The current investigation aimed to offer insight into the underlying sources of metabolism of routine tetrazolium and resazurin agents used in cytotoxicity and proliferation characterization. We used B[a]P as an inducer of the PPP42 and 6AN as an inhibitor of G6PDH43,44 to interrogate the contributory metabolism of the PPP on routine cytotoxicity methods in Beas-2B cells. Present results demonstrated that the enhanced cellular metabolism of routine tetrazolium salts and Alamar Blue upon treatment with B[a]P did not irrevocably implicate either “mitochondrial activity” or “proliferation” – interpretations that remain pervasive in the in vitro toxicology literature45–47 – but rather an aggregate of biological processes, including PPP-derived NADPH (Fig. 6). While this observation has been noted previously for the PPP21, no investigation has examined differential sensitivity of methods presented to a specific biological process in such detail, as well as attempt to examine the underlying modulators of G6PDH-associated metabolism.
The general practice with toxicological in vitro assessment necessitates the confirmatory testing of cytotoxicity/proliferation endpoints using at least one secondary method of a differing mechanistic class, i.e., confirmation of cytotoxicity measured via tetrazolium salt method using a membrane permeability assay. For a given laboratory, LDH typically serves as the confirmatory test for cytotoxicity, while several methods for assessing DNA synthesis, e.g., EdU or BrdU, etc., may be employed to verify proliferation. However, two issues arise during confirmatory testing: 1) appropriateness of specific cytotoxicity assessment employed in the model system and 2) ultimate interpretation of biological effect when inter-assay discordance arises. To the former point, inadequate assessment of assay-specific technical interference for one or both employed method(s) undermines the validity of cytotoxicity/proliferative interpretation48, especially with nanomaterials that are particularly sensitive to interference49. While this interference is pervasive for nanomaterials50, small molecule interference is also common16,51. In pursuant of discordant results, reliance upon a limited selection of confirmatory cytotoxicity assays remains pervasive and relies generally on measurement of LDH release or ATP-derived endpoints52–56. Compounding this issue, significant gaps in knowledge remain about mechanisms of resazurin and tetrazolium metabolism, even though several groups have contributed significantly to the understanding of these endpoints57–60. Initially, our laboratory had observed robust hyperstimulatory WST1 metabolism during cytotoxicity screening of incinerated thermoplastic samples61 similar to those observed with organic extracts of diesel exhaust particles62. Zhu, et al. 63 observed that micromolar B[a]P concentrations induced hyperstimulatory MTT metabolism similar to the current investigation, thus prompting this examination.
Initial observations of WST1 hyperstimulatory tetrazolium metabolism with B[a]P treatment, in the absence of extracellular LDH leakage or PI uptake, suggest enhanced proliferation. Using another tetrazolium salt or resazurin failed to recapitulate the magnitude of metabolic induction measured with WST1 (Fig. 1). Conflicts between cytotoxicity measures for metabolic endpoints, such as tetrazolium salts, and cell-by-cell methods, such as nuclear propidium iodide staining, are known for cytotoxic agents64, thus multiple additional measures of proliferation and cytotoxicity were performed. Immunodetection analysis, doubling time, and cell cycle immunofluorescence (Fig. 2) negated the interpretation of the observed hyperstimulatory metabolism as purely proliferation, as suggested by the significant reduction in the rate of passage through S-phase65. Coupled with elevated (p)-Chk1(S345) and a (p)-p53-to-total p53 ratio via immunodetection analysis and an absence of stark changes in G2 or M-phase prevalence (Figs. 1 and 2), these results indicated transient induction of the inter-S checkpoint due to B[a]P, likely through ATM/ATR axis 66. The results confirmed observations from similarly designed studies in which B[a]P increased the prevalence of S-phase in this cell line41,63. With insignificant changes in G2 prevalence or cdc2 phosphorylation at the tyrosine-15 residue, the results are inconsistent with induction of the G2/M checkpoint67,68. CH223191, a competitive inhibitor of AhR, alone failed to block reductions in EdU uptake, which could be explained as either 1) basal metabolic machinery in Beas-2B cells produced enough genotoxic metabolite of B[a]P69 to initiate inter-S checkpoint or 2) B[a]P and/or its metabolites can activate other nuclear receptors involved in CYP induction and perpetuation of genotoxic stress70. While 6AN inhibited B[a]P-induced effects on EdU uptake, co-treatment led to a paradoxical reduction in S-phase prevalence compared to 6AN treatment-only controls and was larger in magnitude than the effects observed by G6PDH siRNA #2. Together, these results indicate that the activation of the inter-S checkpoint in Beas-2B cells impinged upon an intact CYP450 machinery which included not only CYP metabolism but also AhR and NADPH. Further research is required to attribute the relative contribution of each portion of the system.
