A Dinuclear Persulfide-Bridged Ruthenium Compound is a Hypoxia-Selective Hydrogen Sulfide (H₂S) Donor.

Abstract

Hydrogen sulfide (H₂S) is a gaseous molecule that has received attention for its role in biological processes and therapeutic potential in diseases such as ischemic reperfusion injury. Despite its clinical relevance, delivery of H₂S to biological systems is hampered by its toxicity at high concentrations. Herein, we report the first metal based H₂S donor that delivers this gas selectively to hypoxic cells. We further show that H₂S release from this compound protects H9c2 rat cardiomyoblasts from an in vitro model of ischemic reperfusion injury. These results validate the utility of redox-activated metal complexes as hypoxia-selective H₂S-releasing agents for use as tools to study the role of this gaseous molecule in complex biological systems.

Graphical Abstract

Targeting hypoxia with H₂S Donors: Exploration of the biological activity of a dinuclear persulfide-bridged ruthenium persulfide complex identifies an exciting class of compounds for delivery of H₂S to hypoxic cells.

Hydrogen sulfide (H₂S) has long been known to be a highly toxic gas with a noxious odor. In 1996, this perception was changed by the discovery that this gas is produced endogenously in mammals and functions as a modulator for neurological activity.^[1] Since this discovery, further work has revealed that H₂S is an important signaling molecule involved in angiogenesis and the prevention of oxidative stress.^[2] Furthermore, H₂S has promising therapeutic potential for the treatment of Alzheimer’s disease, Parkinson’s disease, ischemic reperfusion injury (IRI), stroke, and cancer.^[3] The implementation of H₂S in medicine, however, is limited by its gaseous nature, flammability, and toxicity at high concentrations. As such, significant research efforts have focused on developing easily handled prodrugs for this gaseous molecule.^[4–12] Simple sulfide salts such as Na₂S or NaSH rapidly release H₂S upon dissolution in aqueous solution. Although these complexes are more practical for the delivery of H₂S than its direct administration as a gas, their rapid release profiles do not mimic endogenous H₂S production and often elicit toxic side effects.^[13] To circumvent these challenges, several groups have developed synthetic compounds that release H₂S upon activation by external stimuli such as light,^[14–21] pH,^[22,23] and reactive oxygen species.^[24,25] These compounds allow localized and controllable delivery of H₂S in complex biological systems, making them promising therapeutic candidates.

In this study, we sought to develop H₂S donors that could be selectively activated for therapeutic intervention in conditions such as cancer, IRI, or stroke. Under these pathological conditions, cells and tissue exist in a state of hypoxia, causing them to lose the ability to maintain redox balance and the cellular environment becomes reducing.^[26] In this context, the redox chemistry of Cu, Pt, Co, Fe, Ru, Os and Ir has been used to develop prodrugs that are specifically activated in hypoxic cells to produce reactive anticancer compounds.^[27–31] Our strategy to develop a redox-activated H₂S donor invoked the dinuclear ruthenium persulfide (μ-S₂^2-) core [Ru^IIISSRu^III]. This moiety is labile towards reduction in protic solvents, producing H₂S and the related Ru^II species.^[32–34] In this report, we describe our initial evaluation of a ruthenium persulfide complex as a platform for hypoxia-activated delivery of H₂S in cultured cells and demonstrate the ability of this complex to protect against an in vitro model of ischemic reperfusion injury. These results highlight the value of metal-based H₂S donors as tools for understanding the therapeutic utility of this gasotransmitter.

The compound [(H₂O)Ru(NH₃)₄(μ-S₂)Ru(NH₃)₄(OH₂)]⁴⁺ ([1]⁴⁺, Figure 1) was obtained as the chloride salt ([1]Cl₄) by treatment of trans-[Ru(NH₃)₄(SO₂)Cl]⁺ with amalgamated zinc in 0.1 M HCl followed by purification with cation-exchange chromatography.^[32] This complex was characterized by NMR, UV/vis, and resonance Raman spectroscopies in addition to reverse-phase high-performance liquid chromatography (HPLC), elemental analysis, and x-ray crystallography (Figures 1 and S1–S5, Supporting Information, SI).

Figure 1.

(A) Chemical and (B) X-ray crystal structure of [1](SiF₆)₂. The SiF₆^2- counterions have been omitted for clarity. Thermal ellipsoids are depicted at the 50% probability level. Selected geometric parameters (Å,°): S(1)–S(#1) 2.0186(13), Ru–S(1) 2.1659(7), Ru–O(1) 2.177(2), Ru–S(1)–S(#1) 111.63(5).

