A bexarotene-attached Re(I) tricarbonyl complex for NADH oxidation and ROS-mediated cancer phototherapy

Abstract

An axially substituted polypyridyl Re(CO)₃ complex bearing bexarotene triggered caspase-3/7-mediated apoptosis in cancer cells through ROS generation and NADH photo-oxidation.

---

Photoactivated cancer therapy (PACT) has emerged as a promising anticancer strategy for achieving spatial and temporal control over cytotoxicity while minimizing off-target effects.¹ So far, the clinically used photosensitizers are built upon tetrapyrrolic scaffolds, e.g., phthalocyanine, porphyrin, etc., which control their photophysical as well as biological behaviour.² Due to these shared structural motifs, many of them exhibit similar limitations, including low water solubility, poor photostability, tedious synthesis/purification, and slow clearance, often leading to chronic photosensitivity.^2,3 To address these challenges, metal complexes have attracted attention due to their unique photophysical and photochemical properties, which can be finely tuned by ligand design and metal choice.⁴ Moreover, transition metal complexes also possess high photostability, unique excited state activity, and redox behaviour for generating reactive oxygen species (ROS) and multimodal actions, such as NADH photo-oxidation.⁵ In this regard, Re(I) tricarbonyl (Re(CO)₃) complexes have garnered a surge in interest as light-responsive anticancer agents owing to their long-lived excited states and favorable photochemical/photophysical properties.⁶ These complexes primarily operate via light-induced ¹O₂ and other ROS generation to kill cancer cells selectively.⁶ For example, Mao and co-workers developed a carbonic anhydrase IX appended Re(I) photosensitizer showing immunogenic anticancer potential.⁷ The Wilson group realized the light-responsive anticancer potential of a wide range of phenanthroline-based Re(CO)₃ complexes via ROS generation and CO toxicity.⁸ Recently, our group reported phenanthroline- and terpyridine-based Re(CO)₃ complexes showing light-triggered anticancer activity via synergistic ROS generation and NADH photo-oxidation.⁹ Zhang and co-workers also explored the light-activated ROS generation and NADH photo-oxidation mediated immunotherapeutic potential of phenanthroline-based Re(CO)₃ complexes.¹⁰ Despite these advances, very limited Re(CO)₃ complexes demonstrating simultaneous ROS production and NADH photo-oxidation under visible light have been reported, representing a combination that could synergistically disrupt mitochondrial function and trigger apoptotic pathways in cancer cells.⁵ Moreover, the structural diversity of light-responsive Re(CO)₃ complexes has largely been restricted to modifications on α-diimine ligands or axial substitution with pyridine, isonitrile, or phosphine derivatives.^5,6,11 In this study, for the first time, we sought to explore the potential of carboxylate derivatives as axial substituents in Re(CO)₃ complexes to expand the scope of this class of photo-responsive Re(CO)₃ complexes for photoactivated cancer therapy. Successful implementation might allow the use of a wide range of bioactive/photoactive carboxylate derivatives as axial substituents. In addition, although RXR agonists have been explored in cancer therapeutics, their direct conjugation to a PACT agent and their impact on synergistic apoptosis pathways under light irradiation have also not been reported in Re(CO)₃ complexes.

We report two novel [Re(CO)₃(N^N)L] complexes, Re1 and Re2, incorporating bathophenanthroline as the diimine ligand (N^N), and either benzoic acid (Re1) or bexarotene (Re2) as the axial ligand (L) (Fig. 1a). Benzoic acid was utilized as the control for biologically active bexarotene. Bexarotene, a clinically approved retinoid X receptor (RXR) agonist, is known for its established pro-apoptotic activity through modulation of gene expression and transcription factors.¹² In addition, bexarotene is also used in combination with other anticancer drugs to improve their efficacy and performance, making it a rational choice for additional bioactivity to the Re(I) scaffold.¹² Bathophenanthroline was used to exploit its extended π-conjugation and rigid planar structure to increase photosensitivity and excited state stabilization.¹³

Fig. 1 (a) Structures of Re1 and Re2. (b) UV-Vis spectra of Re1 and Re2 in DMSO : H₂O (1 : 9 v : v). (c) X-ray structure of Re1.

