Unexpected Photocatalytic Degeneration of NAD⁺ for Inducing Apoptosis of Hypoxia Cancer Cells

Introduction

Life is based on continuous cellular metabolism processes, where many substrates and products are often involved in various chemical reactions.^1,2 These synergetic reactions, including intracellular redox, enzyme cascade, and reversible reaction, maintain homeostasis of the biological system in individual bodies.³ However, biological metabolic processes with erroneous chemical reactions can cause metabolic disorders and dysfunction, even injury, cancer, and autoimmune disease.^4–7 Customized chemical reactions that could regulate specific biological processes are regarded as an advanced and promising strategy for treating diseases.⁸ However, the complex physiological environments, including the near-neutral pH, mild temperature, pressure, and hypoxic microenvironment, also put many preconditions on developing these reactions.⁹ The additional demands of excellent controllability and biocompatibility make these investigations more challenging.¹⁰ Moreover, cellular self-healing ability¹¹ would further weaken the effects of these reactions because living cells prefer to maintain a chemical balance in the body.

Cancer, with high morbidity and mortality, has significantly threatened human life.^12–14 Hypoxic microenvironments often cause worse disease outcomes by protecting cancer cells against apoptosis and enhancing the ability of tumor metastasis.^3,15,16 Nicotinamide adenine dinucleotide (NAD) is one of the most important coenzymes involved in intracellular redox reactions including cancer metabolism,¹⁷ and it generally exists in two forms including oxidized NAD⁺ and reduced NADH.¹⁸ NAD⁺/NADH levels can affect cancer cell growth by inhibiting the related energy production pathways,⁶ such as glycolysis, the tricarboxylic acid cycle, and the mitochondrial electron transport chain (ETC). Hence, regulation of NAD⁺/NADH levels provides an effective strategy for the inhibition of cancer cells by driving the NAD⁺ and NADH interconversion.^8,19,20 However, the regulated ratio of NAD⁺/NADH can be gradually restored to rebalance the redox levels through the cell repair process.¹¹ Photocatalytic reactions exhibit great potential in achieving desirable intracellular reactions due to the strong controllability, mild reaction conditions,²¹ and noninvasive remote regulation.^22,23 But the poor biocompatibility of noble metal catalysts^24,25 could be harmful to living organisms resulting in limited applications in vivo. In comparison, water-soluble conjugated polymers (WSCPs) possess high biocompatibility^26–29 and electron transfer efficiency that are beneficial to their applications in photoinduced processes in living organisms.^30–32 The ability of WSCPs to catalyze cell-mediated polymerizations^33–36 could be exploited in intracellular photocatalytic reactions, which create NAD⁺/NADH imbalance and regulate cell activities.

In this work, we report an emerging photocatalytic reaction and realize the irreversible decomposition of NAD⁺ to nicotinamide and adenosine diphosphate (ADP)-ribose by the cationic poly(fluorene-co-phenylene) (PFP). The chemical reaction was induced in hypoxic tumor cells to successfully disrupt the cell activity (Scheme 1). The mechanistic investigation shows that the PFP acts as the photocatalyst in the system, and the photoelectrons transfer from the excited PFP (PFP*) to NAD⁺ to yield nicotinamide and ADP-ribose instead of the naturally occurring bioactive NADH. Therefore, the photocatalytic reaction mediated by PFP irreversibly regulates NAD⁺ rather than the common strategy based on a reversible reaction between NAD⁺ and active NADH.³ Inside these tumor cells, the reaction compromises mitochondrial functions that are mainly manifested as blockage of ETC, decreased mitochondrial membrane potential (MMP), and decreased adenosine triphosphate (ATP) level. The disturbed cellular redox cycle and irreversible damage of mitochondria by PFP under irradiation can cause apoptosis and finally disrupt the tumor cell activity.

