Latest Advances in Imaging Oxidative Stress in Cancer

Probes for the Selective Imaging of Individual Reactive Species

In addition to assaying oxidative activity in cells, fluorescent probes have been developed for species-specific ROS detection, including superoxide (dihydroethidium), hydrogen peroxide (2′,7′-dichlorofluorescein), and singlet oxygen (trans-1-(2′-methoxyvinyl)pyrene). Most probes are not truly specific for individual reactive species; rather, they exhibit enhanced selectivity for different ROS. These fluorescent probes rely on varied mechanisms of action to generate contrast. For example, Amplex Red (Thermo Fisher Scientific) is selectively oxidized by hydrogen peroxide in a reaction mediated by horseradish peroxidase. 2′,7′-dichlorofluorescein (H₂DCF) is an alternative dye used for the quantitation of intracellular hydrogen peroxide. In the diacetate form of 2′,7′-dichlorofluorescein, the nonfluorescent probe passively diffuses through the cell membrane, where it is cleaved by esterases to 2′,7′-dichlorofluorescein, resulting in intracellular trapping. 2′,7′-dichlorofluorescein is then oxidized by hydrogen peroxide to produce 2′,7′-dichlorodihydrofluorescein, which is highly fluorescent. Mitochondrion-specific superoxides can also be visualized by MitoSOX Red (ThermoFisher Scientific), a cationic derivative of dihydroethidium that is electrophoretically taken up into actively respiring mitochondria and fluoresces after its oxidation and subsequent binding to DNA.

Imaging of Glutathione

As the most abundant thiol-containing antioxidant, glutathione is a surrogate marker of cellular antioxidant capacity. The fluorescent dyes monobromobimane and monochlorobimane readily react with low-molecular-weight thiols, including glutathione, and in doing so form fluorescent adducts. An additional thiol-tracking dye is ThiolTracker Violet (Thermo Fisher Scientific), which is also retained intracellularly through adduct formation and whose fluorescent signal is 10 times greater than that of bimane compounds (11).

An important consideration is that optical imaging is constrained by overlying tissue both absorbing and scattering the exciting or emitted light. Fluorescence is therefore better suited for cell-based imaging and superficial or intraoperative small-animal preclinical work than for translational applications.

PET Imaging of ROS

The successful application of fluorescent probes to ROS and antioxidant imaging has resulted in the adaptation of these small molecules to PET, often through the incorporation of ¹⁸F. Turn-on mechanisms after radiotracer oxidation, however, cannot be used for the generation of contrast by PET, and alternative methods of intracellular trapping are required. Chu et al. demonstrated the advantages of radiolabeling the fluorescent dye dihydroethidium with ¹⁸F and showed its ability to measure superoxide production after treatment with doxorubicin in cells grown in culture. After its oxidation, ¹⁸F-labeled dihydroethidium becomes charged and can intercalate DNA, intracellularly trapping the tracer (13). Other fluorescent scaffolds, such as hydrocyanines (14), have also been labeled with ¹⁸F as a method to image oxidative stress in vivo. In addition to ROS-sensing fluorophores, chemiluminescent probes based on luminol have been used for ROS detection. Recently, a gallium-labeled luminol derivative (galuminox) was shown to selectively accumulate in the mitochondria of tumor cells after ROS induction, with ⁶⁸Ga-galuminox selectively retained in a model of lung inflammation (15). A radiolabeled ascorbate derivative, ¹⁸F-KS1, is also in the early stages of development for ROS imaging (16).

Imaging the Tumor Antioxidant Response

Given the short-lived nature of ROS, imaging of the durable downstream consequences of this toxic insult may provide a larger detection window with PET. The transmembrane protein system x_C⁻ is a heterodimeric transporter that is placed centrally within the cell’s antioxidant system. The role of system x_C⁻ is to exchange the intracellular amino acid glutamate for the extracellular amino acid cystine. After cystine’s uptake, it is rapidly reduced to cysteine, the rate-limiting precursor for glutathione biosynthesis, placing system x_C⁻ as a central regulator of antioxidant homeostasis (17). Elevated system x_C⁻ activity has been exploited by PET imaging tracers such as (4S)-4-(3-¹⁸F-fluoropropyl)-l-glutamate (¹⁸F-FSPG) (18), ¹⁸F-5-fluoro-aminosuberic acid (19), and ¹⁸F-hGTS13 (20). Tumor retention of ¹⁸F-FSPG is redox-sensitive, mediated by the concentration gradient of cystine across the plasma membrane. In an animal modal of ovarian cancer, ¹⁸F-FSPG tumor retention decreased in proportion to the degree of oxidative stress induced by chemotherapy (Fig. 2A) (21).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

Imaging tumor redox status. (A) Changes in ¹⁸F-FSPG A2780 ovarian tumor retention after oxidizing doxorubicin HCl liposome (Doxil; ALZA Corp.) therapy. (Reprinted from (21).) (B and C) Sequential coronal T2-weighted images and corresponding ¹³C 3-dimensional MR spectroscopic images demonstrating distribution of hyperpolarized [1-¹³C]DHA and vitamin C in TRAMP (transgenic adenocarcinoma mouse prostate) mouse. Regions of liver, kidney, and prostate tumor are segmented and superimposed on spectral grid (color-coded dashed lines). (D) Representative ¹³C spectra from liver, kidney, and prostate tumor in TRAMP mouse. D0 = untreated; D1 = 24-h Doxil treatment; D6 = 6 d after initiation of Doxil treatment. VitC = vitamin C. (Reprinted from (32).)

