Isoform-specific NADPH oxidase inhibition for pharmacological target validation

Unphysiological reactive oxygen species (ROS) formation is considered an important pathomechanism for several disease phenotypes with high unmet medical need. Therapeutically, antioxidants have failed multiple times. Instead, focusing on only disease-relevant, enzymatic sources of ROS appears to be a more promising and highly validated approach. Here the family of five NADPH oxidases (NOX) stands out as drug targets. Validation has been restricted, however, mainly to genetically modified rodents and is lacking in other species including human. It is thus unclear whether the different NOX isoforms are sufficiently distinct to allow selective pharmacological modulation. Here we show for five of the most advanced NOX inhibitors that indeed isoform selectivity can be achieved. NOX1 was most potently (IC50) targeted by ML171 (0.1 μM); NOX2, by VAS2870 (0.7 μM); NOX4, by M13 (0.01 μM) and NOX5, by ML090 (0.01 μM). Conditions need to be carefully controlled though as previously unrecognized non-specific antioxidant and assay artefacts may limit the interpretation of data and this included, surprisingly, one of the most advanced NOX inhibitors, GKT136901. As proof-of-principle that now also pharmacological and non-rodent target validation of different NOX isoforms is possible, we used a human blood-brain barrier model and NOX inhibitor panel at IC50 concentrations. The protective efficacy pattern of this panel confirmed the predominant role of NOX4 in stroke from previous genetic models. Our findings strongly encourage further lead optimization efforts for isoform-selective NOX inhibitors and clinical development and provide an experimental alternative when genetic validation of a NOX isoform is not an option. Graphical abstract


Introduction
Reactive oxygen species (ROS) are considered an important pathomechanism for several common diseases with high unmet medical need including cardiovascular and neurodegenerative diseases as well as cancer. Yet, clinical translation of this hypothesis by therapeutic or even prophylactic use of antioxidants that scavenge ROS has consistently failed [1], in some cases even increasing mortality [2,3] [4].This paradox was initially explained by these compounds being possibly underdosed, thereby not reaching efficacy, or over being overdosed causing reductive stress.
It is now understood, however, that ROS are by no means only harmful metabolic byproducts, but also serve important protective, metabolic and signaling functions, such as the regulation of cell proliferation, differentiation, migration and survival, innate immune response, vascular tone, neuronal signaling as well as inflammation [5][6][7][8]. Anti-oxidants will always simultaneously interfere with both qualities of ROS, the physiological and pathophysiological ones, and the overall neutral or deleterious outcomes of antioxidant studies likely show that the predominant quality of ROS is protective. Thus, ROS should not be interfered with in a systemic manner, but rather by identifying for each disease condition the relevant ROS source and inhibit by classical pharmacology, i.e. enzyme inhibition [9].
Target validation of these NOX isoforms in different disease models has been done primarily by gene knock-out (as reviewed in [22]) and, in the case of NOX5 [18], by knock-in technology limiting most data sets to mice and, in the case of NOX4, in addition to rat [13]. For ultimate clinical translation and target validation in other species, NOX-specific and ideally isoform-selective NOX inhibitors are desirable. It is unclear, however, whether this is pharmacologically achievable given the fact that all five relevant NOX isoforms (NOX1, NOX2, NOX3, NOX4 and NOX5) are similarly structured transmembrane proteins containing conserved heme, FAD and NADPH binding sites. Some degree of functional variability derived from the fact that some isoforms have more (NOX1 and NOX2) or less (NOX4 and NOX5) binding partners or a unique calcium binding domain (NOX5) [23]. In addition, all ROS based activity assays are indirect opening the possibility that any compound exerts some of its apparent effects on ROS formation by direct ROS scavenging [24,25]) or direct assay interference [26,27].
Here, we examine some of the best characterized or most widely used NOX inhibitors, VAS2870, ML171, GKT136901, M13 and ML090. We first analysed the isoform selectivities, then possible assay interference and direct ROS-scavenging capacity, and finally suggest a NOX inhibitor panel for pharmacological target validation, which we then validate an in vitro human model of brain ischemia to identify NOX4 as a therapeutic target. transfection reagent (Promega) was used in a ratio 4:1 with pcDNA3.1 DNA plasmid containing the sequence for either NOX 1, NOXO1, NOXA1 (triple transfection), NOX4 or NOX5. After removal of DMEM medium 240 µl of serum free medium was added to polystyrene tube followed by 10 µl of FuGENE directly into the medium and incubated for 5 min. at room temperature. Thereafter pre-mixed DNA was added to the tube and incubated for 30 min. at room temperature. 250 µl of the mixture was added to the HEK293 cells. Selection was started after three days by changing to DMEM medium containing 200 µg/ml G418. The medium was changed every two to three days before a single clone has been selected. After trypsinisation and dilution a cell suspension of 27 cells/µl was used to prepare a single cell suspension of 1 µl pre-dilution in 12 ml HEK selection medium. Finally, 500 µl of single cell suspension per well seeded in 24-well plate.

