Reactivity-Guided Selective Covalent Targeting of mtDNA over nDNA by an NHC–Rhenium Complex Triggers Bioenergetic Collapse in Cancer Cells

Abstract

Mitochondrial DNA (mtDNA) has emerged as a promising target for the development of next-generation chemotherapeutics, as its damage can disrupt key processes that drive cancer progression and resistance to nuclear DNA (nDNA)-targeting drugs. However, achieving selective and direct targeting of mtDNA remains a significant challenge due to its nucleobase similarity to nDNA and necessitates mitochondria-directed delivery strategies. This study reports the identification of a novel N-heterocyclic carbene (NHC) rhenium (Re) complex 4 that selectively targets and damages mtDNA in cancer cells without the need for mitochondria-directed delivery, while sparing nDNA. This remarkable selectivity arises from the inherent reactivity preference of this complex toward mtDNA, rather than from distribution-driven effects. The in vitro potency of complex 4 outperformed that of nDNA-targeting cisplatin. In-depth mechanistic studies revealed that complex 4 forms covalent adducts with mtDNA, leading to disruption of mitochondrial gene replication and transcription. This triggers various downstream effects, including mitochondrial superoxide generation, dissipation of membrane potential, downregulation of nucleoid-associated proteins and polymerase γ expression, and depletion of OXPHOS proteins and ATP, leading to induction of mitophagy and endoplasmic reticulum stress. It induced necroptosis-mediated cell death and has the ability to overcome cisplatin resistance across multiple cancer cell lines. Furthermore, it significantly reduces the volume of 3D tumor spheroids, demonstrating efficacy in a physiologically relevant tumor model. Overall, this study identifies for the first time the unique mtDNA-targeting capability of [Re­(I)­(bisNHC)­(CO)₃Cl] complexes without the need for mitochondria-targeting vectors.

Introduction

Cancer remains one of the most pressing global health challenges, with approximately 20 million new cases and 10 million deaths in 2022. The high mortality associated with cancer is primarily due to the limited efficacy and frequent failure of frontline therapies, particularly chemotherapy, along with late-stage diagnosis. Platinum (Pt)-based drugs, such as cisplatin, carboplatin, etc., account for nearly 50% of all chemotherapy regimens. These blockbuster drugs primarily exert their anticancer activity by forming covalent adducts with nuclear DNA (nDNA), ultimately inducing apoptosis-mediated cancer cell death. Despite widespread clinical use, the therapeutic efficacy of Pt-based drugs is severely limited by the development of both intrinsic and acquired resistance by cancer cells, which is a primary cause of treatment failure and poor clinical outcomes. Cancer cells use several mechanisms to overcome Pt-induced nDNA damage, including enhanced DNA repair, metabolic and bioenergetic reprogramming, suppression of apoptotic pathways, etc. − Notably, these resistance mechanisms are closely linked to the nDNA-targeting mode of action of Pt drugs, resulting either in poor response at the onset of treatment (intrinsic resistance) or development of resistance during treatment cycles (acquired resistance), ultimately enabling cancer cells to evade the chemotherapeutic effect. The demonstrated clinical utility of Pt drugs, along with their shortcomings due to an nDNA-targeted anticancer mechanism, provides a strong rationale for designing next-generation metal-based chemotherapeutics capable of engaging alternative biomolecular targets or pathways essential for cancer cell survival and progression, yet less susceptible to resistance, thereby offering a promising strategy to improve chemotherapeutic outcomes.

In this context, mitochondrial DNA (mtDNA) has emerged as an attractive molecular target due to its intricate association with cancer incidence and progression along with limited repair capacity. mtDNA is the key regulator of mitochondrial biogenesis and function, thereby playing a central role in essential mitochondrial processes, including energy production, signal transduction, and maintenance of overall cellular homeostasis. − It encodes 37 genes, including 13 essential subunits of oxidative phosphorylation (OXPHOS) complex proteins and RNAs (22 tRNAs and 2 rRNAs) required for protein synthesis. In many cancers, the OXPHOS pathway supports tumor growth through two principal roles: it generates ATP to meet cellular energy demands and serves as a central metabolic node that channels glucose-derived carbon metabolites into macromolecule biosynthesis, thereby integrating both energy-producing and biosynthetic processes. − Both mitochondria and mtDNA cooperatively orchestrate a variety of essential processes required for cancer initiation, progression, metastasis, immune evasion, and development of resistance to therapy. − Therefore, mtDNA-targeted therapeutics may offer several advantages over conventional nDNA-targeted chemotherapy. These include (i) the absence of histone protection and the limited DNA repair capacity in mitochondria, rendering mtDNA more susceptible to damage and enabling the bypass of nDNA repair-mediated resistance mechanisms; (ii) downregulation of OXPHOS, leading to impaired ATP production and depletion of carbon sources required for rapid cell proliferation; and (iii) increased generation of reactive oxygen species (ROS), which further amplifies cellular stress and damage. ,− However, achieving selective and direct targeting of mtDNA remains a significant challenge due to the high nucleobase similarity with nDNA, which represents the primary binding site for functional metal complexes. ,

Consequently, mtDNA-targeted metal-based chemotherapeutics remain largely underexplored, with only a few reported examples to date, all featuring Pt­(II) as the reactive warhead (Figure ). − Lippard et al. demonstrated, for the first time, redirection of the cisplatin warhead to mitochondria from the nucleus using a mitochondria-penetrating peptide (MPP) (Mt1, Figure ). Mt1 exerts its activity through mtDNA damage and was found to overcome cisplatin resistance in A2780 ovarian cancer cells. Wang et al. reported triphenylphosphonium (TPP)-functionalized pyriplatin derivative Mt2. Although Mt2 accumulates 10–11 times higher than cisplatin in mitochondria, the extent of mt-DNA damage induced by Mt2 is only about two times higher than that induced by cisplatin, indicating poor mtDNA damage efficiency. In vivo studies revealed that Mt2 exhibits significant antitumor efficacy in an A549 lung tumor xenograft model, highlighting the clinical competency of mtDNA-targeting agents. The same group recently reported a trinuclear Pt complex, Mt3, that accumulates in mitochondria due to its multiple positive charges and lipophilic character and forms an adduct with mtDNA. Mao et al. reported a heterobimetallic Ru–Pt complex, Mt4, that forms ∼18 times higher adducts with mtDNA compared to nDNA, however, lacks chemotherapeutic activity. Guo et al. reported a naphthalimide-derivatized pyriplatin, Mt5, which exhibited damage to both nDNA and mtDNA via covalent binding as well as intercalation to DNA. In addition, noncovalent targeting of mtDNA by a few metal complexes has also been reported. −

1.

Structures of previously reported covalent mtDNA targeting metal complexes (Mt1–Mt5) and the [Re­(I)­(bisNHC)­(CO)₃Cl] class of complexes (1–4) investigated in this study. The metal center responsible for mtDNA binding is highlighted in yellow.