• Download figure
• Download PowerPoint

Experimental Section

General methods and materials

Cationic PFP was synthesized according to the reported literature procedure.³² All chemical reagents are commercially available and used without further purification. 5,5-Dimethyl-1-pyrroline N-oxide (DMPO) was purchased from Sigma-Aldrich (St. Louis, Missouri, USA). 4T1 cells were purchased from the cell culture center of the Institute of Basic Medical Sciences, Chinese Academy of Medical Sciences (Beijing, China). Roswell Park Memorial Institute 1640 medium (RPMI 1640) was purchased from Hyclone (Beijing, China). Fetal bovine serum was purchased from Sijiqing Biological Engineering Materials (Hangzhou, China). 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) was purchased from Xinjingke Biotechnology Co., Ltd. (Beijing, China). MMP assay kit (JC-1), ATP, and BCA Assay Kits were acquired from Beyotime Biotechnology (Shanghai, China). The NAD⁺/NADH Ratio Kit was purchased from AAT Bioquest (Pleasanton, California, USA). Deionized water was obtained from a Milli-Q system (Millipore, Bedford, MA, United States). 1,1′-dioctadecyl-3,3,3′,3′-tetramethylindodicarbocyanine (DiD) was provided by Life Technologies (Waltham, Massachusetts, USA). UV–vis absorption spectra were measured by a Thermo Scientific Evolution 201 spectrophotometer (Thermo Fisher, Waltham, Massachusetts, USA), and the fluorescence spectra were taken on a Hitachi F-4500 fluorometer (Hitachi, Tokyo, Japan) equipped with a Xenon lamp excitation source. Ultraviolet photoelectron spectroscopy (UPS) was conducted on an AXIS ULTRA DLD instrument (Kratos, Manchester, UK). Electrochemistry measurements were conducted on an electrochemical workstation (Autolab PGATAT 302N, Metrohm, Herisau, Switzerland). NMR spectra were obtained from a Bruker Avance HD 400 MHz and Bruker AVANCE 600 MHz (Rheinstetten, Germany). The electron paramagnetic resonance (EPR) spectra were measured by Bruker Xenoe nano (Bruker EMXnano, Fällanden, Switzerland). The MTT assay was recorded on a microplate reader (BioTek, Winooski, Vermont, USA). Confocal laser scanning microscopy (CLSM) images were conducted on confocal laser scanning biological microscope (Olympus FV 1200-BX61, Olympus, Hataya, Japan).

Intracellular NAD⁺/NADH ratio assay

4T1 cells were treated in the medium at 37 °C at a density of 5 × 10⁵ cells/well in 6-well plates overnight. Then, the medium with PFP (8 μM) and triethanolamine (TEOA; 4 mM) was used to cultivate 4T1 cells at 37 °C for 12 h. The cells were transferred to a hypoxic environment (1% O₂, 5% CO₂, 37 °C) for 6 h. After irradiation with 50 mW/cm² for 20 min, the cells were treated with 250 μL cell lysis buffer. The centrifugal supernatant was collected and used to measure intracellular NAD⁺/NADH ratio by commercial kit.

MMP measurement

The 4T1 were added in confocal dishes at a density of 8 × 10⁴ cells/mL and treated in the RPMI medium containing PFP (8 μM) and TEOA (4 mM) at 37 °C for 12 h. Then the cells were incubated in 1% O₂ and 5% CO₂ at 37 °C for 6 h. After irradiation under white light with 50 mW/cm² for 30 min, the 4T1 cells were incubated with JC-1 (5×) for 20 min. Then cells were rapidly washed three times by JC-1 (1.25×). The CLSM images were collected at 520–550 nm (λ_ex, 488 nm), and 575–675 nm (λ_ex, 488 nm), respectively.

EPR

A mixture of 100 μL PFP (1 mM, dimethyl sulfoxide), 20 μL TEOA (1 M), 100 μL DMPO (1 M), 200 μL NAD⁺ (50 mM), and 580 μL phosphate-buffered saline was irradiated under white light (50 mW/cm²), and the EPR spectra were recorded every 5 min. (Center of Field, 3443.90 G; Sweep Width, 200.0 G; Power, 5 mW; Power Attenuation, 20.00 dB; Modulation Amplitude, 1.0 G; Number of Scans, 6.)