Re1 and Re2 were synthesized from the parent [Re(CO)₃(N^N)Cl] complexes by substituting Cl with L (anionic benzoic acid (Re1); anionic bexarotene (Re2)) in the presence of AgOTf, where N^N was bathophenanthroline (Scheme S1, SI). The complexes were characterized using several spectroscopic techniques (Fig. S1–S8, SI). In ¹H NMR, the peaks between 6.5 and 9.5 ppm were attributed to aromatic protons of Re1/Re2, and the peaks between 0.9 and 1.6 ppm and 4.9 and 5.7 ppm (in Re2) were attributed to the alkyl group and unsaturated protons, respectively (Fig. S1 and S2, SI). The presence of carbonyls in Re1 and Re2 was confirmed by the ¹³C{¹H} NMR peaks observed between 185 and 200 ppm and by characteristic FT-IR bands between 1885 and 2020 cm⁻¹ (Fig. S3–S6, SI). The HRMS spectra in MeOH displayed a molecular ion peak corresponding to [M + H]⁺ (Fig. S7 and S8, SI). The UV-Vis spectra of Re1 and Re2 revealed an absorption band at ∼385 nm with an extended tail (up to 500 nm) in the visible region (Fig. 1b) across different pH values (6–8) (Fig. S9, SI), suggesting their ability to absorb visible light (which might be useful to induce the light-triggered anticancer effect). Re2 displayed a weak emission at ∼625 nm upon 380 nm excitation (Fig. S10, SI). High photostability is crucial for any phototherapeutic to avoid photobleaching or off-target problems.⁴ Interestingly, both complexes displayed excellent photostability up to 4 h, as evidenced by the insignificant change in NMR and UV-Vis spectra (Fig. S11 and S12, SI). Re2 exhibited a logP_o/w (octanol–water coefficient) value of +1.22 ± 0.12 (Fig. S13, SI), indicating its lipophilic nature.

Re1 was crystallized in the P121/n space of the monoclinic crystal system. Re1 had a distorted octahedral shape featuring a ReC₃N₂O coordination core, with axial benzoate, three facial carbonyls, and two nitrogens of bathophenanthroline (Fig. 1c and Fig. S14, SI). The axial Re–O bond length was comparatively shorter than the axial Re–Cl bond length in the corresponding chloride complexes, indicating the formation of a stronger bond.^9,14 Selected crystallographic parameters and selected bond lengths/angles are provided in Tables S1 and S2, SI, respectively. Furthermore, the computational studies were performed to gain insight into the electronic and photophysical behaviour of Re1 and Re2. The complexes were optimized in their different states using the CAM-B3LYP functional with combinatorial (LANL2DZ with pseudo-LANL2 for Re and 6-31g* for the other atoms) basis sets using Gaussian 16 (Fig. 2a and Fig. S15, S16, SI).⁹ The analysis of FMOs revealed that the HOMOs of Re1 and Re2 were localized on the Re(CO)₃ core with slight involvement of bexarotene in Re2, while the LUMOs of Re1 and Re2 were purely distributed on the π* orbitals of bathophenanthroline (Fig. S17 and S18, SI). The ΔE_g (= E_LUMO − E_HOMO) value for Re2 was slightly lower than that for Re1, indicating better photo-sensitivity of Re2 (Table S3, SI).¹⁵ The energies of the ten singlet/triplet excited states were determined to understand the energy differences between the S₀ and the corresponding excited S_n/T_n states (Tables S4 and S5, SI). The intersystem crossing (ISC) efficiently occurs with a small energy gap (ΔE_S1–Tn < 0.3 eV) between the S₁ and T_n states.¹⁵ Thus, based on excitation energy analysis, the possible channels for ISC of Re2 are given in Fig. 2b. The NTO analysis of these transitions indicated the involvement of ¹LLCT-to-³LLCT transitions (Fig. 2c and Fig. S19, S20, SI). The energy difference, ΔE_S0–T1, underlines the efficacy of the lowest-lying triplet state that has adequate energy to generate ¹O₂ (i.e., >0.98 eV) to proceed via the PDT type-II pathway (Fig. 2b). The SOMO plots and spin density plots at the triplet excited state of Re1 and Re2 (Fig. S21 and S22, SI) revealed that the unpaired electrons are distributed around Re(I) and bathophenanthroline, indicating their mixed metal–ligand-based character.