While these studies are highly encouraging and demonstrate the feasibility of developing mtDNA-targeted metallochemotherapeutics, there are significant scopes for improvement. For instance, the reported complexes exhibit comparable or lower potency than cisplatin (Table S3). Owing to the lack of intrinsic mtDNA-selective reactivity of Pt­(II) warheads, these systems rely on mitochondria-targeted delivery strategies, either conjugation with mitochondria-targeting vectors or the use of multivalent designs, which typically require multistep syntheses (Figure ). Furthermore, with the exception of complex Mt2, there is limited experimental evidence supporting the central premise of covalent interactions between the Pt­(II) center and mtDNA. Considering the shortcomings of reported Pt­(II)-based mtDNA-targeting agents, the identification of an appropriate non-Pt-reactive warhead with inherent and preferential reactivity toward mtDNA could eliminate the need for mitochondria-targeted delivery strategies, thereby simplifying molecular design and may offer the opportunity to enhance antitumor potency. However, the potential of non-Pt reactive warheads for covalent mtDNA-targeted chemotherapy as well as their systematic comparison with established Pt­(II)-based systems remains unexplored.

Herein, we report, for the first time, a novel octahedral [Re­(I)­(bisNHC)­(CO)₃Cl] (Re = rhenium, NHC = N-heterocyclic carbene) complex 4 that selectively and covalently damages mtDNA in cancer cells without the need for mitochondria-directed delivery. Remarkably, complex 4 forms 25-fold higher levels of covalent adducts with mtDNA than with nDNA, resulting in selective mtDNA damage while leaving nDNA largely unaffected (Figure ). This unique mtDNA selectivity, despite complex 4 distributing nearly equally between the mitochondria and nucleus, is attributed to the intrinsic reactivity preference of the [Re­(I)­(bisNHC)­(CO)₃] scaffold toward mtDNA. Re complexes have attracted significant interest for applications in photocatalysis, radiotherapy, and, more recently, chemotherapy. − Although several Re complexes bearing a variety of N̂N bidentate ligands have been reported to exhibit excellent antitumor activity, to the best of our knowledge, none have been shown to preferentially target mtDNA over nDNA, as demonstrated by the bisNHC containing complex 4. − Further, in contrast to the extensively studied [Re­(I)­(N̂N)­(CO)₃Cl] complexes, the antitumor potential, mechanism of action, and molecular targets of the [Re­(I)­(ĈC)­(CO)₃Cl] platform have remained largely unexplored. Complex 4 was identified as the most promising candidate from a synthesized library of four [Re­(I)­(bisNHC)­(CO)₃Cl] complexes bearing varying wingtip substituents on the NHC ligand. The in vitro potency of complex 4 outperformed that of nDNA-targeting cisplatin, as well as all previously reported Pt-based mtDNA-targeting chemotherapeutic agents. In-depth mechanistic studies revealed that covalent adduct formation by 4 with mtDNA led to disruption of mitochondrial gene replication and transcription in cancer cells. This triggers various downstream effects, including mitochondrial superoxide generation, dissipation of membrane potential, downregulation of nucleoid-associated proteins and polymerase γ expression, and depletion of OXPHOS proteins and ATP, leading to induction of mitophagy and endoplasmic reticulum (ER) stress (Figure ). Complex 4 induces necroptosis-mediated cell death and has the ability to overcome cisplatin resistance across multiple cancer cell lines. Furthermore, it significantly reduces the volume of 3D tumor spheroids. Overall, this study identifies, for the first time, the unique and covalent mtDNA-targeting capability of a non–Pt complex of the type [Re­(I)­(bisNHC)­(CO)₃Cl] without the need for mitochondria-targeting vectors.

2.

Schematic illustration of the mtDNA-targeted chemotherapeutic mechanism of action of complex 4.

Results and Discussion

Synthesis and Characterization

The synthetic route to the [Re­(bisNHC)­(CO)₃Cl/H₂O] complexes 1–6 is depicted in Scheme , and details are given in the Supporting Information. The bisNHC carbene precursors L ₁ –L ₄ , synthesized from alkylimidazoles, were treated with Ag₂O, followed by reflux with Re­(CO)₅Cl to obtain the [Re­(bisNHC)­(CO)₃Cl] complexes 1–4 as a white powder in 30–42% yield. The bisNHC precursors and complexes 1–4 were thoroughly characterized using ¹H and ¹³C NMR spectroscopy and MALDI-TOF mass spectrometry (Figures S1–S8 and S10–S16), and purities were confirmed to be >96% through elemental analyses and UPLC (Figures S18–S21). Further, the molecular structure of complex 4 was confirmed by single-crystal X-ray diffraction studies. As shown in Figure , the Re atom in 4 features a slightly distorted octahedral geometry with three carbonyl ligands bound trans to the chloride and the bisNHC unit (unit cell packing and crystallographic parameters are presented in Figure S23 and Tables S1 and S2). Furthermore, two aqua complexes 5 and 6 of the type [Re­(bisNHC)­(CO)₃H₂O] were synthesized from 1 and 4, respectively, as reference complexes for aquation/anation studies and were characterized by ¹H and ¹³C NMR spectroscopy and UPLC (Figures S24–S27, S29, and S30).

1. General Synthetic Route for Complexes 1–6 .

3.

ORTEP plot of compound 4 with 50% thermal ellipsoid probability. Hydrogen atoms are omitted for clarity. Color codes: carbon, gray; nitrogen, blue; rhenium, red; oxygen, brown; chlorine, green. CCDC deposition number 2535857 that contains the supplementary crystallographic data for this structure.

Solution Stability, Aquation, and Anation Chemistry

The stark difference in the Cl^– concentration between the extracellular milieu (∼110 mM) and the intracellular environment (cytoplasm ∼ 23 mM, nucleus ∼ 4 mM) is the key factor governing the extracellular stability and intracellular activation through hydrolysis of M–Cl bond of functional metal complexes including cisplatin. ,, Therefore, we assessed the solution stability and aquation/anation behavior of [Re­(bisNHC)­(CO)₃Cl] complexes using 4 as a representative candidate (Figure A). In an aqueous medium containing 100 mM NaCl, which mimics the extracellular Cl^– concentration, >95% of the complex remained intact over 72 h, with negligible hydrolysis of Re–Cl bond (Figure B). In contrast, removal of NaCl markedly accelerated aquation, and by 24 and 72 h, 57% and 78% of 4 had been converted to aqua species 6 (Figure C). Upon addition of 100 mM NaCl to the mixture, the Re–OH₂ species rapidly reverted to the Re–Cl complex 4, with complete conversion within 1 h (Figure C). In contrast, 4 mM NaCl (mimic of intracellular chloride concentration) was unable to convert aqua complex 6 to chloride complex 4 completely. Instead, an equilibrium was established between the Re–OH₂ (6, ∼70%) and Re–Cl (4, ∼30%) species within 5 h, and this distribution remained unchanged over time (Figure S31). The reversible aquation–anation behavior of this [Re­(bisNHC)­(CO)₃Cl] class of complexes, similar to that observed for the [Re­(N̂N)­(CO)₃Cl] family, suggests that an identical activation mechanism through the aquation of the Re–Cl is likely to operate within cells. Next, the robust stability of complex 4 in DMSO over 72 h was confirmed, as DMSO is the solvent frequently used for preparing stock solutions of water-insoluble compounds for cellular studies (Figure S32). Further, the stability of 4 was evaluated in DMEM, a routinely used cell culture medium. The complex remained intact over 72 h, as no noticeable change was observed in ¹H NMR spectra (Figure D). Taken together, these results demonstrated the excellent stability of this class of complexes in the extracellular medium, suggesting that they are likely to enter cells in their intact form and subsequently undergo activation through Re–Cl bond hydrolysis intracellularly.

