Molecular target identification, cell-line selectivity, and preclinical evaluation of the anticancer quinolinethiolato-bridged triosmium carbonyl cluster [Os3(CO)10(μ-H)(μ-S)(C9H6N)]

The organometallic cluster [Os3(CO)10(μ-H)(μ-S)(C9H6N)] — a 2-quinolinethiolate-bridged triosmium carbonyl cluster, termed 'Os' in the source study — was reported in 2013 to kill HCT116 and RKO colon carcinoma cells by a mitochondria-targeting, reactive oxygen species (ROS)-dependent apoptotic mechanism (Bax/Bcl-2-independent mitochondrial permeabilization, dissipation of the mitochondrial membrane potential, electron-transport-chain complex II inhibition, ATP depletion), to inhibit HCT116 xenograft growth in nude mice, and to sensitize cells to cisplatin and doxorubicin in vitro, whereas HT29 and SW620 colon cancer cells showed neither ROS induction nor death. The molecular target of its mitochondrial action — a specific protein or lipid versus indirect perturbation — was not identified, the basis of the cell-line-selective response was left unresolved (only p53 status was excluded), and the authors stated that a thorough toxicological analysis, pharmacodynamic/pharmacokinetic assessment, and comparison with platinum-based drugs remain to be performed. Follow-up work through 2026 on this cluster family established the cellular mode of action (ROS generation, G2/M arrest, apoptosis), showed that potency is governed by cellular uptake and reactivity with serum, and demonstrated preferential reaction of the triosmium core with amine-containing biomolecules, but the precise molecular target 'remained elusive' (2025), the colon-cancer cell-line differential is unexplained, and no pharmacokinetic, toxicological, platinum-comparison, or in vivo combination study exists. The problem is open.

Background

Platinum-based drugs such as cisplatin are cornerstones of cancer chemotherapy, but toxicity and acquired resistance motivate the search for anticancer candidates built from other transition metals. Within this effort, metal carbonyl clusters — molecules containing direct metal-metal bonds between three or more metal atoms, each metal bearing multiple carbon-monoxide (CO) ligands — have been explored since the late 2000s as a chemically distinct class of prospective metallodrugs. The trinuclear osmium family at the center of this problem is accessed from the lightly stabilized precursor $[\mathrm{Os}_3(\mathrm{CO})_{10}(\mathrm{NCCH}_3)_2]$, whose two labile acetonitrile ligands are displaced by thiolates to give clusters of the general form $\mathrm{Os}_3(\mathrm{CO})_{10}(\mu\text{-H})(\mu\text{-SR})$, where $\mu$ marks a ligand bridging two metal atoms and R is an aryl or heteroaryl group; the compound studied here carries a 2-quinolinyl group ($\mathrm{R} = \mathrm{C}_9\mathrm{H}_6\mathrm{N}$). Kong and co-workers showed in 2008 that such clusters induce apoptosis — programmed cell death executed by caspase proteases — in several cancer cell lines.