X-ray diffraction-quality crystals of [1](SiF₆)₂ were obtained by vapor diffusion of acetone into a solution of [1](SbF₆)₄ in 0.1 M DCl (Figure 1). Relevant details are included in the SI (Tables S2 and S3). The complex crystallizes such that the asymmetric unit consists of one half of the molecule, with a center of inversion located in the middle of the persulfide bond. Overall, the Ru–S and S–S interatomic distances agree well with previously explored Ru persulfide complexes (Table S1, SI). The S–S interatomic distance is the longest reported for a persulfide-bridged diruthenium complex (Table S1, SI) and falls between the value expected for a sulfur-sulfur single (2.03 Å for Me₂S₂)^[35] and double bond (1.887 Å for S₂).^[36–38] This intermediate bond order has been observed in other persulfide-bridged diruthenium complexes and is attributed to highly delocalized π-bonding within the persulfide core.^[34]

To assess the suitability of [1]Cl₄ for hypoxia activation, we analyzed the redox activity of this compound using cyclic voltammetry (CV; Figure 2). In pH 7.4 PBS, the CV of [1]Cl₄ displays an irreversible reduction peak with an onset potential of –716 mV vs. SCE (feature II, Figure 2). The irreversibility of this feature may arise from dissociation of the complex upon reduction. Importantly, the onset of the irreversible reduction for [1]Cl₄ lies within the required range (–0.75 to –0.35 V vs. SCE)^[39,40] for hypoxia selectivity. Another feature, an irreversible oxidation, occurs with an onset potential of 490 mV, a potential that is unlikely to be accessible under biological conditions (feature IV, Figure 2). Following oxidation, a weak feature at 150 mV vs. SCE (features III/V, Figure 2) appears, which possibly corresponds to oxidation products of [1]⁴⁺ obtained after sweeping these higher potentials.

Figure 2.

Top: Cyclic voltammogram of [1]Cl₄ in PBS (pH 7.4, 23 °C). Bottom: Cyclic voltammogram of PBS. Conditions: glassy carbon working electrode, Pt wire counter electrode, Ag/AgCl quasi-reference electrode, and 0.1 V s⁻¹ scan rate.

For [1]Cl₄ to be useful as a biological delivery vehicle for H₂S that does not give rise to acute toxicity, the release of this gas molecule should be gradual rather than instantaneous. We examined the potential of [1]Cl₄ to release H₂S upon treatment with a panel of biologically relevant reducing agents. We found that the presence of Ru(III) in solution interferes with the commonly used methylene blue assay for H₂S detection (Figure S6–S9, SI). As such, the turn-on fluorescent probe SF4 was used to monitor release of H₂S from [1]Cl₄ (Figures 3 and S10–S11, SI). No increase in emission intensity was detected when [1]Cl₄ was incubated at 37 °C, indicating that the complex does not release H₂S under these conditions (Figure 3). This result is consistent with UV/vis spectroscopic studies that show this compound to remain >95% intact after incubation in pH 7.4 PBS at 37 °C for 24 h and >75% intact after 72 h (Figure S12, SI). When [1]Cl₄ is treated with a 10-fold excess of the reducing agents HSO₃^–, cysteine, glutathione (GSH), and ascorbate, the emission of the H₂S sensor SF4 increases gradually over the course of 190 min, confirming that reductive activation of [1]Cl₄ triggers H₂S release. Notably the H₂S yield scales directly with the reducing power of the species and incubation with other biologically relevant species, such as anionic nucleophiles (OH^–, Cl^–, aspartate) and oxidants (GSSG, H₂O₂, NaNO₂, NaHClO, ^tBuOOH) does not trigger H₂S release (Figure S11, SI). These results suggest that this compound will be a useful agent for hypoxia-activated delivery of this gas.

Figure 3.

Top: H₂S-release profile from [1]Cl₄ (20 μM) in pH 7.4 PBS at 37 °C in the presence of biologically relevant reducing agents (200 μM unless indicated). Bottom: Quantification of [H₂S] produced by [1]Cl₄ after incubation in pH 7.4 PBS with relevant biological species (200 μM) [1]Cl₄ for 190 min at 37 °C. Results are reported as mean ± SD (ns = not significant; n = 3–4).

Following these initial studies, we sought to further elucidate the pathway of H₂S release from [1]Cl₄. Two possible mechanisms were considered (Figure S13). The first mechanism (Type I) progresses via reduction of the Ru³⁺ centers. The labile Ru²⁺ would undergo aquation to release H₂S₂, which would then disproportionate to form H₂S and higher order polysulfides. The second mechanism (Type 2) considered involves initial reduction of the persulfide bond to yield Ru^III-SH type complexes, which undergo further reduction and ligand substitution to produce H₂S and a Ru^II species.

