines over freshly isolated healthy human cells.

Cell death by necrosis leads to the loss of cell membrane integrity and uncontrolled release of the cellular contents into the surrounding tissue causing inflammation [43] and as such is often considered a toxic process in comparison to apoptosis [44]. It therefore is an advantage to identify anticancer agents that specifically induce apoptosis as opposed to necrosis. After exposure of the granulosa cancer cells KGN, and primary HGC to tyrindoleninone and 6-bromoisatin, caspase-3 and -7 enzymes were activated at much lower concentrations than those required to cause disruption of membrane integrity, indicating cell death occurred predominately by apoptosis as opposed to necrosis. For example, a four hour exposure to 1953 µM (0.5 mg/mL) tyrindoleninone (Figure 2) was required to cause LDH release in both HGC and the corresponding KGN granulosa cell line, with more LDH being released by KGN (1612 ± 771) than primary HGC (876 ± 1084). This effect was more marked after a 24 h exposure, when lower concentrations (195 µM; 0.05 mg/mL) caused LDH release from KGN cells, as opposed to primary HGC (1953 µM; 0.5 mg/mL). In comparison, caspase-3 and -7 activity was noted in HGC after 4 h treatment with 391 µM (0.1 mg/mL) tyrindoleninone (p < 0.01; Figure 3A), whereas the corresponding granulosa cancer cell line, KGN, had significant increases in caspase-3 and -7 activity at 20 µM (0.005 mg/mL; p < 0.001; Figure 3B). In fact, there was greater capsase-3 and -7 activity detected overall by the KGN cells than in the corresponding primary HGC in the presence of tyrindoleninone. Furthermore, there was a dose dependant decrease in KGN caspase activity as the concentration of tyrindoleninone increased, whereas there was a dose dependent increase for HGC (Figure 3B). Finally, necrosis was only detected in the JAr and OVCAR-3 cells after 4 h exposure to 195 µM (0.05 mg/mL) and 391 µM (0.1 mg/mL) of tyrindoleninone respectively (Figure 2).

In a similar manner to tyrindoleninone, 6-bromoisatin significantly affected the reproductive cancer cell lines at lower concentrations than the HGC (Figure 4). Necrosis, as indicated by LDH release, was only significantly increased after 4 and 24 h exposure to 2232 µM (0.5 mg/mL) of 6-bromoisatin in HGC. In comparison, LDH release significantly increased after exposure to 446 µM (0.1 mg/mL) in the KGN, JAr and OVCAR-3 cells lines (Figure 4). Like tyrindoleninone, caspase-3 and -7 activity was only detected in HGC after 4 h treatment with 0.1 mg/mL 6-bromoisatin (p < 0.001; Figure 3C). In comparison KGN cell apoptotic activity was significantly higher after 4 h treatment with 22 µM (0.005 mg/mL) of 6-bromoisatin (p < 0.001; Figure 3D). When KGN cells were treated with 22–2232 µM (0.005–0.5 mg/mL) of 6-bromoisatin for 24 h, caspase-3 and 7 significantly decreased, in a dose-dependent manner (Figure 3D).

Other structurally similar indole and isatin compounds have also been shown to induce apoptosis by the activation of caspase-3 at low concentrations, in a range of cell lines in vitro [31,45,46]. For example 5,6,7-tribromoisatin activates caspase-3 and -7 at 8 µM (0.003 mg/mL) after 5 h in the Jurkat cell line [31]. Other research suggests that indoles and isatins inhibit cell proliferation and activate apoptosis by binding and inhibiting signalling of extracellular protein kinases (ERKs) [47]. As ERK signalling pathways, such as the ERK/MARK phosphatase pathway, are essential in cell proliferation and survival [48] it only follows that an inhibition of ERK would suppress growth and induce apoptosis in the cells.

To further assess whether the two D. orbita fractions containing the bioactive compounds tyrindoleninone and 6-bromoisatin induced cell death by apoptosis, the TUNEL assay was performed on the HGC and KGN cells. The DNase I treated HGC and KGN cells (positive controls) were 100% TUNEL stained (Figure 5 and Figure 6), whereas the negative controls (absence of terminal deoxynecleotide transferase) were TUNEL negative. After 4 and 24 h incubation, 9 and 12% of untreated (DMSO control) HGC, and 10 and 8% of KGN cells were TUNEL labelled (Figure 5 and Figure 6).

