Unchanged NADPH oxidase activity in Nox1-Nox2-Nox4 triple ...

Page 1 of 23

Antioxidants & Redox Signaling Unchanged NADPH oxidase activity in Nox1-Nox2-Nox4 triple knockout mice – What do NADPH-stimulated chemiluminescence assays really detect? (doi: 10.1089/ars.2015.6314) This article has been peer-reviewed and accepted for publication, but has yet to undergo copyediting and proof correction. The final published version may differ from this proof.

1 News and Views

Unchanged NADPH oxidase activity in Nox1-Nox2-Nox4 triple knockout mice – What do NADPH-stimulated chemiluminescence assays really detect? Flávia Rezende1, Oliver Löwe1, Valeska Helfinger1, Kim-Kristin Prior1, Maria Walter1, Sven Zukunft2, Ingrid Fleming2, Norbert Weissmann3, Ralf P. Brandes1,*, Katrin Schröder1,* 1

Institute for Cardiovascular Physiology, Goethe-University, Frankfurt, Germany Institute for Vascular Signaling, Goethe-University, Frankfurt, Germany 3 Excellencecluster Cardiopulmonary System, Justus-Liebig-University Giessen, Germany 2

Short title: Nox membrane assays in Nox1/2/4 triple knockout mice

Word count: Total 4177, Abstract: 185

* corresponding and joined senior authors at Institut für Kardiovaskuläre Physiologie Fachbereich Medizin der Goethe-Universität Theodor-Stern Kai 7 60590 Frankfurt am Main, Germany Tel.: +49-69-6301-85321 Fax.: +49-69-6301-7668 Email: [email protected] or [email protected]

Key Words: NADPH oxidase, NO synthase, lucigenin, superoxide, reactive oxygen species

1

Page 2 of 23


2

Abstract (196 words) NADPH oxidases of the Nox family are considered important sources of cellular reactive oxygen species (ROS) production. This conclusion is in part based on the ability of NADPH to elicit a chemiluminescence signal in tissue/cell homogenates or membrane preparations in the presence of enhancers such as lucigenin, luminol or L012. However, the ability of these particular assays to specifically detect Nox activity and Nox-derived ROS has not been proven. We here demonstrate that combined knockout of the three main Nox enzymes of the mouse (Nox1-Nox2-Nox4 triple knockout) had no impact on NADPH-stimulated chemiluminescence signals in aorta, heart and kidney homogenates. In the NADPHstimulated membrane assays, no effect of in vivo AngiotensinII pretreatment or deletion of Nox enzymes was observed. In in vitro studies in HEK293 cells, the overexpression of Nox5 or Nox4 markedly increased ROS production in intact cells whereas overexpression of Nox5 or Nox4 had no influence on the signal in membrane assays. In contrast, overexpression of nitric oxide synthase or cytochrome P450 enzymes resulted in an increased chemiluminescence signal in isolated membranes. On the basis of these observations we propose the hypothesis that NADPH-stimulated chemiluminescence-based membrane assays, as currently used, do not reflect Nox activity.

2

Page 3 of 23


3

Innovation (83 Words) NADPH oxidases of the Nox family are considered important sources of ROS in mammalian tissue but the assays to determine their activity ex vivo have not been carefully validated. Here we demonstrate using newly generated Nox1-Nox2-Nox4 triple knockout mice that the most frequently used technique – enhanced chemiluminescence in membrane preparations does not detect Nox activity but rather other enzymes such as NO synthase or cytochrome P450 monoxygenase. Thus, this frequently used technique should no longer be applied to determine Nox NADPH oxidase activity.

