Potential interactions between metal-based phenanthroline drugs and the unfolded protein response endoplasmic reticulum stress pathway

Abstract

The unfolded protein response has recently been implicated as a mechanism by which 1,10-phenanthroline-containing coordination compounds trigger cell death. We explored the interaction of two such compounds—one containing copper and one containing manganese—with endoplasmic reticulum (ER) stress. Pretreatment with anisomycin significantly enhanced the cytotoxic activity of both metal-based compounds in A2780, but only the copper-based compound in A549 cells. The effects of pretreatment with tunicamycin were dependent on the nature of the metal center in the compounds. In A2780 cells, the cytotoxic action of the copper compound was reduced by tunicamycin only at high concentration. In contrast, in A549 cells the efficacy of the manganese compound cells was reduced at all tested concentrations. Intriguingly, some impact of free 1,10-phenanthroline was also observed in A549 cells. These results are discussed in the context of the emerging evidence that the ER plays a role in the cytotoxic action of 1,10-phenanthroline-based compounds.

Introduction

The cytotoxic and microbicidal effects of 1,10-phenanthroline (1,10-phen) transition metal coordination complexes are well established (reviewed by McCann et al., 2012; Viganor et al., 2017). The mechanistic details of these effects are not fully determined, but multiple routes of action are likely. A consistent feature of these complexes is their capacity to directly generate or promote the formation of reactive oxygen species capable of inducing DNA damage (Anbu et al., 2013; Kellett et al., 2011a; 2011b; Roy et al., 2008). The accumulation of DNA damage and activation of one or more cell death pathways, is a consistent finding in studies of 1,10-phen-containing compounds (Deegan et al., 2007; Kellett et al., 2011b; Slator et al., 2017). In recent years, however, evidence for the involvement of other cellular processes has begun to emerge.

The ability of 1,10-phen to inhibit the proteasome has been reported, specifically via an effect on the regulatory lid subunit protein RPN11, a zinc-dependent metalloprotease (Mansour et al., 2015; Song et al., 2016; 2017). The strong chelative action of 1,10-phen of transition metals likely underlies this effect. 1,10-Phen compounds containing copper have been shown to inhibit the proteasome and induce apoptosis in PC-3 and MDA-MB-231 cancer cells lines, but not in nontumorigenic origin MCF-10A cells (Crowley, 2019; Zhang et al., 2013; 2017) as well as in SKOV-3 and A549 cells (Crowley, 2019). Slator et al. (2017) demonstrated that a 1,10-phen-based compound containing manganese increased mitochondrial superoxide content and induced autophagy in SKOV-3 cells.

There is also some evidence for a role of endoplasmic reticulum (ER)-stress in the action of 1,10-phen-containing compounds. Wu et al. (2014) reported that a dinuclear 1,10-phen-copper complex could induce the expression of several ER-stress proteins in HepG2 cells, including Grp78, XBP-1, and CHOP. Copper(II)–phenanthroline complexes containing imidazolidine-2-thione derivatives were shown to induce apoptosis via the unfolded protein response (UPR) (Moráň et al. (2019). The toxicity was significantly inhibited in the presence of tauroursodeoxycholic acid (TUDCA), which prevents the dissociation of Grp78 from PERK, and thus attenuates the UPR (Yoon et al., 2016). Very recently, a novel copper(II)-phenanthroline complex containing salubrinal (a protein translation inhibitor) was reported to induce cell death via ER-stress in some, but not all, cell types studied (Masuri et al., 2020).

To expand studies of 1,10-phen and ER-stress, we explored the interaction between two 1,10-phen-containing compounds—{[Cu(3,6,9-tdda)(phen)2]·3H₂O·EtOH}n (Cu-Phen) and {[Mn(3,6,9-tdda)(phen)2]·3H₂O·EtOH}n (Mn-Phen) (where phen = 1,10-phenanthroline and 3,6,9-tddaH2 = 3,6,9-trioxaundecanedioic acid)—and either anisomycin and tunicamycin, both established inducers of ER-stress (Majumder et al., 2012; Mawji et al., 2007; Oslowski & Urano, 2011; Tang et al., 2018).