In the source study (Maillet et al., 2013) this quinolinylthiolate cluster, abbreviated 'Os', was profiled in colon carcinoma cell lines. In sensitive HCT116 cells it triggered the intrinsic (mitochondrial) apoptotic pathway: mitochondrial outer-membrane permeabilization (MOMP, the release of pro-death factors such as cytochrome c from the mitochondrial intermembrane space, normally controlled by the pro- and anti-apoptotic Bcl-2 protein family), activation of caspases 9 and 3, and cleavage of PARP (a caspase substrate whose cleavage is a standard apoptosis marker). Mitochondrial morphology was altered, the mitochondrial transmembrane potential ($\Delta\Psi_m$, the proton gradient across the inner membrane that drives ATP synthesis) collapsed, ATP levels fell, the enzymatic activity of electron-transport-chain (ETC) complex II (succinate dehydrogenase, a respiratory-chain component) decreased, and mitochondrial production of reactive oxygen species (ROS, partially reduced oxygen species such as superoxide and hydrogen peroxide that oxidatively damage macromolecules) rose. Death proceeded without involvement of the pro-apoptotic Bcl-2-family protein Bax or the anti-apoptotic protein Bcl-2, suggesting direct mitochondrial insult rather than upstream death-signaling; a transient autophagy response (bulk degradation of cellular components in double-membrane vesicles) accompanied but did not execute the death program. Pharmacological chronology placed ROS upstream: the antioxidant N-acetylcysteine blocked caspase activation, whereas pan-caspase inhibition did not prevent ROS. Cell-line selectivity was striking — HCT116 and RKO cells responded with mitochondrial ROS and died, while HT29 and SW620 cells produced little ROS and resisted short- and long-term cytotoxicity; p53 status (p53 is a tumor-suppressor protein; both resistant lines carry mutant p53, and knocking out p53 in sensitive HCT116 cells did not confer resistance) was excluded as the determinant. In athymic nude mice (immunodeficient mice) bearing HCT116 flank xenografts (tumors grown from transplanted human cells), intraperitoneal Os twice weekly for three weeks (500 μg or 1 mg per injection) delayed tumor growth, with tumors in the 1 mg group reaching 57% of control size and no weight loss; in vitro pretreatment with Os also sensitized HCT116 cells to cisplatin and to doxorubicin (an anthracycline chemotherapeutic).

What the 2013 study did not establish is the molecular identity of the target: whether the cluster binds a specific mitochondrial protein (for example a component of ETC complex II, whose inhibition was observed only as reduced enzymatic activity) or a mitochondrial lipid, or whether the mitochondrial damage is secondary to interactions elsewhere in the cell. Its Discussion explicitly called for a thorough toxicological analysis, pharmacodynamics/pharmacokinetics assessment 'to gauge the true therapeutic potential of Os', and a determination of whether the activity is comparable to platinum-based compounds. Subsequent chemistry-driven work clarified parts of the picture without closing these questions. A 2014 structure–activity relationship (SAR) study across the wider cluster family found that cytotoxicity requires solubility and at least one vacant or labile coordination site, and stated that 'a biological target has not yet been identified'. A 2020 enzymology study showed that the precursor cluster $[\mathrm{Os}_3(\mathrm{CO})_{10}(\mathrm{NCCH}_3)_2]$ — but none of the thiolato-bridged analogues tested — irreversibly inhibits mammalian thioredoxin reductase (TrxR, a selenocysteine-containing enzyme of the cellular antioxidant system; IC$_{50}$ (the half-maximal inhibitory concentration) of 5.3 μM, selective over glutathione reductase), excluding TrxR as the target of the thiolato class to which the 2013 compound belongs. A 2024 SAR study established an inverse correlation between cluster reactivity toward serum components and antiproliferative potency, with electron-donating substituents favoring persistence and cellular entry. A 2025 mode-of-action study confirmed ROS generation, G2/M cell-cycle arrest (arrest at the checkpoint between DNA replication and mitosis), and apoptosis for the family lead in MDA-MB-231 breast-cancer cells; quantification of intracellular osmium by ICP-MS (inductively coupled plasma mass spectrometry) showed that serum reduces cellular uptake and that uptake tracks potency; reactions with small-molecule mimics localized biomolecular reactivity to the triosmium core, preferring amine-containing species; and a density-functional-theory (DFT, a standard quantum-chemical method for computing molecular energies) calculation giving a free energy of $\Delta G^\circ = +87$ kJ/mol for shifting the bridging thiolate from a $\mu$-S (bridging two metal atoms) to a terminal $\kappa$S (bound to a single metal atom through sulfur) bonding mode supports a ligand-substitution mechanism that opens a vacant metal coordination site. Yet the same paper concludes that 'the precise nature of this species remained elusive', it studied breast rather than colon cells, and none of the preclinical assessments requested in 2013 — toxicology, pharmacokinetics/pharmacodynamics, in vivo platinum comparison, long-term efficacy, or in vivo combination validation — has been performed for this compound class anywhere in the literature.