A consequence of ROS-induced membrane peroxidation is the intracellular production of reactive aldehydes that if left unchecked result in catastrophic DNA damage. Many cancer cells upregulate aldehyde dehydrogenases in response to this oxidative stress, which mediates aldehyde detoxification (22). The enzymatic activity of aldehyde dehydrogenase 1A1 has recently been quantified with a novel substrate-based radiotracer (23). Using a complementary strategy, Kirby et al. developed ¹⁸F-NA₃BF₃ for the imaging of total aldehydic load through radiotracer–aldehyde complex formation (24). Together, these tracers may provide insight into oxidative stress–mediated lipid peroxidation during anticancer therapy.

Hyperpolarized Spectroscopic MRI

Dynamic nuclear polarization is an emerging technique that increases the sensitivity of MR experiments by more than 10,000 times, allowing dynamic imaging of administered ¹³C-labeled substrates and their metabolic products in vivo (28). Flux through the PPP has been estimated using this technique through the conversion of U-²H,U-¹³C-glucose to the PPP intermediate 6-phosphogluconate (29). PPP metabolic activity is upregulated in cancer, which generates NADPH to maintain the antioxidant capacity of cells. However, the short polarization lifetime of uniformly labeled glucose at relatively low levels (∼15%), along with overlapping ¹³C resonances of 6-phosphogluconate and 3‐phosphoglycerate (a glycolytic intermediate), is a challenge that currently restricts the widespread use of U-²H,U-¹³C-glucose. Alternatively, hyperpolarized [1-¹³C]dehydroascorbic acid ([1-¹³C]DHA), the oxidized form of ascorbic acid (vitamin C), has been used to probe tumor redox potential (30). After its uptake by the facilitative glucose transporters, hyperpolarized [1-¹³C]DHA was rapidly converted to [1-¹³C]vitamin C in lymphoma (30) and prostate tumors (Figs. 2B–2D) (32), the rate of which was determined by the levels of both glutathione and NADPH (31). Despite the promise of [1-¹³C]DHA to assess total tumor antioxidant capacity, administration of a 10 mg/kg dose of DHA to tumor-bearing mice resulted in transient respiratory arrest and cardiac depression (31). Optimization of dosing regimens and a greater understanding of DHA toxicity are therefore a prerequisite for clinical imaging with [1-¹³C]DHA.

An important consideration for all MRI-based redox probes is the requirement of high mass doses of contrast agent. Given that these agents are frequently either stable radicals or potent radical scavengers, redox-active MRI probes may also perturb the system that they are measuring, possibly accounting for DHA-induced toxicity.

Response Monitoring

In conjunction with DNA damage, chemotherapies and ionizing radiation produce high levels of oxidative stress in tumors, with cell death induced in those sensitive to treatment (6). Consequently, redox imaging agents have the potential to assess the efficacy of a wide range of therapies that converge with the induction of oxidative stress. In a recent proof-of-principle study, the tumor antioxidant response to doxorubicin was shown to be an earlier maker than both ¹⁸F-FDG and tumor volume (21). Furthermore, the imaging window for the measurement of tumor antioxidant response is not limited by a temporally unstable marker (e.g., cell death) (34). Additionally, several therapies have been developed whose primary mechanism of action is the induction of lipid ROS and concurrent membrane peroxidation (35). Redox imaging probes may therefore play an important role in monitoring the response to these novel agents.

Prediction of Drug Resistance

Elevated antioxidant capacity and the ability to buffer oxidative stress are a hallmark of drug-resistant cancer (36). A noninvasive measure of drug resistance will facilitate early intervention, allowing the clinician to adapt the treatment regimen, with the potential to improve patient outcomes. For widespread utility, the imaging biomarker ideally should be causal to drug resistance, be tumor-specific, result in a positive imaging signal, be generalizable to multiple drugs with different mechanisms of action, have expression that is independent of other factors or conditions, and require a single imaging scan. To date, ¹⁸F-FSPG imaging of system x_C⁻ activity has proven to be a good surrogate marker of drug resistance in animal models of ovarian cancer, reporting on the elevated glutathione found in these tumors (37). Further work, however, is needed to determine whether ¹⁸F-FSPG is a robust marker of drug resistance for multiple cancer types with discrete driver mutations.

Metastases

Tumor cells experience substantial oxidative stress when they detach from the extracellular matrix and enter the circulation. Anoikis, a form of programmed cell death after loss of anchorage, frequently follows intravasation and restricts the metastatic capabilities of tumor cells (38). The oxidative environment of the bloodstream further limits metastatic efficiency. In anoikis-resistant cells, PPP-generated NADPH mitigates the ROS that accompanies loss of attachment to permit cell survival (39). Suppressing oxidative stress by increasing endogenous and exogenous antioxidant availability in vivo further promotes metastasis in multiple models of cancer (1). Consequently, by imaging tumor antioxidant capacity before membrane detachment, it may be possible to determine the metastatic potential of primary tumors.