4.

(A) Schematic of reversible aquation and anation of complex 4 in the absence and presence of Cl^–. (B) ¹H NMR spectra at different time points of complex 4 in acetone-d ₆/D₂O (9/11, v/v) mixture containing 100 mM NaCl. (C) Reversible aquation and anation of complex 4 studied using ¹H NMR spectroscopy in an Acetone-d6/D2O (9/11, v/v) mixture. (D) Stacked ¹H NMR spectra at t = 0 and 72 h of complex 4 in a DMEM cell culture medium. Signals of complex 4 and complex 6 are marked with # and *, respectively.

Next, we examined the influence of wingtip alkyl substituents on aquation rates by monitoring the aquation kinetics of 1 and 4 in water using UPLC. Complex 1 aquated much faster (100% conversion in 15 min) than complex 4 (∼9 h for complete conversion), which can be attributed to the stronger electron-donating methyl group in 1 compared to the benzyl group in 4 (Figure S33). Consistently, the carbene carbon of 1 appears 1 ppm upfield shifted than that of 4 in the ¹³C NMR spectra (Figures S10 and S12), indicating a higher electron density on the Re center in 1. In addition, the role of steric bulk of the wingtip substituents (Me and Bn) in the aquation chemistry of complexes 1 and 4 could not be excluded, as observed previously for Re­(I)­(N∧N)­(CO)₃ complexes. Nevertheless, these results show that modifying the wingtip substituent on the imidazole ring can be an effective strategy for tuning Re–Cl bond ligand lability.

In Vitro Antitumor Activity, Cellular Uptake, and SAR

The antitumor activities of [Re­(bisNHC)­(CO)₃Cl] complexes 1, 3, and 4 were first screened in nonsmall lung (A549) and human cervical (HeLa) cancer cells. Complex 2 was excluded due to poor solubility and precipitation in the cell culture medium. Cells were treated with increasing concentrations of the test complexes and incubated for 24, 48, or 72 h, after which cell viability was assessed using the MTT assay. The corresponding IC₅₀ values are summarized in Figure A and Table S3, with representative dose–response curves provided in the Supporting Information (Figures S34 and S35). All complexes exhibited potent antitumor activity across all incubation times (24, 48, and 72 h) against both A549 and HeLa cells (Table S3). Comparison of IC₅₀ values at 72 h revealed that complex 1 (3.2 μM in HeLa and 1.6 μM in A549) was the least potent, whereas complex 4 (1 μM in HeLa and 0.5 μM in A549) was the most active among the three complexes (Figure A). Notably, complex 4 showed 3.2 to 18.2-fold higher potency in HeLa and 4.5 to 11.3-fold higher potency in A549 cells (24–72 h incubation) compared to cisplatin (Table S3). Interestingly, A549 lung cancer cells presented slightly higher sensitivity to the [Re­(bisNHC)­(CO)₃Cl] complexes as compared to HeLa cells (Figure A). Comparison of IC₅₀ values with previously reported covalent mtDNA-targeting metal-based chemotherapeutics revealed that complex 4 is 3-, 7-, >83-, 5-fold more potent compared to Mt2, Mt3, Mt4, and Mt5, respectively, in A549 cells (Table S3). ,− IC₅₀ values of Mt1 were not reported in HeLa or A549 cells, thus direct comparison was not possible. However, comparison of the reported IC₅₀ values of Mt1 (7.5 μM) and cisplatin (0.6 μM) in A2780 cells suggested Mt1 has ∼12-fold weaker potency as compared to cisplatin. Taken together, complex 4 represents the most potent mtDNA covalent targeting metal-based chemotherapeutic agent reported to date.

5.

(A) IC₅₀ values of complexes 1, 3, 4, and cisplatin in A549 and HeLa cells in 72 h incubation assay (extended data are presented in Table S3). (B) Log P and whole cell uptake in A549 cells of complexes 1, 3, 4, and cisplatin (10 μM, 6 h exposure). Extended data are presented in Table S4. All data presented as mean ± SD from three independent experiments, each performed in triplicate.

Lipophilicity is the master physicochemical property that profoundly influences cellular uptake and antitumor activity of chemotherapeutics. To assess the influence of wingtip substituent on lipophilicity, the octanol–water partition coefficients (P) of complexes 1, 3, and 4 were determined. Of note, 100 mM NaCl was added to suppress hydrolysis of the Re–Cl bond. As shown in Figure B and Table S4, the measured log P values (−0.2 to +0.5) were markedly higher than those of cisplatin (−2.2), and followed the expected trend: 1 (Me) ≪ 3 (Bu) ≈ 4 (Bn). Next, the cellular uptake of the 1, 3, 4, and cisplatin was determined in A549 cells by quantifying total intracellular Re/Pt content using inductively coupled plasma mass spectrometry (ICP–MS). As illustrated in Figure B, the Re content in the cell treated with 1 (1.3 nmol Re/mg protein) was significantly lower than that treated with 3 (6.3 nmol Re/mg protein), 4 (4.4 nmol Re/mg protein), and cisplatin (6.9 nmol Pt/mg protein).

Understanding structure–activity-relationships (SARs) is key to the rational optimization of any new class of metal-based antitumor agents and can be inferred from correlations among the molecular structure, lipophilicity (log P), cellular uptake, and cytotoxicity. Our results indicate that the presence of Bu or Bn wingtip substituents, respectively, in 3 and 4, leads to ∼0.6 log unit higher lipophilicity compared to the Me-substituted 1. Subsequently, the higher cellular uptakes of 3 and 4 compared to 1 were attributed to their increased lipophilicity (Figure B). Moreover, the enhanced uptake of 3 and 4 appears, at least in part, to account for their higher antitumor potency relative to 1 in A549 cells. Nevertheless, this SAR analysis highlights lipophilicity as one of the critical determinants of cellular internalization and antitumor potency of [Re­(bisNHC)­(CO)₃Cl] complexes, providing a handle for further tuning the antitumor properties of this class of molecules. However, the cellular uptake of cisplatin is known to involve transporter-mediated pathways, and therefore its uptake cannot be directly correlated with a log P value alone.