Results and Discussion

Characteristics of photocatalyst

To verify the feasibility of photocatalytic reaction of PFP to catalytically convert NAD⁺, the photophysical information, energy level, and fluorescence decay kinetics of PFP were collected accordingly. As shown in Figure 1a, PFP exhibited a maximum absorption wavelength of 380 nm, and the peak of fluorescence emission was 429 nm. The band gap of PFP was calculated with a standard method to be 2.97 eV.⁴ The photoelectric response under a solar simulator demonstrated the possible conversion of absorbed photons into excited electrons ( Supporting Information Figure S1). The highest occupied molecular orbital (HOMO) was determined to be −5.82 eV by UPS (Figure 1b), and the lowest unoccupied molecular orbital (LUMO) was calculated to be −2.85 eV with an optical band gap of 2.97 eV. The E_red and E_ox of PFP were +1.32 V and −1.65 V (vs normal hydrogen electrode [NHE]), respectively, obtained by calculations described in the literature.^37,38 The reduction potential of NAD⁺/NADH is −0.54 V versus NHE,³⁹ which is lower than E_red of PFP (+1.32 V vs NHE), and, therefore, suitable to capture the electron from PFP*. The fluorescence decay kinetic curves show that the fluorescence lifetime of PFP decreased from 104 to 23 ns with the addition of NAD⁺ (Figure 1c). The electron-transfer pathway from PFP to NAD⁺ under illumination was predicted based on the HOMO/LUMO potential (Figure 1d). The electron captured from the HOMO of PFP could be transferred to NAD⁺ in the reaction process. All the above characterizations indicate that under illumination PFP generates the holes and electrons, and then the photoelectrons from the PFP HOMO to LUMO transition might reduce NAD⁺ into the product. The holes are consumed by TEOA whose redox potential (+0.53 V vs NHE) is more negative than PFP (+1.32 V vs NHE).

• Download figure
• Download PowerPoint

The photocatalysis reaction to decompose NAD⁺

According to NMR and mass spectrometry, the photocatalytic reaction of NAD⁺ generated nicotinamide and ADP-ribose as the products (Figure 2a and Supporting Information Figure S2). Before illumination, ¹H NMR peaks corresponding to the nicotinamide unit (a–d), adenine unit (g and h), and the two glycosidic protons (e and f) in NAD⁺ were identified (black trace), whereas PFP peaks were invisible due to the low concentration. After illumination, the complicated ¹H NMR spectral pattern consisted of free nicotinamide (a′–d′, red trace), ADP-ribose (e′–h′), and unreacted NAD⁺ (not labeled). Major changes of the peaks of nicotinamide and the glycosidic proton e/e′ indicate that the photocatalytic cleavage occurred between nicotinamide and its adjacent ribose. In contrast, minor peak shifts of adenine and its adjacent glycosidic proton f/f′ suggest an intact glycoside on adenine. To further resolve the products, diffusion ordered spectroscopy (DOSY) was performed on the sample after 1 h illumination (Figure 2b). Peaks corresponding to the free nicotinamide exhibited two- to threefold higher diffusion coefficients than other major peaks, and the trend was consistent with the smaller molar mass of nicotinamide (122 g/mol) compared to NAD⁺ and ADP-ribose. Although the latter two compounds could not be resolved in ¹H NMR or DOSY spectra, they were identified on the electrospray ionization-mass spectrometry (ESI-MS) spectrum. The peaks at m/z of 558.1 and 540.1 correspond to ADP-ribose and its dehydrated derivatives, respectively, (Figure 2c), and the peak at 662.1 is unreacted NAD⁺. To monitor the reaction progress to characterize the kinetics of the photocatalysis, the ¹H NMR integrations of H(d) in NAD⁺ and H(d′) in nicotinamide were calculated (Figure 2d). After 1 h irradiation, 113 μM nicotinamide was obtained, and the products continued to accumulate with extended irradiation time.