3

Page 4 of 23


4

Main text (2074 words) NADPH oxidases of the Nox family are important sources of reactive oxygen species (ROS). Much evidence for this came from chemiluminescence assays that made use of the enhancers lucigenin (4), L012 or luminol / horseradish peroxide (HRP). In these, tissues or cells were homogenized and the chemiluminescence was measured after addition of “substrate” i.e. NADH or NADPH to the crude homogenates or membrane preparations. To our knowledge, the true cellular ROS formation by Nox enzymes in living cells has never been systematically compared to the signal obtained in homogenates after the addition of NADPH. To determine whether the currently used assays are detecting Nox activity, we generated a triple knockout mouse lacking Nox1, Nox2 and Nox4 (3N-/-) and compared the signals generated in the different assays. Moreover, two separate approaches i.e. in vivo angiotensin II (AngII) treatment and ex vivo transforming growth factor β (TGFβ) application were used to increase Nox expression. Triple Nox knockout mice are viable and lack gross abnormalities 3N-/- mice showed no obvious phenotype. There was a trend towards lower systolic blood pressure and a significantly higher basal heart rate in 3N-/- as compared to wildtype (WT) mice (Fig.1A-E). As shown by PCR, the 3N-/- animals lacked Nox1, Nox2 and Nox4, whereas Nox3 was expressed at extremely low levels. Vascular expression of eNOS was significantly decreased in 3N-/- (Table 1). Following exposure to AngII, the expression of Nox3 increased in 3N-/- but not in WT mice. In the heart, eNOS and Duox2 expression were significantly higher in 3N-/- mice. In the kidney, Nox3 and Duox2 were increased in the 3N-/- mice but no difference in eNOS expression was detected. eNOS expression was decreased in the lungs of 3N-/- mice whereas Duox1 was increased (Table 2). These data indicate that some compensation occurs in 3N-/- mice which might be relevant for eNOS, Duox1/2 and Nox3. Triple knockout of Nox results in an attenuated response to AngII In response to AngII, the heart rate was more markedly decreased in the 3N-/- mice, so that in the presence of AngII, the heart rate in both strains was similar (Fig.1D). The increase in blood pressure in response to AngII was similar in both strains, although there was a trend towards an attenuated response in 3N-/- mice (Fig.1A-D). As judged by heart to body weight ratio, only WT animals developed cardiac hypertrophy to AngII (Fig.1E). Similarly, AngII induced a rightward shift in the endothelium-dependent relaxation of mesenteric arteries from WT mice but not 3N-/- mice (Fig.1F). Interestingly, the constrictor response in mesenteric arteries from WT mice was less than that of all the other groups, potentially indicating that these animals had the highest basal NO availability (Fig.1G). The compensations occurring in a triple knockout mouse are complex and the genetic situation is hard to control. Nevertheless, it is fascinating that the 3N-/- mice comprise a mixed phenotype of the individual Nox knockout: similar as the Nox1 and Nox2 knockout mice (7), they do not develop endothelial dysfunction in response to AngII but due to the lack of Nox4 (9), their endothelium-dependent relaxation under basal conditions is not enhanced. Similar as in Nox4-/- mice, expression of eNOS is attenuated in tissue rich of endothelial cells, whereas the failure to induce cardiac hypertrophy in response to AngII was previously noted in Nox2-/mice (3).

4

Page 5 of 23


5 Aortic rings from wild-type chemiluminescence assays

and

triple

knockout

give

similar

signals

in

To correlate these functional data with Nox activity assays, measurements with lucigenin and L012 were performed. There was no detectable difference in the basal lucigenin signal of the rings in the absence of NADPH but its addition resulted in a massive increase in the signal. The latter was, however, similar in tissues from WT and 3N-/- mice (Fig.2A), indicating that the lucigenin signal generated by isolated mouse aortic rings is not derived from Nox1, Nox2 or Nox4. In the L012 assay, aortic rings from 3N-/- mice yielded a slightly (but nonsignificantly) higher signal than WT mice. In vivo treatment with AngII had no effect on the chemiluminescence signal but the acute application of PMA, to stimulate Nox2, induced a significant increase in the signal from WT aortic rings, but not rings from 3N-/- mice (Fig.2B). The decrease in the signal in response to PMA in the 3N-/- group most probably reflects the fact that the solvent DMSO is a strong quencher of the L012 signal. Thus, L012 chemiluminescence can detect the PMA-stimulated activation of Nox2 but is unable to detect the increase in Nox1 expression/activity associated with AngII treatment with the current amount of tissue and number of samples included. Triple Nox knockout has no effect on the chemiluminescence signal of cardiac membranes. Membrane preparations from wild-type and 3N-/- mouse hearts were compared using all three enhancers. Whereas the membranes produced no basal chemiluminescence with lucigenin alone (data not shown), a signal was readily detectable after the addition of NADPH. This, however, was identical in samples from WT and 3N-/- mice and not affected by in vivo AngII treatment (Fig.3A). Likewise, with luminol/HRP there was no difference in the NADPHinduced signal between samples from WT and 3N-/- mice (Fig.3B). Interestingly, the L012 assay yielded a signal by membranes even in the absence of NADPH. This signal might reflect an alteration of the autoxidation of L012 mediated by peroxidases or small molecules (F.R., unpublished observation). In response to NADPH the signal increased by almost 10 fold and extracts from 3N-/- hearts generated more signal as those from WT heart. Nevertheless, there was no difference in the signal to tissues from AngII-treated mice (Fig.3C). On this basis, it has to be concluded that all three assays failed to detect the activity of Nox1, Nox2 and Nox4 in cardiac membranes. Kidney membranes generate a complex chemiluminescence signal Kidney is the organ with the highest Nox4 expression. Amplex red/HRP assays in freshly isolated tissue revealed lower levels of H2O2 in samples from 3N-/- mice (Fig.4A). Interestingly, this effect was not observed when the samples were further processed and the assay repeated using renal membrane preparations stimulated with NADPH in the presence of luminol/HRP, which is generally assumed to detect H2O2 (Fig.4B). Unexpectedly, a lower NADPH-amplified lucigenin chemiluminescence signal was detected in membranes from 3N-/mice versus WT counterparts, despite the fact that Nox4 produces H2O2 rather than superoxide (Fig.4C). As with the cardiac membrane preparations, L012 chemiluminescence yielded an autoxidation signal in membrane samples without NADPH, which was higher in 3N-/- than WT preparations. Upon addition of NADPH, the L012 chemiluminescence increased and was similar between the WT and 3N-/- preparations (Fig.4D). In vivo AngII treatment slightly increased the NADPH-stimulated signal in the luminol/HRP assay and L012 assay of WT mouse kidney membranes, and this effect was not observed in membranes prepared from 3N-/- mice. In contrast, the lucigenin signal of renal membranes was similar between fractions obtained from mice with and without in vivo AngII treatment 5