Intracellular Distribution and DNA Metalation

To obtain mechanistic insights, we selected the most potent complex 4 and first investigated its intracellular organelle distribution in A549 cells. Cells treated with 4 or cisplatin were harvested, and nuclear, cytosolic, and mitochondrial fractions were isolated, followed by quantification of metal content in each fraction using ICP–MS. As shown in Figure A, complex 4 accumulated nearly equally in the nucleus (2.5 ± 0.5 nmol/mg protein) and mitochondria (1.7 ± 0.2 nmol/mg protein). In stark contrast, cisplatin accumulated mainly in the nucleus (2.5 ± 0.8 nmol/mg protein), with substantially lower mitochondrial accumulation (0.5 ± 0.2 nmol/mg protein). Notably, while nuclear accumulation of 4 and cisplatin was comparable, absolute mitochondrial accumulation of 4 was ∼3-fold higher than that of cisplatin.

6.

(A) Intracellular distribution of complex 4 and cisplatin in A549 cells (10 μM, 24 h exposure). (B) Comparison of LC–MS chromatograms of a mixture of complex 6 and 2.5 equiv GMP at different time points and separately synthesized authentic Re­(I)-GMP adduct. (C) Stacked ¹H NMR spectra of free GMP, complex 6, and authentic Re­(I)-GMP adduct showing the shift of H-8 proton upon complexation (Figure S37A for full spectra). (D) ESI-MS spectrum corresponding to the new peak at 14.5 min when 6 is treated with GMP incubated for 144 h at 37 °C. (E) Metalation of nDNA and mtDNA of A549 cells by complex 4 and cisplatin (10 μM, 24 h exposure). Data represent the mean ± SD of three replicates. Asterisks denote differences that are statistically significant (***p < 0.001) and n.s. denotes statistically nonsignificant.

Given that complex 4 undergoes activation through aquation, similar to the clinical DNA-targeting drug cisplatin, we investigated whether the aqua adduct of 4, namely complex 6, can form covalent adducts with DNA bases. A few previous reports showed the GMP binding ability of [Re­(N∧N)­(CO)₃(OH₂)] class of complexes. ,− To investigate this, a mixture of complex 6 and 2.5 equiv guanosine 5′-monophosphate (GMP, a small molecule model for the DNA base) was incubated at 37 °C and the progress of the reaction was followed using LC–MS. As shown in Figure B (Figure S36 for extended data), the appearance of a new peak at retention time ≈14.5 min with a concurrent disappearance of complex 6 was observed with increasing incubation time. Complex 6 was completely consumed when incubated for 144 h. The new peak was confirmed to be the Re­(I)-GMP adduct by comparing the retention time in the LC–MS chromatogram with a separately synthesized authentic Re­(I)-GMP adduct (synthesis and characterization presented in the Supporting Information). Consistently, mass spectrometry showed an adduct at m/z = 962.16 corresponding to [Re­(I)-GMP]⁺ species (Figure D). The stability of the Re­(I)-GMP adduct in LC–MS suggested the interaction of complex 6 with GMP is covalent in nature. Further, ¹H NMR analysis revealed a dramatic upfield shift of the H8 GMP proton upon binding to the Re­(I) center (Figure C), which is in line with previous literature reporting a similar upfield shift of H2 imidazole proton upon complexation with Re­(CO)₃ core. Further, HSQC 2D-NMR data supports this observation (Figure S28). No significant change in ³¹P NMR was observed when GMP was treated with complex 6 (Figure S37B). This suggested GMP binds to the Re­(I) center of complex 6 through the N7 position not through the phosphate group. To further support the covalent interaction of complex 6 with DNA, ct-DNA was incubated with either complex 6 or cisplatin for 24 h. Following incubation, the ct-DNA was precipitated, and the metal content remaining in the supernatant was quantified by ICP–MS. Similar to the well-established covalent DNA binder cisplatin, more than 90% of complex 6 was found to be bound to the DNA fraction (Figure S38). These results provide further evidence that complex 6 interacts covalently with DNA.

The potential of activated complex 4 to directly engage DNA as a biomolecular target, together with its substantial accumulation in both the nucleus and mitochondria, prompted us to investigate its ability to bind nDNA and mtDNA in cells. A549 cells were treated with 10 μM complex 4 or cisplatin (as a control) for 24 h, mtDNA and nDNA were isolated, and repeatedly washed to remove unbound complexes. The metal contents in each DNA fraction were measured using ICP–MS, and results are presented in Figure E. Cisplatin was found to form adducts with nDNA (0.63 ± 0.2 nmol Pt/mg DNA) and mtDNA (0.62 ± 0.4 nmol Pt/mg DNA) with comparable efficiency, despite cisplatin accumulation being 5-fold higher in the nucleus compared to mitochondria. This observation can be explained by the greater susceptibility of mtDNA to DNA-damaging agents, which arises in part from the absence of histone protection compared to nDNA. In sharp contrast, complex 4 exhibited striking selectivity toward mtDNA, forming adducts almost exclusively with mtDNA (18.6 ± 1.5 nmol Re/mg DNA), accounting for ∼25-fold higher mtDNA adduct formation compared to nDNA (0.75 ± 0.15 nmol Re/mg DNA). Notably, this pronounced mtDNA targeting occurs despite slightly lower mitochondrial accumulation of complex 4 relative to the nucleus (Figure A), highlighting its preferential reactivity for mtDNA rather than preferential mitochondrial accumulation driven effects. This can be rationalized by considering fundamental topological, organizational, and repair differences between mtDNA and nDNA. Unlike nDNA, which is linear and tightly packaged into chromatin by histone proteins, mtDNA is circular and relatively exposed due to the lack of histone protection, rendering it more accessible to DNA-reactive agents. In addition, metal-induced lesions in nDNA can be repaired through nucleotide excision repair (NER) pathways, whereas equivalent repair mechanisms are absent for mtDNA. To gain preliminary insights into the influence of DNA topology on the interaction of complex 4 with DNA, we performed gel electrophoresis studies using circular pUC19 DNA and linear ct-DNA (Figure S39). Gel electrophoresis has previously been employed to investigate various aspects of DNA binding by functional Pt-based agents. , Our results suggested that cisplatin has minimal effect on both pUC19 and ct-DNA, causing only a slight increase in the electrophoretic mobility of the open circular form (OC) band of pUC19 DNA at the highest concentration tested (500 μM). In contrast, complex 4 displayed a striking preference for pUC19 circular DNA. At 100 μM, complex 4 caused the complete disappearance of the dominant supercoiled (SCI) band and a substantial reduction in the OC band, accompanied by the appearance of a faint, faster-migrating band. Upon increasing the concentration to 500 μM, all DNA bands disappeared from the gel, with no detectable DNA retained in the loading well (Figure S39A). These observations suggest extensive damage to the circular DNA, including fragmentation and disruption of the native DNA structure, which likely impairs ethidium bromide intercalation and consequently reduces fluorescence staining. In sharp contrast, treatment of linear ct-DNA with either 100 or 500 μM complex 4 produced no significant changes in electrophoretic mobility or band intensity (Figure S39B). Collectively, these data imply that complex 4 has an inherent reactivity preference toward circular DNA, resembling the topology of mtDNA, which could be one potential factor contributing, at least in part, to its selective targeting of mtDNA. Importantly, this selectivity is expected to be enhanced in the cellular environment due to relatively exposed nucleobases of mtDNA toward reactive agents and lack of NER. ,