• Download figure
• Download PowerPoint

Intracellular catalysis study and therapeutic efficacy

The intracellular catalytic effect of PFP photocatalyst on 4T1 cancer cells was investigated. The stability of the cationic conjugated PFP in various physiological solutions was measured ( Supporting Information Figure S3). The similar absorption spectra indicate excellent stability of the conjugated PFP skeleton as a photocatalyst. Meanwhile, the efficiency of this photocatalytic reaction was measured under hypoxic and acidic conditions to simulate the tumor microenvironment. As shown in Supporting Information Figure S4, the efficiency of the photocatalytic reaction was not affected by the low oxygen concentration or acidic pH. It indicates that this photocatalytic reaction mediated by PFP worked in the tumor microenvironment. To exclude the photosensitization effects of PFP, the reactive oxygen species (ROS) production capacity of PFP was measured by 2,7-dichlorodihydrofluorescein (DCFH; Supporting Information Figure S5). In contrast to the efficient photogeneration of ROS under aerobic conditions, negligible ROS were produced by the illumination of PFP under hypoxic conditions. It indicates that the influence of common ROS is not a major contributing factor in this antitumor system. As illustrated in Supporting Information Figure S6, CLSM images of 4T1 cells showed that the blue signal from PFP was surrounded by an entire membrane (labeled with DiD and emitting red signal), which elucidates the successful internalization of PFP. As shown in Supporting Information Figure S7, the PFP, which located in the lysosome, was taken up into cells over the incubation time and retained intracellularly for more than 72 h. This indicates that PFP has the potential to act as a long-term catalyst in cancer cells. To verify that the catalytic reaction could occur in cells, the intracellular NAD⁺/NADH ratio in hypoxic 4T1 cells treated with PFP under light irradiation was measured. As shown in Figure 3a, the NAD⁺/NADH ratio of the optical irradiation group treated with PFP decreased from 73% to 65% compared to the dark group. Thus, PFP can catalytically convert NAD⁺ into nicotinamide and ADP-ribose under white illumination. As an important coenzyme pair in the mitochondrial ETC, and a raw material for photocatalytic reactions, NAD⁺ can affect the function of mitochondria. The continuous depletion of NAD⁺ can lead to mitochondrial apoptosis, thereby affecting cell growth. So, the related MMP was measured to further investigate the influence of the explored photocatalytic reaction in hypoxic 4T1 cells. The commercial MMP probe JC-1 was used to visually observe the apoptosis of mitochondria after treatment with PFP and optical illumination. Typically, JC-1 characterizes the state of MMP (ΔΨm) by its monomer or J-aggregate. The J-aggregate initially attaches to the outer membrane of the normal mitochondria and emits red fluorescence. However, the ΔΨm decreases once the NADH-dependent oxidative phosphorylation (OXPHOS) process is inhibited,⁴⁰ and the JC-1 changes to the monomer and shows green fluorescence, indicating the mitochondria goes into apoptosis (Figure 3b). As displayed in Figure 3c and Supporting Information Figure S8, 4T1 cells exhibited strong green fluorescence after being treated with PFP upon optical illumination in the hypoxic environment, suggesting that the ΔΨm decreased and the mitochondria may be undergoing apoptosis. The variations of the intracellular ATP level were subsequently measured to confirm the mitochondria damage. As shown in Figure 3d, the ATP level of 4T1 cells in the control group had negligible changes no matter with or without white-light illumination. However, the ATP level of the white-irradiation group decreased to 0.11 μg/mg protein (from 0.20 μg/mg protein), which indicates that the process of NAD⁺ cleavage and product production led to mitochondrial damage, further affecting the normal metabolic process and finally causing cell apoptosis. To visually show the therapeutic effect of the PFP system on 4T1 cancer cells, LIVE/DEAD cell staining was performed accordingly. As shown in Figure 3e and Supporting Information Figure S9, almost all the cells treated with PFP in the dark emitted strong green signals of living cells, indicating a high cell-survival rate and the promising biosafety of PFP, whereas over half of the cells treated with PFP under illumination showed the red signal indicative of dead cells, signifying that PFP had an obvious therapeutic effect on the life activities of tumor cells under white-light illumination. The MTT assay was performed to evaluate the cell viability in the presence of different concentrations of PFP. 4T1 cells had high cell viability even treated with a high concentration of PFP (16 μM) in dark conditions (Figure 3f), further confirming its biocompatibility. However, upon white-light illumination, the 4T1 cell viability apparently decreased as the concentration of PFP increased (over 2 μM). 4T1 cell viability fell below 60% after being treated with 16 μM PFP, reinforcing the good therapeutic efficacy of the artificial photocatalytic system due to the catalytic decomposition of NAD⁺ by this system that results in a disordered metabolic process. Electron sacrificing agent TEOA and the nicotinamide involved in the system also exhibited negligible toxicity ( Supporting Information Figures S10 and S11). As illustrated in Figure 3g and Supporting Information Figure S12, the apoptosis cell ratio of 4T1 cells after being treated with PFP under white-light illumination increased significantly from 0.5% to 42.4%, whereas the control groups remained negligibly or slightly changed, including the white-light illumination-only group and PFP-only group. This indicates the good therapeutic efficacy of the artificial photocatalytic system as well as the noninvasiveness and biocompatibility of optical operation manner and PFP photocatalyst, respectively. To study the universality of the described photocatalytic system, two other tumor cell lines, HeLa and HepG2, were characterized by MTT assay ( Supporting Information Figure S13). With increasing concentration of PFP (over 1 μM), the decreasing cell viability of HeLa and HepG2 demonstrate that the photocatalytic system was effective for various cancer cells, including human cancer cells. Compared to the common strategy of regulation between NAD⁺ and active NADH, the metabolism regulation catalyzed by PFP is an irreversible reaction, which can disturb the cellular redox cycle and promote irreversible damage to mitochondria and even cells. The cytotoxicity of the photocatalytic reaction to human foreskin fibroblasts (HFF) cells was measured to study the effect of this photocatalytic reaction on normal cells ( Supporting Information Figure S14). The HFF cells maintained high viability under light and dark conditions.