Methods

Materials

The synthesis of {[Cu(3,6,9-tdda)(phen)2].3H₂O.EtOH}n (Cu-Phen) and {[Mn(3,6,9-tdda) (phen)2]·.3H₂O·EtOH}n (Mn-Phen) (where phen = 1,10-phenanthroline and 3,6,9-tddaH2 = 3,6,9-trioxaundecanedioic acid) is described in Gandra et al. (2017). 1,10-Phenanthroline was purchased from Sigma Aldrich (Merk), while anisomycin and tunicamycin were purchased from Cayman chemical. Anisomycin and tunicamycin were dissolved in DMSO, aliquoted, and stored at −20°C until required. 1,10-Phen was dissolved in methanol up to 24 hr before use.

Cell culture conditions

A2780 (ovarian cancer) and A549 (alveolar adenocarcinoma cancer cell lines were cultured in RPMI1640, supplemented with 10% fetal bovine serum and 1% l-glutamine, and incubated in a humidified atmosphere of 5% CO₂ in air at 37°C. Both cell lines were obtained from the internal cell bank at Technological University Dublin (TU Dublin City Campus) and commercially through the European Collection of Authenticated Cell Cultures (ECACC). Cells were plated at 1 × 10⁴ cells/well for A2780 and 5 × 10³ cells/well for A549 cells in black-walled 96-well plates. These cell densities are based on previous optimization trials in our laboratory.

Treatment with ER-stress-inducing agents

A2780 and A549 cells were pretreated with either 100 nM of anisomycin for 24 hr or 5 nM of tunicamycin for 8 hr, followed by treatment with Cu-phen, Mn-phen, or 1, 10-phen for 24 hr. Pretreatment concentrations and incubation times were optimized in previous experiments. To ensure sensitivity to changes in cytotoxic action, concentrations of 1,10-phen-based compounds were chosen to include non- and modestly toxic concentrations, based on previous work in our laboratory which determined the IC₅₀ and cytotoxic range of the compounds (Table 1).

Table 1. IC₅₀ values for 1,10-phenanthroline-based compounds in A2780 and A549 cells

Note. Cells were treated for 24 hr with each compound and assayed using the Alamar blue cell proliferation assay. IC₅₀ values were calculated using GraphPad prism 5.0.

Cell viability

Cell viability was evaluated via the resazurin reduction assay using empirically optimized conditions for each cell type. A 0.44 mM solution of resazurin salt (sigma) in PBS (pH 7.4), equivalent to the commercial Alamar Blue reagent (O’Brien et al., 2000), was prepared in house. Following drug exposure for the specified time, 10 μl of resazurin was added to each well at 10% of final well volume and incubated for 2 hr with A2780 cells and 4 hr with A549 cells. Fluorescence was detected using a MULTISKANGO spectrophotometer (ThermoFischer) at excitation of 530 nm and emission of 590 nm.

Statistical analysis

Data was first expressed as a percentage of nontreated control cells and then normalized to cells treated with anisomycin or tunicamycin only to account for affects either compound had on cell viability. Data were analyzed using Microsoft excel and graphs were generated using GraphPad Prism Version 5.00. Results were compared using the Students t-test with values p < .05 considered significant.

Results

Anisomycin treatment

All results shown below are corrected for the direct impact of anisomycin. Pretreatment of A2780 cells with anisomycin resulted in a modest yet significant increase in the cytotoxic action of Cu-phen, Mn-phen, and 1,10-phen (Figure 1). The apparent loss of the effect at 5 μM of Cu-phen is attributed to the pronounced cytotoxic action at this concentration masking any anisomycin-induced effects. This also prevented further studies at this concentration of Cu-phen. In A549 cells an almost identical effect was observed following 500 nM Cu-phen treatment (Figure 2). In contrast, there was no impact on the efficacy of Mn-phen and only a very minor impact on 1,10-phen. Anisomycin reduced the viability of A2780 cells by 42 ± 2.1% and A549 cells by 30 ± 5.1%.

Figure 1. Effects of Anisomycin (100 nM) on the efficacy of 1,10-phenanthroline-based compounds in A2780 cells. Cells were pretreated with anisomycin as described in methods for 24 hr, followed by treatment with Cu-phen, Mn-phen, and 1,10-phen for a further 24 hr. Results are expressed as mean percent viability compared to cells without cotreatment, ±1 standard deviation of a representative sample of two independent experiments ( **p < .01, and ***p < .001).