mtDNA Damage, Replication, and Transcription Inhibition

Given its preferential binding to mtDNA, the ability of complex 4 to induce mtDNA damage was evaluated by quantifying lesions using the real-time quantitative polymerase chain reaction (RT-qPCR). This method exploits the inhibitory effect of DNA damage on polymerase-based amplification, either by hindering primer annealing or by blocking polymerase-mediated synthesis of the complementary DNA strand. We chose two primers belonging to two different regions of the mtDNA, R1 and R2, housing the genes encoding ND4 and ND5 subunits (primer R1) and COX2, COX3, ATPase6, and ATPase8 subunits (primer R2). To compare the levels of mtDNA lesion in each region of the mitochondrial genome, two mtDNA fragments of different lengths (long fragments comprising 1000 bp each and small fragments comprising 54 and 55 bp, respectively), both located in the same mtDNA region were utilized. The adjacent shorter DNA fragment serves as an internal normalization control, adjusting for any influences that affect both fragments of the same mitochondrial genomic region equally. Amplified shorter DNA fragments represent intact mtDNA, as the probability of lesion formation within small nucleotide sequences remains low, data analysis was done based on the measurement of the crossing point (Cp) of each qPCR cycle. Isolated total DNA from the untreated sample was taken as a reference. For each mtDNA region, the difference in crossing point (Cp) values between treated and untreated samples was determined separately for the longer and shorter strands, giving the Δ_long and Δ_short, respectively. mtDNA lesion frequency (MLF) was then calculated according to the 2^–ΔΔCp method in correlation to the fragment size and normalized to a standard DNA length of 10,000 base pairs. As shown in Figure A, a significant increase in MLF was observed upon treatment with complex 4, in comparison to the untreated cells. Furthermore, a dose-dependent increase in MLF was observed in the case of R1 primer, upon treatment with 4. These observations suggested that 4 is a potent mtDNA-damaging agent.

7.

(A) Quantification of mtDNA lesions in A549 cells incubated with complex 4 (2.5 μM and 5 μM) for 24 h using qPCR. (B) Western blot analysis of mtTFA and POLG in A549 cells untreated or treated with 4 for 24 h. Uncropped blots are presented in Figure S40A. (C) Fluorescence images of A549 cells untreated or treated with 4 for 4 h and costained with SYBR gold (1st column) and MitoTracker (2nd column). A white arrow indicates nDNA and yellow arrow indicates mtDNA. (D) Dose–response plots of 4 and cisplatin in A549^WT (with intact mtDNA) and A549ρ⁰ (with depleted mt-DNA) cells. (E) Comparison of IC₅₀ values of 4 and cisplatin in A549^WT and A549ρ⁰ cells (24 h assay). (F) Illustration of the mitochondrial genome and qPCR-amplified fragments are presented in boxes. (G) Mitochondrial gene expression profile of selected OXPHOS genes. (H) Western blot analysis of nDNA damage marker proteins γ-H2AX, PARP, and cleaved PARP in A549 cells treated with complex 4 or cisplatin for 24 h. Images of uncropped full blots are presented in Figure S40B. Data represent the mean ± SD of three replicates. The asterisks and n.s. denote differences that are statistically significant (*p < 0.05, **p < 0.01) and nonsignificant.

These nucleoids consist of mtDNA–protein assemblies primarily organized by mitochondrial transcription factor A (mtTFA), which plays central roles in mtDNA packing, replication, and transcription. ,, In addition to mtTFA, DNA polymerase gamma (POLG) serves as an essential enzyme responsible for mtDNA replication and its structural integrity. The coordinated regulation and functional interplay between mtTFA and POLG are critical for preserving mtDNA homeostasis. Consequently, alterations in mtTFA and POLG expression are widely recognized as indicators of mitochondrial DNA damage and dysregulation. On this basis, we examined the expression levels of mtTFA and POLG in A549 cells following treatment with complex 4 using Western blot analysis. As shown in Figure B, the treatment of A549 cells with complex 4 led to a dose-dependent reduction in both mtTFA and POLG levels. Collectively, these results demonstrate that complex 4 not only induces mtDNA lesions but also perturbs the expression of key proteins governing the mtDNA integrity, replication, and transcription.

Because mtDNA damage and downregulation of mtTFA and POLG are often accompanied by depletion of mtDNA content due to replication arrest, we next examined whether complex 4 induces mtDNA loss in A549 cells. , Following a previously reported protocol, we employed SYBR Gold probe for labeling mtDNA, in addition to MitoTracker Red for locating mitochondria in cells. , At low concentrations, SYBR Gold preferentially stains mtDNA in live cells, with minimal nuclear accumulation. Accordingly, untreated cells showed strong colocalization of cytoplasmic SYBR Gold green fluorescence with MitoTracker Red fluorescence (Figure C), along with a weaker SYBR Gold signal from the nucleus. In contrast, cells treated with complex 4 exhibited a pronounced reduction in a mitochondrial SYBR Gold signal, evidenced by decreased cytoplasmic green fluorescence and poor overlap with MitoTracker Red, accompanied by increased nuclear SYBR Gold staining (Figure C). These results indicated that complex 4 caused mtDNA damage-mediated reduction of genomic content in mitochondria.

To evaluate the contribution of mtDNA to the antitumor activity of complex 4, A549 cells were depleted of mtDNA by ethidium bromide treatment, as confirmed by SYBR Gold staining (Figure S41). mtDNA depletion reduced the cytotoxicity of complex 4 by approximately 2.9-fold, while having little effect on cisplatin activity (Figure D,E). These results identify mtDNA as an important cellular target of complex 4.

Damage to mtDNA is expected to compromise OXPHOS protein expression, leading to a disruption of oxidative respiration and mitochondrial function. To investigate this, we selected one subunit belonging to each complex (ND4, complex I; CyB, complex III, COX1, complex IV; and ATPase 6, ATP synthase) and evaluated their mRNA expression levels using the reverse transcriptase-quantitative polymerase chain reaction (RT-qPCR) method (Figure F). As shown in Figure G, significant downregulation of the mRNA of CyB at IC_50/24h and all four proteins at 2 × IC_50/24h concentration was observed. These observations highlight the irreversible damage inflicted by complex 4 on mtDNA and the consequent suppression of the gene expression of mitochondrially encoded OXPHOS subunits via transcriptional inhibition, ultimately disrupting mitochondrial bioenergetics.