Fig. 2 (a) Optimized structure of Re2 in the ground state. (b) Calculated excited state energy and possible ISC channels of Re2 (energy in eV). (c) NTOs for the S₀ → S₁/S₂ transition for Re2.

The absorbance within the visible range, high photostability, and potent theoretical ¹O₂ generation ability of Re1 and Re2 inspired us to investigate them as PACT agents. In PACT, ¹O₂ production plays a critical role in causing oxidative stress, disrupting membranes, denaturing proteins, and damaging DNA.^1,6,8 The ¹O₂ generation ability was determined using diphenylisobenzofuran (DPBF) as a ¹O₂ probe.¹⁶ The absorbance of DPBF remained unchanged in the presence of Re1 and Re2 (Fig. S23, SI) under dark conditions, indicating no detectable ¹O₂ generation. However, there was a gradual decrease in DPBF-based absorption peaks upon light exposure (400–700 nm, 10 J cm⁻²), exhibiting ¹O₂ generation (Fig. 3a and Fig. S24, SI), even under different pH conditions (Fig. S25, SI). The ¹O₂ quantum yield (Φ_Δ) of Re1 and Re2 was 0.12–0.14 with [Ru(bpy)₃]Cl₂ as the standard (Φ_Δ = 0.22).⁹ The Kim group has reported that several photosensitizers capable of generating light-induced ¹O₂ can also facilitate photocatalytic NADH oxidation.¹⁷ NADH is a vital coenzyme that actively participates in the electron transport chain (ETC) and plays a crucial role in metabolism.^5,17,18 Thus, its oxidation can impair ETC function in cancer cells, ultimately leading to cell death.^5,18 The obtained results revealed that there was no notable change in NADH (150 μM) absorbance in the presence of Re1 and Re2, indicating no oxidation of NADH (Fig. S26, SI) in the dark. Upon light (400–700 nm, 10 J cm⁻²) exposure, the characteristic absorbance of NADH at ca. 339 nm gradually decreased; at the same time, the absorbance at ca. 256 nm corresponding to NAD⁺ progressively increased in the presence of Re1/Re2, indicating the photo-oxidation of NADH to NAD⁺ (Fig. 3b and Fig. S27, SI). Moreover, the NADH photo-oxidation ability of Re2 did not change in the presence of different ROS scavengers (D-mannitol for radical-based ROS and NaN₃ for ¹O₂) (Fig. S28, SI), indicating ROS-independent Re2-mediated NADH photo-oxidation. The turnover frequency (TOF) of Re2 (TOF = 13.6 h⁻¹) was comparatively higher than that of Re1 (TOF = 9.0 h⁻¹) for NADH to NAD⁺ oxidation. These findings suggested that Re1/Re2 could act as a phototherapeutic agent that can produce ROS, such as ¹O₂, and oxidize NADH upon light exposure.

Fig. 3 (a) ¹O₂ generation by Re2 under light exposure in DMSO : PBS (2 : 98 v/v). (b) NADH photo-oxidation by Re2 under light exposure in DMSO : PBS (2 : 98 v/v).