In the presence of 0.05 mg/mL tyrindoleninone after 4 h treatment, 31% of HGC (Figure 5A,C) and 66% of KGN cells were TUNEL positive (Figure 5B,C). After 24 h exposure to 20 µM (0.005 mg/mL) of tyrindoleninone only 16% of HGC were labelled, in comparison to 62% of KGN cells (p < 0.001; Figure 5A,B). DNA fragmentation after 4 h exposure to 10 tyrindoleninone in both HGC and KGN therefore occurred at a lower concentration (0.05 mg/mL) than that which caused LDH release (1953 µM or 0.5 mg/mL; Figure 2).

When HGC and KGN cells were treated with 6-bromoisatin at 223 µM (0.05 mg/mL) for 4 h, 27% of primary HGC were TUNEL positive (Figure 6A,C), whereas 58% of KGN cells were positive (p < 0.001; Figure 6B,C), and after 24 h of treatment with 6-bromoisatin (22 µM or 0.005 mg/mL), 10% of HGC were TUNEL labelled in comparison to a statistically significant increase in TUNEL positive KGN cells of 53% (p < 0.001; Figure 6A,B). In a similar manner to tyrindoleninone, when treated with 6-bromoisatin TUNEL fluorescence indicating DNA fragmentation was generally detected at lower concentrations than required to induce necrosis by LDH release (2232 µM or 0.5 mg/mL; Figure 4).

The consensus is that apoptosis can be activated via two main pathways, the extrinsic (cell surface activation) or intrinsic (mitochondrial activation) pathways [44]. Apoptosis in the KGN and COV434 human granulosa cancer cell lines treated with cisplatin was attributed to the extrinsic receptor pathway [49]. The extrinsic pathway via the Fas-Fas mediated ligand system was also implicated in KGN apoptosis when KGN cells exposed to interferon-γ underwent Fas-induced apoptosis [38]. Furthermore, the Fas-Fas pathway was involved in primary granulosa-luteal cell apoptosis [50]. TRAIL apoptosis via the intrinsic pathway has also been identified in primary granulosa and theca cells of the ovary, regulated by Bcl-212 proteins [51]. Therefore, it could be hypothesized that tyrindoleninone and 6-bromoisatin inhibited cell proliferation and activated apoptosis in the KGN granulosa cancer cell line via the TRAIL-induced or Fas-Fas pathways. Future studies, including investigation of initiator caspase 8 and 9, could help discriminate between the extrinsic and intrinsic pathway for the induction of apoptosis by these brominated indoles.

At higher concentrations of tyrindoleninone and 6-bromoisatin (≥0.05 mg/mL) a dose dependent increase in LDH release was observed in the KGN cells, HGC, JAr and OVCAR-3 cells. Our caspase-3/7 and TUNEL results suggest that tyrindoleninone and 6-bromoisatin activate cell death by apoptosis at low concentrations, but cause secondary necrosis or necrosis at higher concentrations [42,52]. In other words, at higher concentrations the apoptotic pathway was terminated and secondary necrosis may have been triggered. Benkendorff et al. [30] have shown that a crude extract (0.5 mg/mL) from the egg masses of D. orbita induced both apoptosis and necrosis in the T-cell lymphoma, Jurkat cell line, and necrosis alone in the HT29, solid colon cancer cell line after a 4 h exposure [30]. The authors of that study pointed out that cell death by apoptosis in lymphoma cell lines is congruent with published studies, whereas necrosis is more common in solid tumour cell lines [30].

Interestingly, in vivo studies have demonstrated that crude D. orbita extracts successfully induced apoptosis in cells of the distal colon in response to genotoxic damage [32]. High concentrations of the D. orbita extracts (1 mg/mL) induced a significantly greater rate of apoptosis than low concentrations (0.125 mg/mL) in vivo [32]. This is likely to be related to in vivo degradation and metabolism of the active compounds, thus requiring higher concentrations for oral delivery than is necessary to induce an effect in vitro. The induction of apoptosis at lower concentrations in our study compared to previous studies may also be related to testing the purified compounds rather than the crude extract in the apoptosis assays. Different concentrations of these muricid compounds possibly stimulate different cell death pathways in different cell lines and therefore in the KGN reproductive cells, apoptosis is being activated at low concentrations, whereas secondary necrosis or necrosis is activated at high concentrations.