Material and Methods
In some experiments, HEK293 cells were transiently transfected with NOX4 or NOX5 or were triple transfected with NOX1, NOXO1 and NOXA1 using FuGENE6 followed by ROS measurement after 48h. As described previously [28] in HL-60 cells was measured by cytochrome C reduction as described previously [28]. Briefly, cytochrome C (100 µM) was added to the DMSO-differentiated HL-60 cell suspension. After Statistical differences between means were analysed by one-way ANOVA followed by Bonferroni correction for multiple comparisons. A value of P < 0.05 was considered statistically significant.

NOX pharmacological isoform selectivity
To characterize the NOX isoform selectivity of the current second generation NOX inhibitors we determined concentration-dependency and efficacy of GKT136901, Ml171, VAS2870, M13 and ML090 on NOX1, 2, 4, and 5 (Fig. 1A). Specifically, NOX2 expressing HL-60 cells or HEK293 cells transfected with NOX1, NOX4 or NOX5 were incubated in presence of increasing concentrations of each compound. ROS generation was subsequently induced by phorbol 12-myristate 13-acetate (PMA) to stimulate NOX1, NOX2 and NOX5, which was assayed in presence of ionomycin. Thereafter, ROS production was assayed using a panel of cellular assays with structurally unrelated probes; i.e. Amplex red, luminol or cytochrome C ( Figure 1B). Amplex red was used to quantify extracellular H 2 O 2 [29], O 2 •generation was detected by luminol-based chemiluminescence and cytochrome C-reduction was quantified to assess NOX2 activity.
The same holds true for ML171 (Fig. 2B) which is more selective for NOX1 compared to NOX4 and NOX5. VAS2870 (Fig. 2C) displayed NOX2 over NOX1 and NOX4 selective inhibition and also slightly inhibited NOX5. The GlucoxBiotech compound M13 (Fig. 2D) showed almost selective NOX4 inhibition but NOX2 (with low E max ) and NOX1 (at high concentrations) inhibition was observed as well. Finally, Ml090 (Fig. 2E) inhibited NOX5, NOX1 and NOX4 with comparable IC 50 but enhanced E max for NOX5 inhibition. These data suggest that NOX inhibitors indeed display differential isoform targeting and therefore should be considered ideally by combined analysis of an inhibitor panel and ranking potencies.

Non-specific anti-oxidant and assay artefacts
To screen for possible assay interference between GKT136901, ML171, VAS2870, M13 or ML090 and the NOX1 assay, a cell-free system was performed in which each inhibitor was screened in presence of luminol, 1mU/ml xanthine oxidase (XO) and 1mg xanthine (X) generating O 2 •-. At concentrations 1µM GKT136901 did not show interference with either the molecular probe or with the X/XO system (Fig. 3A). Similar findings were obtained with 0.1 µM ML171, 10 µM VAS2870 and 30 nM ML090 (Fig. 3A). In contrast, 1 µM M13 enhanced chemiluminescence (Fig.   3A). To identify direct interactions between assay components and GKT136901, ML171, VAS2870, M13 or ML090 a cell-free counter screen was performed without X/XO-generated ROS.
Indeed, ML171 and VAS2870 showed reduction of the luminol-based signal suggesting direct interference with luminol-based chemiluminescence (Fig. 3B). In contrast, in presence of ML090 the signal was enhanced (Fig. 3B).
To study non-specific antioxidant effects of VAS2870, the effect of 10 µM VAS2870 was studied in a cell free system in presence of cytochrome C, and X/X0-derived ROS.
To study whether the effects of GKT136901, ML171, VAS2870, M13 and ML090 on NOX4 and NOX5 are specific, a cell-free assay was performed with NOX4 or NOX5 effective concentrations of each inhibitor in presence of 0.25µM H 2 0 2 in an Amplex red assay. ML171, VAS2870, M13 and ML090 did not directly affect Amplex red-based detection of H 2 O 2 (Fig. 3D).
Hence, the assay was repeated without H 2 0 2 to study potential direct assay component interference.
As predicted from the inhibitor screen, M13, GKT136901 or ML171 protect against ischemia-induced hyperpermeability in a human brain ischemia model. Our data suggested that a NOX inhibitor panel may be used for target validation of specific NOX isoforms. We thus tested our NOX inhibitor panel in an in vitro human model of brain ischemia (Fig. 4A) where NOX4 is involved in subacute hypoxia-induced increases in cell permeability whereas NOX1, NOX2 do not [30] and NOX5 only acutely (Casas et al., unpublished observation (0.01 µM; mainly targeting NOX5). Hypoxia increased cell permeability after 24hrs of reoxygenation. GKT136901, and ML171 treatment prevented this detrimental effect (Fig. 4C), suggesting protection against hyperpermeability via NOX4 inhibition while, as expected given their respective IC 50 values, VAS2870, M13 or MI090 treatment showed no effect (Fig. 4C). These data provided proof-of-concept for pharmacological target validation using NOX inhibitor panels.