Problem Statement

The problem arising from the source study of Maillet et al. (2013) and left open by all subsequent literature is to establish how the thiolato-bridged triosmium carbonyl cluster $[\mathrm{Os}_3(\mathrm{CO})_{10}(\mu\text{-H})(\mu\text{-S})(\mathrm{C}_9\mathrm{H}_6\mathrm{N})]$ kills cancer cells and whether it merits development as an anticancer drug. It comprises three linked components: components 1 and 2 address the mechanism and the selectivity that the source's results leave unresolved, and component 3 performs the assessments the source explicitly identified as required in order to gauge the compound's true therapeutic potential:

1. Molecular mechanism of mitochondrial targeting. Identify the molecular target(s) the cluster engages and the causal chain producing its mitochondrial phenotype. Specifically: does the cluster bind a specific mitochondrial protein (for example, a component of electron-transport-chain complex II, whose enzymatic inhibition was observed in 2013 but never traced to a binding event), a mitochondrial lipid, or another cellular species (the amine-containing species implicated for the triosmium core in 2025) — or are the mitochondrial structural and functional changes (collapse of the mitochondrial transmembrane potential, complex II inhibition, ATP depletion, ROS rise, Bax/Bcl-2-independent permeabilization) induced indirectly, for example through a proximal lesion elsewhere in the cell? The proposed bonding-mode mechanism (a μ-S to κS shift opening a vacant metal coordination site for a biomolecule, supported by a density-functional-theory (DFT) calculation) remains to be validated against real biomolecules.

2. Determinants of cell-line selectivity. Explain why HCT116 and RKO colon carcinoma cells respond to the cluster with mitochondrial ROS production and apoptosis, whereas HT29 and SW620 cells show neither ROS induction nor sensitivity (p53 status already excluded; candidate determinants such as drug uptake, antioxidant capacity, and electron-transport-chain composition remain untested across this panel).

3. Preclinical assessment of therapeutic potential. Perform the evaluations the source authors explicitly identified as required — thorough toxicological analysis; pharmacodynamic (drug exposure versus effect) and pharmacokinetic (absorption, distribution, metabolism and excretion of the drug) characterization; and a determination of whether the anticancer activity is comparable to that of broadly used platinum-based compounds such as cisplatin, judged on antiproliferative potency in the colon-carcinoma cell panel and on in vivo tumour-growth inhibition — together with two further evaluations that follow from gaps in the source's own results: long-term in vivo efficacy and potential off-target effects beyond the single three-week xenograft experiment, and validation of the cisplatin/doxorubicin synergy in preclinical models with appropriate dosing schedules (the existing synergy data are in vitro only). All of these evaluations are required. The platinum comparison is fixed to anticancer activity — antiproliferative potency in the colon-carcinoma cell panel and in vivo tumour-growth inhibition — while toxicity and the cisplatin/doxorubicin sensitization are each addressed by their own evaluation listed above.

Each of the three components is required and independently refutable: any one of them can be advanced or refuted on its own, but answering component 1 or component 2 alone, or performing only part of component 3, is partial progress and does not resolve the task, which is complete only when all three components have been addressed. Components 1 and 2 are addressed by an evidence-grounded mechanistic and selectivity account, not by new wet-lab data. Because the task environment supplies no compound, cell line, animal or dataset, component 3 is addressed by an evidence-grounded preclinical assessment: an evaluation counts when it is justified to the standard the field would apply — specifying the model, exposure, endpoints and comparison at issue and stating the expected outcome and the evidence it would provide — rather than only when it has been executed.