AGK2 partially attenuated B[a]P effects on EdU uptake, while inhibition of CYP1/AhR by αNF, G6PDH catalytic activity by 6AN, and G6PDH siRNA #2, categorically reversed B[a]P-induced cell cycle changes. EX527 failed to alter the response, as expected. Since AGK2 increases G6PDH catalytic activity post-translationally via SIRT240,41, blocking SIRT2 did indeed partially reverse hyperstimulation indicating SIRT2 and, thus, likely plays a role in the B[a]P bioactivation response (Fig. 5). Since EX527 is not associated with post-translational modification of G6PDH, its inability to modulate the B[a]P response was expected. Other known cytosolic sources of NADPH, such as isocitrate dehydrogenase-1 (IDH1), malic enzyme-1 (ME1), and other compartmental stores22–24,71, could not generate NADPH in sufficient concentrations to sustain microsomal metabolism for B[a]P. Ultimately, the basal microsomal metabolic machinery and PPP-derived cytosolic NADPH concentrations in Beas-2B cells were sufficient to initiate and maintain genotoxic stress by B[a]P metabolites. G6PDH catalytic enhancement by SIRT2 likely initiated a feed-forward loop but does not seem to account for the entire regulatory mechanism of G6PDH; the exact contribution of SIRT2 under basal conditions are less clear but warrants further investigation. Furthermore, a myriad of other regulators of G6PDH are known, including ATM72, whose contributions in this system were not explored, and may explain only partial attenuation by AGK2. Examination of proliferative capacity discounted an interpretation of proliferation as the source of hyperstimulatory metabolism observed in MTS and WST1 assays.
To establish a link between resazurin/tetrazolium metabolism and their respective dependence upon the PPP, we examined the contributory effects of proliferation, cytotoxicity, and assay interference to isolate the effect of the PPP on each of the endpoints tested. Here, we demonstrated that B[a]P did not interfere with the assay components, analyte chromophore, or detection of wavelength transmittance (optical activity) of the endpoints examined, thus they were non-contributory to the observed elevation in optical density (Supplementary Fig. 9). The nano- to low-micromolar concentrations of B[a]P did not induce significant lytic cytotoxicity in concordance with previous studies using Beas-2B cells63,73−75. Examination of high-content imaging (Fig. 2; Supplementary Fig. 10 and Supplementary Fig. 11) used for cell cycle analysis did not suggest substantial apoptosis induction at frequencies enough to significantly alter the interpretations drawn, thus underscoring the major effect of B[a]P on alterations in proliferation rate. 6AN dose-dependently reduced metabolism of the measured endpoints listed in order of magnitude: WST1 > MTS > MTT > Alamar Blue without overt cytotoxicity. The hierarchy of metabolic inhibition was linearly correlated with B[a]P-induced metabolic induction of each endpoint, thus providing correlative evidence of endpoint-specific sensitivity to the metabolic status of G6PDH. This correlation was supported by co-incubation with 6AN (Fig. 4H), further supporting each endpoint’s sensitivity to PPP-derived NADPH. These results are not perplexing as several studies implicate NADPH in the metabolism of WST176, MTS4, and MTT6, though strict comparisons between each method have not been made. The contribution of water-soluble tetrazolium reliance may be due to differential reduction potentials among the two-electron cycling substrates used to enhance tetrazolium development (5-methylphenazium methyl sulfate for MTS and 1-methoxy-5-methylphenazium methyl sulfate for WST1). Further testing would be required to confirm differential reduction potentials among these cycling substrates. Nevertheless, at the most extreme, WST1 was almost exclusively metabolized by PPP-derived NADPH, followed by MTS, and then MTT; Alamar Blue was comparatively insensitive to NAPDH-derived sources in this model. Overall, CYP450 translational control and G6PDH competitive inhibition proved universally effective in abrogating B[a]P-mediated metabolic induction. These results demonstrated that, while the PPP was the source of reducing equivalents responsible for metabolic induction, an intact metabolic machinery, i.e., CYP450, of a defined enzymatic capacity was also required to precipitate a state of enriched PPP-derived NAPDH that could be readily reduced by resazurin and tetrazolium protochromophores.