As shown in Figure E, although the extent of nDNA metalation induced by complex 4 is significantly lower than its mtDNA metalation, the level of nDNA metalation remains comparable to that caused by cisplatin. Therefore, we investigated whether this also triggers an nDNA damage response. nDNA damage caused by cisplatin through Pt-nDNA covalent adduct formation is known to trigger the upregulation of specific marker proteins such as γ-H2AX, PARP, and cleaved PARP. Therefore, we performed Western blot analysis to examine the effect of complex 4 treatment on the level of these markers in A549 cells. As presented in Figure H, cisplatin treatment significantly upregulated PARP and cleaved PARP, along with a slight increase in γ-H2AX at 10 μM, which corresponds to 0.36 fold of its IC_50/24h. In contrast, complex 4 did not produce any alteration of these proteins in cells treated up to a conc. of 2-fold IC_50/24h (5 μM), with noticeable upregulation of γ-H2AX and PARP being observed only at 4-fold IC_50/24h (10 μM) dose, without noticeable changes in the cleaved PARP level. These results suggest that a substantially higher dose of complex 4 is required for activation of nDNA damage response proteins as compared to cisplatin, implying nDNA metalation is unlikely to be the primary contributor to the antitumor activity of complex 4.

Oxidative Stress and Mitochondrial Dysfunction

mtDNA damage and subsequent impairment of OXPHOS protein expression are known to trigger mitochondrial superoxide (O ^₂•−) generation, which further amplifies oxidative damage to mtDNA and other mitochondrial components, establishing a self-amplifying vicious cycle. ,, Confocal imaging of complex 4-treated (4 h exposure) A549 cells stained with a mitochondrial O₂ ^{• −} probe MitoSox-red revealed a substantial increase in mitochondrial superoxide, which was further quantified using flow cytometry, confirming elevated mitochondrial oxidative stress (Figures A,B and S42 for dot plots). Consistently, complex 4 treatments dose dependently elevated the whole-cell ROS levels, as revealed by DCFDA imaging, which can be quenched by N-acetyl cysteine (Figure S43).

8.

(A) Confocal microscopy images of untreated and complex 4 treated (for 4 h) A549 cells stained with mitochondrial superoxide probe MitoSOX Red. (B) Quantification of relative mitochondrial superoxide levels using MitoSOX Green in untreated and 4 treated (4 h) cells measured using flow cytometry. Dot plots are presented in Figure S42. (C) Quantification of ΔΨm, as a ratio of JC-1 red/green, in cells left untreated or treated with complex 4 for 4 h or positive control CCCP for 1 h and analyzed using flow cytometry. Dot plots are presented in Figure S44. (D) Quantification of ATP in A549 cells treated with increasing concentrations of complex 4 for 4 and 24 h. Data represent the mean ± SD of three replicates. The asterisks and n.s. denote differences that are statistically significant (*p < 0.05, **p < 0.01, ***p < 0.001) and nonsignificant, respectively.

Superoxide generation is tightly associated with mitochondrial membrane depolarization. Hence, we measured the mitochondrial membrane potential (ΔΨm) in untreated and 4-treated A549 cells utilizing a JC-1 (5,5,6,6′-tetrachloro-1,1′,3,3′-tetraethyl-imidacarbocyanine iodide) probe with flow cytometry. Mitochondria with intact ΔΨm emit in red from JC-1 aggregates, while damaged mitochondria with depleted ΔΨm emit in green from JC-1 monomers. The results depicted in Figure C (Figure S44 for dot plot) suggested a significant and dose-dependent reduction in ΔΨm caused by 4, akin to the positive control carbonyl cyanide m-chlorophenyl hydrazone (CCCP).

Depletion of the mitochondrial membrane potential (ΔΨm) is closely associated with a decline in intracellular ATP levels. Notably, the provision of glycolytically generated ATP has been shown to preserve ΔΨm under conditions of respiratory inhibition, indicating that ATP consumption by ATP synthase and the adenine nucleotide translocator (ANT) is required to maintain ΔΨm when respiration is compromised. Conversely, the loss of the ΔΨm, reflecting dissipation of the electrical gradient across the inner mitochondrial membrane, directly compromises ATP generation, as this gradient provides the driving force for ATP synthase. Prolonged or extensive depolarization therefore results in a shutdown of mitochondrial ATP production. Based on this interdependent ΔΨm and ATP homeostasis, intracellular ATP levels were measured at early and late time points (4 and 24 h) in untreated and complex 4-treated A549 cells. The results indicate a significant ATP depletion in cells treated with ≥5 μM of complex 4 after 4 h exposure, and across all tested concentrations following 24 h of treatment (Figure D). Taken together, mtDNA damage induced by complex 4 leads to elevated oxidative stress, which in turn triggers mitochondrial dysfunction and impairs ATP-dependent mitochondrial bioenergetics.

Induction of Mitophagy and ER Stress

Mitophagy is a specialized form of selective autophagy in which dysfunctional mitochondria are specifically recognized and engulfed by autophagosomes and subsequently delivered to lysosomes for degradation. Mitophagy can serve either a protective or a cell death-promoting role, depending on the biological context. For example, cisplatin-induced mitophagy has been reported to protect against cisplatin-associated nephrotoxicity and ototoxicity , Conversely, excessive or sustained mitophagy can lead to severe mitochondrial damage, bioenergetic collapse, and ultimately cell death. Notably, our confocal imaging experiments using cells with MTS-DsRed-labeled mitochondria revealed treatment with complex 4 caused progressive cytosolic vacuolization, which was confirmed to originate from mitochondria (Figure A). This finding aligns with an earlier literature study demonstrating a correlation between mitochondrial damage and vacuole formation. Mitophagy is associated with mitochondrial fission, a strategy to segregate dysfunctional mitochondria from the network of healthy and functional mitochondria. Therefore, inhibition of mitochondrial fission could rescue cells from undergoing mitophagy and vacuolization. In line with our expectation, treatment of cells with the nontoxic dose of mitochondrial fission inhibitor m-divi abolished vacuolization of cells by complex 4 (Figure B). Furthermore, whole cell ROS generated by complex 4 was also significantly depleted by m-divi, suggesting mitophagy to be one of the primary causes of vacuolation and ROS generation (Figure B,C). During mitophagy, damaged mitochondria are selectively sequestered into autophagosomes, which subsequently fuse with lysosomes for degradation. To confirm this, A549 cells were transfected with the mitochondrial marker Su9-EGFP-PCDNA3.1 and stained with LysoTracker Red for lysosomal visualization. In untreated cells, mitochondria exhibited an elongated filamentous network with minimal overlap with lysosomes. In contrast, complex 4-treated cells (10 μM, 4 h) displayed severely dilated mitochondrial structures that accumulated within enlarged lysosomal vesicles (Figure D), suggesting the formation of mito-lysosomal structures, a hallmark of late-stage mitophagy. These findings are consistent with previous reports demonstrating that damaged mitochondria are delivered to lysosomes for degradation during mitophagic flux. Further, since mitophagy represents selective autophagic clearance of damaged mitochondria, the dose-dependent conversion of LC3I to LC3II provided further support for induction of mitophagy by complex 4 (Figure E).