Primary-derived human reproductive granulosa cells (HCG) were isolated from the follicular aspirates donated by women (n = 3) undergoing assisted reproductive technology (ART) at Flinders Medical Centre, Adelaide, South Australia under a protocol approved by the Flinders Clinical Research Ethics Committee (260/067). Granulosa cells were donated by women with normal reproductive physiology who were undergoing ART to treat male infertility. Granulosa cell isolation has been described previously [53]. Briefly, pooled aspirates for each woman were isolated by centrifugation at 107× g for 10 min, followed by two cycles of rinsing and then the cell suspension was re-suspended in Dulbecco’s modified Eagle’s HAMS F12 (DMEM-F12; GIBCO, Invitrogen Corporation) mediumsupplemented with 10% FBS (GIBCO Invitrogen Corporation), penicillin/streptomycin 5000 IU/mL and 5000 μg/mL respectively (Thermo Scientific), insulin (5 μg/mL), transferin apohuman (5 μg/mL), selenium sodium selenite (5 ng/mL) buffered with 1.2 g/L NaHCO₃ (Pfizer). Granulosa cells were separated from the red blood cells using a lymphoprep (Ficoll-Paque TMPlus) column. The purified granulosa cells were rinsed and re-suspended in DMEM/F12 complete medium.

The KGN granulosa cell line [38] was maintained in the same DMEM/F12 complete media as the HCG. The JAr [40] and OVCAR-3 cell lines [39] obtained from the (Global Bioresource Centre™ American Tissue Culture Collection) were maintained in RPMI-1640 medium supplemented with 10 and 20% FBS respectively (GIBCO, Invitrogen Corporation), sodium pyruvate (1 mM), HEPES (10 mM), glucose (4.5 g/L), L-glutamine (2 mM), and penicillin and streptomycin (5000 IU/mL and 5000 μg/mL, respectively; Thermo Scientific). Insulin solution from bovine pancreas (0.01 mg/mL) was also added to OVCAR-3 medium and both JAr and OVCAR-3 medium was buffered with 1.5 g/L NaHCO₃ (Pfizer). Cell lines were maintained in 75 cm² sterile tissue culture flasks (NUNC, Thermo Fisher Scientific) at 37 °C in a humidified atmosphere with 5% CO₂ and sub-cultured every 2–3 days as required. When the cells were 80% confluent they were either passaged or used in experiments. Viable cell numbers were determined using the trypan blue exclusion assay on a haemocytometer [54].

After 4 and 24 h exposure to the D. orbita fractions in opaque 96-well plates, a combined membrane integrity and caspase 3/7 assay were performed on the primary-derived granulosa cells and KGN cells. The CytoTox-ONE Homogeneous Membrane Integrity Assay kit (Promega) was applied according to the manufacturer’s instructions. After the fluorescence was read at 535_EX/590_EM, the Caspase-Glo 3/7^® assay (Promega) was applied to the primary granulosa cells and KGN cells according to the manufacturer’s instructions for1h and the plates were read on full light to capture total luminescence. The LDH-membrane integrity assay alone was performed on the JAr and OVCAR-3 cells using the CytoTox-ONE Homogeneous Membrane Integrity Assay kit (Promega). This assay was repeated on three separate occasions (n = 3).

After 4 and 24 h exposure to D. orbita fractions in clear 96-well plates, media and all dead unattached cells were removed. The remaining adherent cells were rinsed with sterile 1× PBS and the crystal violet assay was performed [55]. The crystal violet assay is a colourmetric assay in which only the nuclei of live cells take up the crystal violet stain (0.5%) [55]. The absorbance in the wells of the crystal violet plates were read on an automatic spectrophotometer at 570 nm, with reference absorbance 630 nm using KC Junior Software. This assay was repeated on three separate occasions (n = 3).

After 4, 24, 48 and 72 h exposure to D. orbita fractions in clear 96-well plates, media and all dead unattached cells were removed. The remaining adherent cells were rinsed with sterile 1× PBS and the MTT assay was performed. The MTT assay measures the reduction of a yellow substrate to formazan, a purple precipitate, by the mitochondrial succinate dehydrogenase enzymes [56]. A solution of 0.5 mg/mL MTT (was added to wells in a final volume of 0.1 mL. The plates were then incubated for 1 h (JAr cells) and 18 h (KGN, OVCAR-3 and GC cells) at 37 °C and 5% CO₂. At the end of the incubation period, 80 μL/well 20% SDS in 0.02 M HCl was added and plate was incubated for a further 1 h at room temperature in the dark. The absorbance was the measured at 570 nm, with reference absorbance 630 nm, using an automatic spectrophotometer using KC Junior software. This assay was repeated on three separate occasions (n = 3).