Figure 2. Effects of Anisomycin (100 nM) on the efficacy of 1,10-phenanthroline-based compounds in A549 cells. Cells were pretreated with anisomycin as described in methods for 24 hr, followed by treatment with Cu-phen, Mn-phen, and 1,10-phen for a further 24 hr. Results are expressed as mean percent viability compared to cells without cotreatment, ±1 standard deviation of a representative sample of two independent experiments (*p < .05, ns, not significant).

Tunicamycin treatment

All results shown below are corrected for the direct impact of tunicamycin. Pretreatment with tunicamycin resulted in very variable effects, dependent on both cell type, compound type, and concentration. In A2780 cells, treatment with low concentrations of Cu-phen and Mn-phen slightly increased the cytotoxic action of both compounds. Similar potentiation of cytotoxic action was observed at high concentrations of 1,10-phen. In contrast, treatment of cells with a higher concentration of Cu-phen inhibited the action of Cu-Phen and eliminated the effect of Mn-phen (Figure 3). Tunicamycin reduced A2780 cell viability by 35 ± 2.2% and A549 cell viability by 40 ± 2.1%.

Figure 3. Effects of Tunicamycin (5 nM) on the efficacy of 1,10-phenanthroline-based compounds in A2780 cells. Cells were pretreated with Tunicamycin as described in methods for 8 hr, followed by treatment with Cu-phen, Mn-phen, and 1,10-phen for a further 24 hr. Results are expressed as mean percent viability compared to cells without cotreatment, ±1 standard deviation of a representative sample of two independent experiments (*p < .05, **p < .01, and ***p < .001).

In A549 cells, there were marked differences between the effects of pretreatment on the action of Cu-phen and Mn-phen. There was only a very modest impact of tunicamycin on Cu-phen at the highest concentration tested, in contrast to the situation in A2870 cells (Figure 4). Surprisingly, we observed a robust protective effect of both Mn-phen and 1,10-phen in A549 cells at all concentrations tested. Indeed, in these cases, tunicamycin pretreatment appeared to prevent all cytotoxic action at all tested concentrations.

Figure 4. Effects of Tunicamycin (5 nM) on the efficacy of 1,10-phenanthroline-based compounds in A549 cells. Cells were pretreated with Tunicamycin as described in methods for 8 hr, followed by treatment with Cu-phen, Mn-phen, and 1,10-phen for a further 24 hr. Results are expressed as mean percent viability compared to cells without cotreatment, ±1 standard deviation of a representative sample of two independent experiments. (*p < .05, ***p < .001).

Discussion

The potential for metal-phen-based compounds to interact with the UPR pathway as a novel mechanism of action has recently been explored, particularly regarding copper-containing phenanthroline derivatives (Masuri et al., 2020; Moráň et al., 2019; Wu et al., 2014). This work investigated whether the cytotoxic potential of novel copper and manganese-containing phenanthroline compounds could be increased using the UPR-inducing drugs Anisomycin and Tunicamycin. Overall, these data indicate that Cu-Phen, Mn-Phen, and 1,10-phen interact with both UPR-inducing drugs to a modest degree and in a cell type specific manner. While major changes in cytotoxicity were not seen, the modest yet significant enhancement of metal-phen compound efficacy suggests that the UPR may play a role in the mechanism of action of the metal-based compounds, although it may not be the primary mediator of cell death.

Both Mn-Phen and 1,10-phen reacted with Anisomycin and Tunicamycin treatment in cell type specific manners. In A549 cells the efficacy of 1,10-phen is significantly enhanced by Anisomycin with a similar trend observed in A2780 cells indicating that 1,10-phen likely interacts with the UPR pathway albeit modestly. Mn-Phen interaction with the UPR-inducing drugs is highly contextual in relation to the cell type implemented. Mn-Phen did not interact with Anisomycin in A549 cells yet experienced modest increases in cytotoxicity with a high degree of significance in A2780 cells following Anisomycin treatment. Tunicamycin appeared to inhibit both Mn-phen and 1,10-phen activity in A549 cells while again modest yet highly significant increases in cytotoxicity were observed in A2780 cells. Collectively these data indicate that in a similar fashion to 1,10-phen Mn-phen does interact with the UPR pathway to a modest degree, however, the somewhat contradictory results observed between A549 and A2780 cells regarding Tunicamycin pretreatment are somewhat perplexing.