References

1. Maillet A, Yadav S, Loo YL, Sachaphibulkij K, Pervaiz S. A novel Osmium-based compound targets the mitochondria and triggers ROS-dependent apoptosis in colon carcinoma. Cell Death & Disease 4(6):e653 (2013). DOI: 10.1038/cddis.2013.185. https://doi.org/10.1038/cddis.2013.185
2. Kong KV, Leong WK, Ng SP, Nguyen TH, Lim LHK. Osmium Carbonyl Clusters: A New Class of Apoptosis Inducing Agents. ChemMedChem 3(8):1269-1275 (2008). DOI: 10.1002/cmdc.200800069. https://doi.org/10.1002/cmdc.200800069
3. Lee HZS, Leong WK, Top S, Vessières A. Cytotoxic Triosmium Carbonyl Clusters: A Structure-Activity Relationship Study. ChemMedChem 9(7):1453-1457 (2014). DOI: 10.1002/cmdc.201300394. https://doi.org/10.1002/cmdc.201300394
4. Lee JJY, Leong WK. Towards understanding the mode of action of the cytotoxic triosmium carbonyl cluster Os3(CO)10(NCCH3)2: its reactivity with amino acids and oligopeptides. Journal of Organometallic Chemistry 924:121459 (2020). DOI: 10.1016/j.jorganchem.2020.121459. https://doi.org/10.1016/j.jorganchem.2020.121459
5. Koh WX, Coppo L, Holmgren A, Kong JW, Leong WK. Inhibition of Thioredoxin Reductase by Triosmium Carbonyl Clusters. Chemical Research in Toxicology 33(9):2441-2445 (2020). DOI: 10.1021/acs.chemrestox.0c00214. https://doi.org/10.1021/acs.chemrestox.0c00214
6. Kong JW, Lam Z, Chan KH, Ganguly R, Joey Lee JY, Loo LH, Webster RD, Wong ZX, Leong WK. Group VIII Metal Carbonyl Cluster-Boronic Acid Conjugates: Cytotoxicity and Mode of Action Studies. ACS Omega 6(43):29045-29053 (2021). DOI: 10.1021/acsomega.1c04116. https://doi.org/10.1021/acsomega.1c04116
7. Liang X, Leong WK. SAR Study on Thiolato-Bridged Triosmium Carbonyl Clusters: Higher Reactivity Does Not Equal Higher Antiproliferative Activity. Journal of Medicinal Chemistry 67(23):20980-20985 (2024). DOI: 10.1021/acs.jmedchem.4c01500. https://doi.org/10.1021/acs.jmedchem.4c01500
8. Liang X, Bertrand HC, Delsuc N, Huang B, Leong WK, Lopez-Sanchez A. Decoding the mechanism for cytotoxicity in thiolato-bridged triosmium carbonyl clusters: from serum reactivity to cellular impact. JBIC Journal of Biological Inorganic Chemistry 30(6-8):537-544 (2025). DOI: 10.1007/s00775-025-02127-8. https://doi.org/10.1007/s00775-025-02127-8
9. Liang X, Elumalai K, Zhang L, Tran DL, Leong WK. SAR study on the cytotoxic triosmium carbonyl clusters Os3(μ-H)(μ-EC6H4X)(CO)10: The effects of E and X. Journal of Organometallic Chemistry 1045:123983 (2026). DOI: 10.1016/j.jorganchem.2025.123983. https://doi.org/10.1016/j.jorganchem.2025.123983
10. Konkankit CC, Marker SC, Knopf KM, Wilson JJ. Anticancer activity of complexes of the third row transition metals, rhenium, osmium, and iridium. Dalton Transactions 47(30):9934-9974 (2018). DOI: 10.1039/c8dt01858h. https://doi.org/10.1039/c8dt01858h
11. Tsvetkova D, Ivanova S, Obreshkova D. Potential Activity of Non-Platinum Metal-Based Organic Complexes Against Different Cancer Cell Types. Pharmaceuticals 19(6):925 (2026). DOI: 10.3390/ph19060925. https://doi.org/10.3390/ph19060925