9.

(A) Confocal microscopy images of untreated and complex 4 treated A549 cells with RFP-labeled mitochondria (exposure 4 h). Cytosolic vacuoles are pointed with white arrows. (B) Bright field and DCFDA-stained images of A549 cells either left untreated or treated with either complex 4 only or complex 4 (10 μM) and m-divi (50 μM). Vacuoles are shown with white arrows. (C) Quantification of relative ROS level from 9B. (D) Confocal imaging of untreated and complex 4 treated A549 cells expressing the mitochondrial marker Su9-EGFP (green) and stained with LysoTracker Red (red). White arrows indicate dilated mitochondria accumulated within enlarged lysosomal vesicles (merged image). (E) Western blot analysis LC3I/II in A549 cells treated with complex 4 for 24 h. Celastrol, a known inducer of ER stress, is used as a positive control. (F) Western blot analysis of marker proteins of ER stress in A549 cells treated with complex 4 for 24 h. The uncropped full images of all blots are presented in Figure S45. (G) Effect of m-divi and cycloheximide on the viability of A549 cells treated with complex 4 (2.5 μM, 24 h). Data represent the mean ± SD of three replicates. The asterisks denote differences that are statistically significant (***p < 0.001).

ER and mitochondria form close contact through cholesterol-rich microdomains known as mitochondria-associated ER membranes (MAMs), and both organelles need to interact for them to function properly. Therefore, mitochondrial damage and dysfunction caused by complex 4 may interfere with the function of the ER, thus amplifying ER stress. Cells treated with complex 4 had an increase in polyubiquitinated proteins in a dose-dependent manner, suggesting accumulation of unfolded proteins (Figure S45A). IRE1α, ATF6, and PERK are the three primary ER transmembrane proteins that initiate signaling in response to ER stress. IRE1α, ATF6, and phosphorylated eIF2α (a downstream protein of the PERK pathway) were all markedly increased upon treatment with complex 4, confirming ER stress induction by complex 4 (Figure F).

To delineate the relative contributions of mitophagy and ER stress to the antitumor activity of complex 4, cell viability was assessed in the absence and presence of the mitophagy inhibitor m-divi or the ER stress inhibitor cycloheximide, and the results are presented in Figure G. Treatment with complex 4 alone in the absence of inhibitors reduced the cell viability to approximately 40%. However, the presence of a nontoxic concentration of either m-divi or cycloheximide completely rescued the cells from viability loss caused by complex 4. These results indicated that both mitophagy and ER stress play critical roles in mediating the antitumor activity of complex 4.

Mode of Cell Death

We then focused on characterizing the mode of cell death induced by complex 4. Cell death pathways involving apoptosis, necroptosis (or necrosis), and autophagy are typically triggered by oxidative damage, mitochondrial dysfunction, and ER stress. , First, we conducted flow cytometry analysis using dual Annexin-V-APC (AV) and propidium iodide (PI) staining in A549 cells, which were either untreated or treated with increasing concentrations of complex 4. Cisplatin and H₂O₂ were included as controls for gating purposes. While PI stains the nucleus of necrotic cells with compromised plasma membranes, AV binds to apoptotic cells with phosphatidylserine flipped outside the plasma membrane. As illustrated in Figure A, similar to the positive control H₂O₂, treatment with complex 4 led to a dose-dependent increase in the AV–/PI+ necrotic population (4.9% in untreated vs 7.6% and 11.1% in 2.5 μM and 5 μM complex 4 treated cells, respectively). No significant increase in the population of AV+/PI– early apoptotic and AV+/PI+ late apoptotic populations was observed. This data suggests that necrosis (or necroptosis), not apoptosis, is one of the main mechanisms of cell death induced by complex 4. Necrosis has traditionally been considered accidental and unprogrammed cell death without the involvement of molecular regulators. However, recent evidence strongly supports the idea that necrosis can occur in a regulated fashion mediated through RIP-1 and/or RIP-3 kinases, termed necroptosis. To investigate whether complex 4 induces necrosis or necroptosis, we performed Western blot analysis to assess the levels of RIP-1 and RIP-3 proteins in A549 cells treated with complex 4. As depicted in Figure B, complex 4 treatment resulted in a significant upregulation of both RIP-1 and RIP-3, similar to the known necroptosis inducer shikonin. Necroptosis can proceed through two separate and distinct routes, independent of each other: tumor necrosis factor mediated and via the poly­(ADP-ribose)­polymerase (PARP) pathway. To understand whether the necroptosis induced by complex 4 is PARP-dependent or independent and to reaffirm the absence of apoptosis, we blotted PARP and cleaved PARP (pro-apoptotic protein). As previously shown in Figure H, PARP was upregulated only at higher concentrations of 4 (>5 μM), but there was no upregulation of cleaved PARP even post-treatment with significantly higher concentrations of 4 (10 μM). These results indicate that treatment with up to 5 μM of 4 triggers PARP-independent necroptosis; however, treatment with higher concentrations of 4 (10 μM) shifts the necroptosis to a PARP-dependent pathway. Moreover, treatment with various cell death inhibitors revealed that the RIP1 inhibitor necrostatin-1 and the autophagy inhibitor 3-methyladenine significantly reduced the cytotoxic potency of complex 4, whereas the pan-caspase inhibitor Z-VAD-FMK and the ferroptosis inhibitor ferrostatin-1 had negligible effects. These findings support necroptosis as the primary mode of cell death induced by complex 4, accompanied by autophagy induction associated with mitophagy (Figure S46).

10.

(A) Flow cytometry analysis of apoptotic and necrotic populations of A549 cells left untreated or treated with complex 4 (2.5 μM-10 μM, 24 h), cisplatin (10 μM, 24 h), and H₂O₂ (2 mM, 2 h). (B) Western blot analysis of necroptosis marker protein in A549 cells treated with complex 4, celastrol, and shikonin (positive control) for 24 h. The uncropped blots are presented in Figure S47.