Discussion
Our results clearly show that using NOX inhibitor panels for isoform-selective pharmacological target validation is feasible. In this study, we provide the isoform (NOX 1, 2, 4 and 5) preferences as IC 50 values and maximal efficacies of the currently five most advanced NOX inhibitors, GKT136901, ML171, VAS2870, M13 and ML090. We also apply some of the most commonly used ROS assays and detect for GKT136901 and ML171 relevant, non-specific antioxidant effects and assay artefacts by either interacting with the assay or directly with ROS in concentrations similar to the IC 50 values. Highest potencies were observed for GKT136901 and ML171 with NOX1 over NOX 4 and 5, VAS2870 with NOX2 over NOX1, 4 and 5, M13 with NOX4 and ML090 with NOX5 over NOX1, 4 and 5 (table 1). None of these small molecule inhibitors are isoform selective but inhibited several NOX isoforms with different ranking potencies.
In detail, VAS2870 has been claimed a pan NOX inhibitor [31] because NOX1, 2, 4 and 5 [23,30] [32,33] activity was inhibited but only IC 50 values for NOX2 were published [33,34]. We showed that VAS2870 is highly potent for NOX2 (IC 50 ~ 0.7µM), followed by IC 50 values in low micromolar range for NOX1 and 4 and higher micromolar range for NOX5 although E max was higher for NOX5 than NOX4 verifying that VAS2870 is not isoform selective but inhibits all four NOX isoforms. In addition, we found that M13 is a first-in-class NOX  While VAS2870, M13, Ml090 did not show assay interferences, the clinically most advanced compound, GKT136901, claimed to be NOX1 and 4 specific, showed direct assay interference with Amplex Red (NOX4 and 5 assay) and luminol (NOX1 assay) and ROSscavenging effects. The latter has been verified in the alternative H 2 O 2 measuring HVA assay (tested from >10µM) [36]. Likewise, non-specific antioxidant effects of GKT136901 have been published earlier such as inhibition of xanthine oxidase at high concentrations (100 µM) [37] or peroxynitrite scavenging [25]. In addition, it could also act like a peroxidase inhibitor, thus limiting application of common used ROS assays. This may thus complicate the interpretation of results regarding the contribution of NOX enzymes in disease models, questioning the proposed isoform selective inhibition of GKT136901. In this context, the 2-acetylphenothiazine ML171 which has been published to be NOX1 specific and has xanthine oxidase activity [38], showed assay interferences with i) Amplex Red in similar NOX1 effective concentrations, with ii) the HVA assay (tested >10µM) and interfered with the iii) luminol NOX1 assay. These data are in line with a previous publication that discovered ML171 to be inactive on all NOX isoforms but interact with the Amplex red assay by either an intrinsic antioxidant activity or inhibition of horseradish peroxidase [39]; the latter would support previous observations that phenothiazines are peroxidase substrates [40]. Facing different assay problems that may limit interpretations, we suggest here to use a NOX inhibitor panel at IC 50 concentrations for isoform-specific target validation. Importantly, we could validate this approach by GKT136901, ML171 and M13 being neuroprotective in a manner that suggested NOX4 as responsible isoform as had previously been validated genetically [13].
In summary, all tested NOX inhibitor compounds displayed different isoform profiles suggesting differential chemical targeting. This knowledge will be valuable to advance further lead optimization to allow single compound target validation with potential for progress towards specific therapeutic development. We also provide an immediately applicable inhibitor panel approach that allows target validation of NOXs under conditions where gene knock-out or knock-in is not feasible or desirable.    inhibited ROS production with NOX isoform specific IC 50 . Data are presented as the mean±SEM;

Tables
All ROS measurements were performed in triplicate in at least three independent experiments.

Figure 3: Several inhibitors display inhibit ROS assays. A/B) GKT136901 and VAS2870
inhibit luminol-based assays. To study possible interference of GKT136901, ML171, VAS2870, M13 or ML090 with luminol-based measurements cell-free luminol assays were used. In these assays, ROS production generated by X/XO was enhanced by M13 (A) but not by GKT136901, ML171, VAS2870 or ML090. Moreover, chemiluminescence produced by luminol only was inhibited by ML171 and VAS2870, enhanced by ML090 and not affected by GKT136901 or M13, respectively (B). C) Possible assay interference by NOX2-specific VAS2870 was assessed by studying cytochrome C reduction in presence of X/XO-derived ROS in a cell free system. Presence of VAS2870 slightly, but significantly reduced chemiluminescence. D/E) The possibility of interference between GKT136901, ML171, VAS2870, M13 or ML090 and Amplex Red-based assays was studied using cell-free Amplex Red assays. In these assays, ROS production generated by H 2 O 2 was inhibited by GKT136901 (D) but not by ML171, VAS2870, M13 or ML090 (D).
Moreover, fluorescence produced by Amplex Red only was inhibited by GKT136901, ML171, M13 and ML090 while it was enhanced by VAS2870 (E). Data are presented as the mean±SEM; All ROS measurements were performed in triplicate in at least three independent experiments.