The ability of Re1 and Re2 to produce ¹O₂ and oxidize NADH under light exposure prompted us to investigate their anticancer efficacy against A549 (lung cancer) and MCF-7 (breast cancer) cells, and normal HEK-293 (human embryonic kidney) cells under both dark and light conditions. The dark and light IC₅₀ values of Re1 and Re2 are provided in Table S6, SI. Re1 and Re2 exhibited minimal cytotoxic effect against both lung (IC₅₀ > 50 μM for Re1 and ∼36 μM for Re2) and breast (IC₅₀ = ∼21 μM for Re1 and ∼19 μM for Re2) cancer cells under dark conditions (Fig. S29–S32, SI). However, their cytotoxicity was notably improved upon light exposure as the IC₅₀ for Re2 was found to be ∼0.27 and ∼1.02 μM against MCF-7 and A549 cells, respectively, while IC₅₀ for Re1 was found to be ∼3.91 and 12.01 μM against MCF-7 and A549 cells, respectively. The phototoxicity index (PI = dark IC₅₀/light IC₅₀) of Re2 (PI up to 71) was much higher than that of Re1 (PI up to 5.6), possibly due to its higher ¹O₂ generation/NADH photo-oxidation efficacy. Importantly, Re1 and Re2 did not exhibit cytotoxicity (IC₅₀ > 50 μM) towards normal HEK-293 cells, irrespective of light or dark conditions (Fig. S33 and S34, SI), aligning well with the clinical rationale of photosensitizers. This result indicated that Re2 selectively killed cancer cells without harming normal cells with a high selectivity index (SI = light IC_{50 normal cells}/light IC_{50 cancer cells}) of up to 185. The high selectivity of Re2 can be attributed to the RXR targeting nature of the appended bexarotene motif.¹² To gain insight into the effect of bexarotene on the Re(CO)₃ complex, a molecular docking study was performed with Re1, Re2, and bexarotene with the RXRα receptor (PDB ID: 1MVC) (Fig. S35, SI).¹⁹ The docking results (Table S7, SI) suggested an improvement in the RXRα receptor-binding capability of Re2. Furthermore, the interaction analysis revealed that both Re2 and bexarotene formed similar hydrogen bonds with ARG A:316 of the RXRα receptor, a key residue involved in RXRα activation and increased hydrophobic interaction within the RXRα receptor's binding pocket (Fig. S36–S38, SI).

For cell death mechanistic studies, the DCFH-DA (2,7-dichlorodihydrofluorescein diacetate) assay was used to determine in-cell ROS generation in MCF-7 cells.^9,20 The obtained results revealed that Re2 caused significant in-cell ROS generation in MCF-7 cells upon light (400–700 nm, 10 J cm⁻²) exposure, as indicated by bright green fluorescence in Fig. 4a. In contrast, Re2 alone (dark conditions) did not induce notable ROS production, reaffirming its light-dependent ROS-generating capability (Fig. 4a). Previous reports suggested that the photosensitizers showing light-triggered ROS production and NADH photo-oxidation can effectively disrupt MMP (mitochondrial membrane potential) and trigger apoptosis in cancer cells.^5,9 Hence, we assessed the change in the MMP of Re2-treated MCF-7 cells under light and dark conditions by the JC-1 assay (Fig. 4b). In this method, the JC-1 dye accumulates in mitochondria and displays red emission at higher MMP, whereas at lower MMP, it displays green emission and gets dispersed in the cytoplasm. Under dark conditions, Re2-treated MCF-7 cells displayed red emission, revealing no notable MMP changes (Fig. 4b). In contrast, Re2-treated MCF-7 cells displayed green emission upon light (400–700 nm, 10 J cm⁻²) exposure, reflecting a significant loss and change in MMP (Fig. 4b). Thus, Re2 effectively altered the MMP of MCF-7 cells under light rather than dark conditions. The change in the MMP of Re2-treated MCF-7 cells under light exposure suggests a higher probability of mitochondrial damage triggering apoptosis.^20,21 Thompson and coworkers have shown a direct correlation between mitochondrial integrity, MMP, and apoptosis.²⁰ Also, bexarotene derivatives are known to induce apoptosis in cancer cells.¹² Hence, the mechanism of cell death induced by Re2 was investigated in MCF-7 cells by AO (acridine orange)/EtBr (ethidium bromide) staining.²⁰ As shown in Fig. 4c, MCF-7 cells treated with Re2 or dark/light only exhibited a well-organized cytoplasm and intact green-stained nuclei, indicating that Re2 is mostly non-toxic under dark conditions. However, light-exposed Re2-treated MCF-7 cells produced bright green/yellowish nuclei along with membrane blebbing, indicating the occurrence of early and late apoptosis. Annexin V-FITC/PI dual staining assays revealed that the controls and Re2 under dark conditions presented negligible cell death. However, under Re2 + light treatment, the cell death was notably increased, with ca. 10% early apoptosis, 31% late apoptosis, and 25% necrosis (Fig. S39, SI). Caspase 3/7 are the key executioners of apoptosis, and their activation ultimately results in programmed cell death.¹⁸ The potential of Re2 for caspase activation was determined in MCF-7 cells using caspase 3/7 and SYTOX red assays.²⁰ The results revealed that the only Re2 + light-treated group demonstrated caspase 3 activation (Fig. S40, SI). Based on these findings, it could be concluded that Re2 caused ROS generation and NADH photooxidation under light exposure, which compromised MMP and activated caspase-3/7 to induce apoptosis in MCF-7 cells.