Platinum Cross Resistance Profile and Efficacy in 3D Tumor Spheroids

Cisplatin and carboplatin remain the mainstay treatments for various cancers, including lung cancer, in the clinic. However, their clinical effectiveness is increasingly compromised by both intrinsic and acquired resistance, which contributes to treatment failure. Since complex 4 exerts its activity through mtDNA damage and induces cell death partly via necroptosis, we anticipated that it would be expected to circumvent cisplatin resistance in the lung as well as other cancers. To investigate this, the IC_50/72h values of 4 in cisplatin-resistant lung A549 Cis and prostate DU145 Cis cancer cells were determined and compared to respective cisplatin-sensitive pairs, A549 and DU145, respectively (Figure S48 for dose–response plots). The resistance factor (RF) was calculated as the ratio of IC₅₀ values in resistant versus sensitive cell lines and is presented in Table . In the A549/A549 Cis pair, complex 4 presented a markedly reduced RF (1.4) as compared to that of cisplatin (5.4). Similarly, in the DU145/DU145 Cis model, complex 4 showed an RF of 0.7, indicating greater efficacy in the resistant DU145 Cis cells compared to their cisplatin-sensitive counterparts. In contrast, cisplatin exhibited a 3-fold reduction in potency when applied to resistant cells relative to the wild type. Additionally, complex 4 was evaluated in MDA-MB-231 cells, a triple-negative breast cancer line known for its innate resistance to platinum-based drugs. Intriguingly, as compared with cisplatin, complex 4 demonstrated approximately 13.6-fold greater potency in MDA-MB-231 cells (Table ). These findings underscore the potential of complex 4 to circumvent both innate and acquired cisplatin resistance in cancer cells.

1. IC₅₀ Values (μM) of Complex 4 and Cisplatin against Cisplatin-Sensitive and -Resistant Cell Lines of A549 and DU145 When Incubated for 72 h .

cell lines/RF	complex 4 (μM)	cisplatin (μM)
A549	0.5 ± 0.1	4.9 ± 1.2
A549 Cis	0.6 ± 0.1	25.5 ± 0.1
RF(A549 Cis/A549)	1.4 ± 0.4	5.4 ± 1.4
DU145	1 ± 0.1	5.7 ± 0.2
DU145 Cis	0.8 ± 0.1	17.2 ± 1.1
RF(DU145 Cis/DU145)	0.7 ± 0.01	3 ± 0.1
MDA-MB-231	2.3±0.1	31.2±0.7

^aDose–response curves are presented in Figures S48.

Next, we evaluated the antitumor potency of 4 in a 3D tumor spheroid model. Compared to 2D cultured assays, the 3D models provide a more physiologically relevant approximation of in vivo tumor microenvironment, thus offering greater predictive accuracy in preclinical drug testing. − To this end, A549 3D tumor spheroids (∼0.5 mm diameter) were treated either with media alone or with media containing complex 4 (5 μM, 10 μM, or 20 μM), and spheroid diameters were recorded on alternate days. After 10 days of treatment, spheroids exposed to complex 4 showed a marked reduction in volume (Figures A,B and S49 for extended spheroid images). Notably, treatment with 20 μM complex 4 resulted in complete spheroid disintegration by day 6 (Figure A), whereas cisplatin has been reported to exert negligible effects on a similar model at the same concentration. These findings strongly suggest that complex 4 is highly effective in inducing cell death in 3D tumor models.

11.

Antitumor efficacy of complex 4 in A549 3D tumor spheroids. (A) Representative phase contrast images of 3D tumor spheroids, untreated or treated with complex 4 (extended data in Figure S49). (B) Change in tumor volume (mm³, mean ± SD) from days 0 to 10 in untreated and complex 4-treated spheroids (n = 3).

Conclusion

In summary, the development of metal-based chemotherapeutics capable of preferential and covalent mtDNA targeting represents a promising strategy to overcome resistance against nDNA-targeted Pt-based and organic drugs, thereby improving the chemotherapeutic outcome. However, only four reports featuring Pt­(II)-based reactive warheads have been documented to covalently target mtDNA, which exerted comparable or lower potency than cisplatin. Owing to the lack of intrinsic mtDNA-selective reactivity of Pt­(II) warheads, these systems rely on mitochondria-targeted delivery strategies. Our study reports the identification of a novel N-heterocyclic carbene (NHC) rhenium (Re) complex (complex 4) that selectively targets and damages mtDNA in cancer cells without the need for mitochondria-directed delivery while sparing nDNA. This remarkable selectivity seems to arise from the inherent reactivity preference of this class of molecules toward mtDNA over nDNA rather than from distribution-driven effects. Complex 4 was identified as the most promising candidate from a synthesized library of four [Re­(I)­(bisNHC)­(CO)₃Cl] complexes with varying lipophilicity. The antitumor potency of complex 4 outperformed that of nDNA-targeting cisplatin as well as previously reported all Pt-based mtDNA-targeting agents. SAR analysis suggested that lipophilicity is the critical determinant of cellular uptake and antitumor activity of this class of complexes. In-depth mechanistic studies revealed, despite nearly equal intracellular distribution between the two DNA-containing organelles, mitochondria and nucleus, complex 4 selectively forms covalent adducts with mtDNA. This remarkable mtDNA selectivity can be attributed to the unique chemical reactivity preference of the [Re­(I)­(bisNHC)­(CO)₃Cl] scaffold toward mtDNA, in addition to the increased vulnerability of mtDNA to undergo metalation. Covalent mtDNA targeting by 4 triggers various important downstream events, which include inhibition of mtDNA transcription and replication machineries (mtTFA and POLG) and downregulation of OXPHOS proteins. Further, mitochondrial impairment was substantiated by pronounced membrane depolarization, elevated levels of mitochondrial superoxide production, and intracellular ATP depletion. Notably, this multimodal mitochondrial collapse promoted mitophagy and ER stress, culminatively triggering necroptotic cell death. This novel mechanism of action enabled complex 4 to overcome both innate and acquired cisplatin resistance in the lung, prostate, and triple-negative breast cancer cells. Moreover, complex 4 presented significant activity in 3D tumor spheroid models, highlighting its efficacy in more clinically relevant systems. Overall, this study reports for the first time the unique ability of a non-Pt-reactive [Re­(I)­(bisNHC)­(CO)₃Cl] warhead to covalently target mtDNA without the need for mitochondria-targeting vectors, thereby simplifying molecular design and paving the way for the development of next-generation metal-based mtDNA-targeted potent chemotherapeutics with improved clinical potential.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/jacsau.6c00579.

• Synthesis of complexes, characterization data including crystallographic details, NMR spectra, mass spectra, and procedures for all in vitro cellular assays (PDF)

†.

A.D. and S.D. contributed equally to this work. M.P. designed and supervised the project. A.D. synthesized and characterized the compounds and performed analytical experiments. SD and S.S. performed cellular experiments and interpreted the data. S.C. performed the RTPCR experiments and analyzed the data. P.S.N performed mitophagy imaging and assisted with DNA gel electrophoresis. S.P solved the X-ray structures. All authors contributed to preparing and editing the manuscript and also approved the submission.

The authors declare no competing financial interest.