• Download figure
• Download PowerPoint

The radical and biological action characterizations of the photocatalysis reaction

EPR assay was conducted to reveal the specific mechanism of the above photocatalytic process. Because the transient radical would be trapped by DMPO and changed to the EPR-active species DMPO-ADP-ribosyl•, the collected carbon-free radical signal under irradiation was recognized as ADP-ribosyl• radical (Figure 4a). It demonstrates that the NAD⁺ transferred to ADP-ribosyl• radical in this photocatalysis process under illumination. Regarding the discussion of Figure 1d, the electron that NAD⁺ captured was from PFP*. Based on the above experimental results, a presumptive mechanism of photocatalytic reaction by PFP is proposed (Figure 4b). Under illumination, electrons from PFP* transfer to NAD⁺, generating ADP-ribose• radical and nicotinamide; then the holes are captured by TEOA. After electron rearrangement, ADP-ribose is obtained from the NAD^+· by the combination of water. To test the irreversibility of this system in vivo, the enzymatic activity of production was characterized by monitoring the catalytic reaction of pyruvic acid with lactic dehydrogenase (LDH). Generally, LDH catalyzes the conversion of pyruvic acid to lactic acid with the oxidation of NADH, and the reaction should be inhibited when NADH is replaced by photocatalytic products. The enzymatic activity of production was characterized by monitoring the absorption at 340 nm, which is the characteristic absorption of NADH. As shown in Figure 4c, the absorption at 340 nm of photocatalytic products exhibited negligible change with the addition of the pyruvic acid and LDH, demonstrating that the photocatalytic product could not be used as a coenzyme of LDH. For active NADH, the absorption decreased significantly after adding the pyruvic acid and LDH. Thus, photocatalytic products are unable to participate in the normal enzymatic reactions, which proves that the products obtained in the photocatalytic reaction can not re-enter the enzymatic cycle of the cells. Therefore, the photocatalytic reaction interferes with the function of cancer cells and induces hypoxic tumor apoptosis.

• Download figure
• Download PowerPoint

Conclusions

In summary, we report a new photocatalytic reaction where irradiated PFP irreversibly converts NAD⁺ to nicotinamide and ADP-ribose. The chemical reaction was used to promote apoptosis in hypoxic tumor cells of 4T1. The cell experiments demonstrated that the intracellular NAD⁺ is converted into nicotinamide and ADP-ribose instead of naturally bioactive NADH, which is different from the original enzymatic pathway. As a competitive reaction of the cellular redox cycle, the continuous depletion of NAD⁺ disrupted the balance of coenzyme recycling. In addition, the disturbance decreased the MMP of mitochondria under hypoxic conditions, so the photocatalytic reaction in cells results in compromised mitochondria and the mitochondrial electron respiratory chain, eventually promoting apoptosis. Our approach utilizes an oxygen-independent photocatalytic reaction that may circumvent the issues of hypoxia and is different from conventional photodynamic therapy. The proposed photocatalytic reaction mechanism involves the photoinduced electron obtained from PFP* could transfer to NAD⁺, so the NAD⁺ can efficiently generate the ADP-ribosyl• radical and nicotinamide. Finally the ADP-ribosyl• radical is converted to ADP-ribose. This work provides a proof of concept that applying conjugated polymers in photocatalytic reactions could be an effective approach to regulating biofunctions for treating diseases. Designing customized chemical reactions would further pave the way to explore some more advanced cancer treatments and cross the frontiers between photochemistry and biomedical application.

Supporting Information

Supporting Information is available and includes experimental procedures, additional microscopic images, and Figures S1–S14.

Conflict of Interest

The authors declare no competing financial interests.