At the current state of knowledge, it is difficult to speculate about the mechanisms for this increase in viability. It may be a consequence that the mechanism of action of Mn-Phen is dependent on one or more glycoproteins, the concentration of which is reduced following Tunicamycin treatment in A549 cells but not A2780 cells. Glycans and glycoproteins are involved in the complex regulation of cell fate either promoting or inhibiting cell death in a context dependent manner (reviewed by Lichtenstein & Rabinovich, 2013; Seyrek et al., 2019). As such alterations to glycosylation status of proteins involved in regulating cell death pathways by Tunicamycin could have unexpected effects regarding cytotoxic insult by metal-phen compounds.

Furthermore, Shenkman et al. (2007) demonstrated that inhibition of both proteasomal and nonproteasomal protein degradation pathways by 1,10-phen was alleviated by the presence of cobalt and manganese, likely owing to metal dependant degradation enzymes. As inhibition of protein degradation would further signal ER-stress mediated cell death, the presence of manganese supplied both by Mn-phen and 1,10-phens chelative properties could reduce the accumulation of misfolded proteins thereby decreasing ER-stress and associated cell death signals. It may therefore be the case that disruption of glycoprotein cell death signaling by tunicamycin and alleviation of protein degradation pathways coincide to significantly reduce the cytotoxic potential of Mn-Phen and 1,10-phen in a cell type specific context. While the exact reasons for these effects are outside of the scope of the current work, further exploring this avenue of investigation in more detail would likely prove fruitful.

Cu-Phen is seen to interact with Anisomycin and Tunicamycin in both cells lines and supports the hypothesis that this compound interacts with the UPR stress pathway, possibly due to a combination of previously reported proteasome inhibition by copper-phen complexes (Zhang et al., 2017). While a higher degree of activity between Cu-Phen and ER stressinducing agents was expected, these findings are supported by previous observations reported within the literature (Masuri et al., 2020; Moráň et al., 2019; Wu et al., 2014), and further, add to the mounting evidence of interactions between ER-stress mediated cell death and Cu-Phen compounds. Using an alternative inhibitory approach, for example, using ER stress alleviators such as TUDCA over ER stress inducers (Anisomycin, Tunicamycin) may demonstrate a higher degree of interaction between compound and pathway.

Interestingly higher concentrations of Cu-Phen appeared to experience an inhibition of cytotoxicity similar to that of Mn-Phen following treatment with Tunicamycin. However, this inhibition is observed in A2780 cells regarding Cu-Phen yet Mn-Phen is inhibited only in A549 cells and is enhanced in A2780 cells following Tunicamycin treatment. Many of the reasons for these phenomena have already been speculated and this specific instance of compound inhibition may be a combination of both cell type specific effects regarding not just Tunicamycin as previously thought but also cytotoxic mechanisms of the compounds in both cell lines.

To fully determine the extent the UPR plays in compound cytotoxicity the implementation of a multipronged approach would likely be required. Expression analysis of UPR-associated proteins (IRE1α, XBP-1, ATF4, and CHOP) (Walter & Ron, 2011) and implementing various inhibitors involved in UPR signaling would likely add supporting evidence to the observations, particularly regarding Cu-Phen. Inhibitors of the autophagic process would also shed light on whether the Mn-Phen compound described shares similarities with the analogous manganese-phen compound already shown to induce autophagic cell death as described by Slator et al. (2017). Sequentially combining both UPR inhibitors with inhibitors of autophagy could also shed light on the degree of any potential crosstalk between the two pathways in relation to signaling cell death following treatment with Mn-phen. A change in cytotoxicity in response to one or both inhibitors would likely determine this.

Taken together these results implicate the potential of the UPR pathway in mediating the effects of novel metal-based phen compounds.

Conclusions

Emerging evidence suggests that 1,10-phen-based drugs may interact with ER stress pathways. Pretreatment with the stress-inducing agents anisomycin or tunicamycin followed by treatment with novel copper and manganese compounds revealed evidence of potential interaction, although cell-specific effects were apparent. However, these data may also highlight a potentially conserved mechanism of action for 1,10-phen-based compounds—interaction with the ER and triggering of ER stress and the UPR.