Fig. 4 (a) ROS generation induced by Re2 in MCF-7 cells. Scale bar = 400 μm. (b) Mitochondrial depolarization induced by Re2 in MCF-7 cells. Scale bar = 100 μm. (c) AO/EtBr assay indicating apoptosis in MCF-7 cells induced by Re2. Scale bar = 100 μm.

Overall, we report two Re(CO)₃ complexes (Re1 and Re2) featuring bathophenanthroline as the diimine ligand and either benzoate (Re1) or bexarotene (Re2) as the axial carboxylate ligand. Upon light exposure, these complexes induced the oxidation of NADH to NAD⁺ and generated ¹O₂, thereby activating dual mechanisms of cancer cell death. Re2 demonstrates potent photocytotoxicity against A549 and MCF-7 cancer cells via ROS-mediated mitochondrial dysfunction and caspase-3/7-dependent apoptosis while sparing normal HEK 293 cells. This work highlights the phototherapeutic potential of integrating an axial carboxylate bioactive ligand with a photoactive Re(I) scaffold to develop next-generation phototherapeutic agents for photoactivated cancer therapy.

This work was supported by the SERB (now ANRF), India (SRG/2022/000030). R. K. thanks the GOI for the PMRF. R. K. performed the synthesis, characterisation, in-solution studies, X-ray crystallography, TD-DFT calculations, and MD studies. V. S. and B. K. performed the biological studies. S. B. formulated the concept and overall project.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data supporting this article have been included as part of the SI.

Experimental protocols, synthesis, characterization data, in solution studies, and bioassays data. See DOI: https://doi.org/10.1039/d5cc03374h.