To probe these mechanisms, we first treated [1]Cl₄ with 40-fold excess GSH or 10-fold excess ascorbate in the presence of the polysulfide-selective fluorescent probe DSP-3 (Figure S14–S15, SI).^[41] No change in DSP-3 fluorescence after 40 min of treatment was observed, suggesting that polysulfide species are not produced during the reduction of [1]Cl₄ To further confirm these results, we monitored the reaction between [1]Cl₄ and 10-fold excess GSH by UV/vis spectroscopy in the presence of 0.5 M isonicotinamide (isn), a technique previously employed to study the reduction reaction between GSH and Ru^III ammine complexes (Figure S16, SI).^[42] Upon addition of GSH to a solution containing [1]Cl₄ and 0.5 M isonicotinamide (isn), a peak rapidly appears at 427 nm, which is characteristic of trans-[isn(NH₃)₄Ru(SH)]²⁺.^[43] Taken together, these results suggest that decomposition of [1]Cl₄ upon reduction proceeds through a mechanism similar to Type 2 (Figure S13, SI) to selectively release H₂S without initial production of polysulfide species. This reactivity pattern contrasts that of many organic H₂S donors that contain polysulfide bonds produce reactive polysulfide species as intermediate products.^[44–49]

Given the promising H₂S-release profile and selectivity of [1]Cl₄, we investigated the biological activity of this complex. The complex is effectively nontoxic at concentrations up to 200 μM in cervical cancer (HeLa) and rat cardiomyoblast (H9c2) cells (Figure S17, SI). Additionally, [1]Cl₄ is taken up by cells effectively (Figure S18, SI), as determined by graphite furnace atomic absorption spectroscopy. Based on its cell permeability and low toxicity, we next investigated the ability of [1]Cl₄ to selectively release H₂S in hypoxic cells using the cell-trappable, H₂S-responsive fluorescent probe, SF7-AM.^[50] Cells that were only treated with [1]Cl₄ or subjected to hypoxic conditions in the absence of the complex showed no significant increase in fluorescence intensity compared to control cells. In contrast, we observe a significant increase in fluorescence intensity in cells treated with both [1]Cl₄ and hypoxic conditions, indicating that both components are required for intracellular release of H₂S (Figure 4). Additionally, when HeLa cells were treated with a spent solution of the complex (See SI for details, Figure S19–S20, SI), we observed no increase in fluorescence intensity with incubation under normoxic or hypoxic conditions (Figure S20, SI). This result confirms that the fluorescence enhancement observed for cells treated with [1]Cl₄ under hypoxic conditions arises from H₂S produced by the complex.

Figure 4.

(A) Representative images of H₂S-release from [1]Cl₄ in vitro. HeLa cells were loaded with 5 μM SF7-AM for 30 min, washed, and loaded with 0 or 50 μM [1]Cl₄ for 1 h. Cells were then incubated in GBSS under either hypoxic (95:5 N₂/CO₂) or normoxic (95:5 Air/CO₂) conditions for 3 h. (B) Corrected total fluorescence of HeLa cells incubated in the conditions described (see SI for details). Results are reported as mean ± SD (***p<0.001, n = 3–4).

H₂S has therapeutic properties for preventing the damaging effects of IRI in heart disease and stroke.^[51–55] We therefore investigated the ability of [1]Cl₄ to protect H9c2 rat cardiomyoblast cells from this condition using an in vitro model for IRI. hen cells were pretreated with [1]Cl₄ prior to hypoxia, a dose-dependent increase in cell viability relative to the untreated cells was observed, indicating that this compound gives rise to cytoprotective effects (Figure 5). One the major mechanisms of the therapeutic effect of H₂S for the treatment of IRI is the activation of the mitochondrial K_ATP channel,^[56–59] an energy-dependent transporter of mitochondrial K⁺ ions. H₂S will activate this channel, causing the mitochondria to expunge K⁺ ions to decrease the mitochondrial membrane potential (MMP). The decreased MMP will lead to diminished uptake of Ca²⁺ ions, preventing mitochondrial calcium overload, the primary cause of the cytotoxicity of IRI. To confirm that H₂S mediates the protective effects of [1]Cl₄, the H9c2 cells were incubated with the K_ATP channel inhibitor glibenclamide^[60] (10 μM) prior to subjecting them to IRI. In the presence of glibenclamide, [1]Cl₄ fails to protect the cells from death due to IRI (Figure 5). This result indicates that the cytoprotective effects of [1]Cl₄ arise from its ability to release H₂S, which acts directly on the mitochondrial K_ATP channel. Furthermore, treatment with the spent solution described above did not give rise to any of the observed protective effects, confirming that the cytoprotective effects of [1]Cl₄ arise from its ability to produce H₂S in hypoxic cells (Figure S21, SI).

Figure 5.

Protective effects of [1]Cl₄ at various concentrations in H9c2 cells exposed to hypoxia-reoxygenation injury after preincubation with 0 or 10 μM glibenclamide. Results are reported as mean ± SD (ns = not significant; **p<0.01; ***p<0.001; n = 3).

In summary, by applying the “activation by reduction” principle that has been leveraged for the design of metal-based anticancer agents, we have been able to use [1]Cl₄ as the first H₂S donor that is activated selectively by reduction in hypoxic cells. This work demonstrates that Ru persulfide complexes are viable platforms for H₂S delivery. Furthermore, it highlights how transition metal compounds, in general, may serve as viable candidates for releasing H₂S and other reactive-sulfur species, adding to their prior roles as delivery agents for the more well-known gasotransmitters CO and NO.^[61–67]