We observed significant ΔΨm depolarization in B[a]P-treated cells compared to controls (Fig. 1). Generally, JC-1-reported ΔΨm depolarization is positively correlated with markers of cytotoxicity in vitro77–80; however, this was not presumed in the current investigation. The observed peak depolarization occurred between 1 and 10 µM B[a]P (48 hours exposure) – the dose range associated with maximal hyperstimulatory metabolism of all tested tetrazolium salt methods. If tetrazolium reduction was correlated with mitochondrial function, as remains contemporaneously purported45–47, then one would expect significant reductions in tetrazolium metabolism in association with ΔΨm depolarization. To this point, the authors purport the changes in ΔΨm, including 6AN-mediated apparent attenuation of ΔΨm depolarization by B[a]P, directly support an active B[a]P-induced metabolic shift from glycolysis to PPP as has been previously observed in RT4 bladder cancer cells using an untargeted metabolomic approach42. Results of the current investigation suggest redirection of glucose equivalents from glycolysis to the PPP to support NADPH-dependent microsomal systems by B[a]P. Additionally, fructose-6-phosphate (F-6-P), an intermediate of the PPP, typically undergoes further glycolytic metabolism to eventually enter the TCA cycle to support retention of ΔΨm. However, F-6-P can be metabolized back to G-6-P for re-entry into the PPP. Given the ΔG° for conversion of G-6-P to F-6-P is slightly positive81, the reverse reaction can become thermodynamically favorable as to become spontaneous, especially if G-6-P concentrations are kept sufficiently low through rapid tonic metabolism by G6PDH. This is the most plausible explanation for the metabolic data in the current investigation, especially since B[a]P-induced WST1 metabolism can be explained sufficiently by Michaelis-Menten kinetics. This suggests a single rate-limiting step that constrained the metabolic induction rate in the present system. Further studies would be required to confirm these assertions as well as to identify this presumed rate-limiting step in the observed process. The present results in this model offer more evidence to support previous assertions that mitochondrial function and tetrazolium metabolism can be uncoupled in viable cells6,15,82.
Based on the presented Beas-2B cell proliferation determinations and resazurin/tetrazolium salt metabolic data, B[a]P induces a metabolic shift from glycolysis to the PPP. This shift was detected most sensitively via water-soluble tetrazolium methods (WST1 and MTS), while MTT and Alamar Blue showed minimal metabolism induction in this system. Considering only the observed hyperstimulatory WST1 metabolism and lack of B[a]P-induced LDH release without context could erroneously lead to a conclusion of enhanced cell proliferation. However, the present results demonstrate that the magnitude of B[a]P-induced assay metabolism is strictly associated with sensitivity to PPP inhibition, thus warranting preliminary validation in method utilization for agents which may perturb PPP metabolism. Without complete examination of the nuanced method-specific heterogeneities among routine tetrazolium/ resazurin methods utilized in toxicity testing, qualifying results of specified cytotoxicity measures under test conditions become untenable. For example, a choice confirmatory test of LDH would have failed in recognizing the reduced proliferative capacity of B[a]P-treated Beas-2B cells. Therefore, delineating and weighing discrepancies between screening and confirmatory testing becomes equally untenable, leading to uncertainties in rationalizing cytotoxic outcomes. As shown, investigations examining cytotoxic and proliferative changes due to AhR ligands, such as PAHs, should consider the potentially confounding contribution of metabolic reprogramming. Further, methods for even routine cytotoxicity detection which rely on NADPH-mediated reduction should be preliminarily validated for potential interactions from PPP metabolism prior to applications under test conditions.