CCDC 2463073 contains the supplementary crystallographic data for this paper.²²

References

1. (a) H. Shi and P. J. Sadler, Angew. Chem., Int. Ed., 2025, 64, e202423335; (b) K. M. Kuznetsov, K. Cariou and G. Gasser, Chem. Sci., 2024, 15, 17760–17780.
2. (a) B. Zheng, Q. He, X. Li, J. Yoon and J. Huang, Coord. Chem. Rev., 2021, 426, 213548; (b) W. Yina, P. Lib, H. Huang, L. Fengb, S. Liub, X. Liu and F. Bai, RSC Adv., 2024, 14, 29368–29383.
3. X. Zhao, J. Liu, J. Fan, H. Chao and X. Peng, Chem. Soc. Rev., 2021, 50, 4185–4219.
4. (a) J. Karges, Angew. Chem., Int. Ed., 2022, 61, e202112236; (b) S. A. McFarland, A. Mandel, R. Dumoulin-White and G. Gasser, Curr. Opin. Chem. Biol., 2020, 56, 23–27; (c) L. Skos, Y. Borutzki, C. Gerner and S. M. Meier-Menches, Curr. Opin. Chem. Biol., 2023, 73, 102257.
5. A. K. Yadav, R. Kushwaha, A. A. Mandal, A. Mandal and S. Banerjee, J. Am. Chem. Soc., 2025, 147, 7161–7181.
6. (a) K. Schindler and F. Zobi, Molecules, 2022, 27, 539; (b) A. S. Sharma, N. Vaibhavi, B. Kar, U. Das and P. Paira, RSC Adv., 2022, 12, 20264–20295; (c) R. Kushwaha and S. Banerjee, Future Med. Chem., 2025, 17, 1101–1103.
7. X. Su, W. Wang, Q. Cao, H. Zhang, B. Liu, Y. Ling, X. Zhou and Z. Mao, Angew. Chem., Int. Ed., 2022, 61, e202115800.
8. (a) K. M. Knopf, B. L. Murphy, S. N. MacMillan, J. M. Baskin, M. P. Barr, E. Boros and J. J. Wilson, J. Am. Chem. Soc., 2017, 139, 14302–14314; (b) S. C. Marker, A. P. King, R. V. Swanda, B. Vaughn, E. Boros, S. Qian and J. J. Wilson, Angew. Chem., Int. Ed., 2020, 59, 13391–13400.
9. (a) R. Kushwaha, V. Singh, S. Peters, A. K. Yadav, T. Sadhukhan, B. Koch and S. Banerjee, J. Med. Chem., 2024, 67, 6537–6548; (b) R. Kushwaha, A. Upadhyay, S. Saha, A. K. Yadav, A. Bera, A. Dutta and S. Banerjee, Dalton Trans., 2024, 53, 13591–13601.
10. D. Li, G. Wen, H. Wang, Q. Ren, D. Wang, A. Dao, H. Huang and P. Zhang, J. Med. Chem., 2025, 68, 3749–3763.
11. (a) L. C. Lee and K. K. Lo, J. Am. Chem. Soc., 2022, 144, 14420–14440; (b) B. Kar, U. Das, N. Roy and P. Paira, Coord. Chem. Rev., 2023, 474, 214860.
12. (a) D. Shen, X. Yu, Y. Wu, Y. Chen, G. Li, F. Cheng and L. Xia, Expert Rev. Anticancer Ther., 2018, 18, 487–499; (b) L. T. Farol and K. B. Hymes, Expert Rev. Anticancer Ther., 2004, 4, 180–188.
13. A. C. Hachey, D. Havrylyuk and E. C. Glazer, Curr. Opin. Chem. Biol., 2021, 61, 191–202.
14. B. Machura, M. Wolff, M. Jaworska, P. Lodowski, E. Benoist, C. Carrayon, N. Saffon, R. Kruszynski and Z. Mazurak, J. Organomet. Chem., 2011, 696, 3068–3075.
15. G. Xu, C. Li, C. Chi, L. Wu, Y. Sun, J. Zhao, X.-H. Xia and S. Gou, Nat. Commun., 2022, 13, 3064.
16. A. Linero-Artiaga, L. Servos, V. Rodríguez, J. Ruiz and J. Karges, J. Med. Chem., 2025, 68, 7792–7806.
17. M. Li, Y. Xu, Z. Pu, T. Xiong, H. Huang, S. Long, S. Son, L. Yu, N. Singh, Y. Tong, J. L. Sessler, X. Peng and J. S. Kim, Proc. Natl. Acad. Sci. U. S. A., 2022, 119, e2210504119.
18. A. Chiarugi, C. Dölle, R. Felici and M. Ziegler, Nat. Rev. Cancer, 2012, 12, 741–752.
19. (a) G. M. Morris, R. Huey, W. Lindstrom, M. F. Sanner, R. K. Belew, D. S. Goodsell and A. J. Olson, J. Comput. Chem., 2009, 16, 2785–2791; (b) Y. N. Nosova, D. S. Karlov, S. A. Pisarev, I. A. Shutkov, V. A. Palyulin, M. Baquié, E. R. Milaeva, P. J. Dyson and A. A. Nazarov, J. Organomet. Chem., 2017, 839, 91–97.
20. A. Mandal, V. Singh, S. Peters, A. A. Mandal, T. Sadhukhan, B. Koch and S. Banerjee, Dalton Trans., 2025, 54, 6785–6789.
21. E. Gottlieb, S. Armour, M. H. Harris and C. B. Thompson, Cell Death Differ., 2003, 10, 709–717.
22. R. Kushwaha, V. Singh, B. Koch and S. Banerjee, CCDC 2463073, Experimental Crystal Structure Determination, 2025 DOI:10.5517/ccdc.csd.cc2np0zc.

---