Page 6 of 23


6 (Fig.4B-D). Collectively, these data suggest that renal membranes of 3N-/- have a reduced NADPH-stimulated lucigenin, which can, however, not be linked directly to Nox activity. Nox4 is not the source of the signal of the NADPH-stimulated luminol/HRP assay In order to further dissect the previous data, experiments were performed in murine dermal fibroblasts in the absence and presence of TGFβ (16 hours, 10ng/mL), which markedly increased Nox4 protein expression (Fig.5A). Intact WT cells generated a higher basal signal than 3N-/- cells in the Amplex red/HRP and luminol/HRP assays (Fig.5B&C). Paralleling the induction of Nox4, TGFβ increased the chemiluminescence signals from WT but not 3N-/cells (Fig.5B&C). In order to determine whether measurements could also be made in disrupted cells, NADPH enhanced luminol/HRP chemiluminescence was assessed in homogenates from the same cells. The signals generated by homogenates from WT and 3N/fibroblasts, however, were similar and unaffected by TGFβ (Fig.5D). Thus, although HRPbased assays are capable of detecting Nox4-dependent H2O2 production in living cells, the signal obtained in homogenates is unrelated to Nox4. Chemiluminescence assays in membrane fractions detect CYP and eNOS activity rather than Nox-derived ROS production Among the membrane-anchored enzymes, NO synthases (NOS) and cytochrome P450 enzymes produce ROS. To assess whether assay signals can be attributed to these enzymes, HEK293 cells overexpressing endothelial NOS (eNOS) or CYP2C8 were compared to Nox4 and Nox5. In intact HEK293 cells, Nox4 and Nox5 produced a strong signal in the presence of luminol/HRP, whereas in the presence of L012 or lucigenin, which are thought to detect primarily O2- , the signal was strongest in the Nox5-expressing cells (Fig.6A-C). Neither CYP2C8 nor eNOS overexpression resulted in chemiluminescence under these conditions. This changed after addition of NADPH, so that there was a pronounced increase in the signal in the eNOS-transfected cells as well as a clear increase in the CYP2C8-expressing. In contrast, the chemiluminescence of WT or Nox4-transfected cells only doubled and the WT, Nox4 and Nox5 overexpressing cells all generated similar signal intensities. Thus, NADPH obscures the signal generated by Nox enzymes (Fig.6D). A similar picture was obtained using membrane fractions whereby the Nox4- and Nox5overexpressing cells were indistinguishable from WT cells in NADPH-enhanced luminol/HRP or L012 chemiluminescence, while a 5-fold higher signal was detected in eNOS or CYP2C8overexpressing cells (Fig.6E&F). In the lucigenin assay using isolated membranes, there was no clear signal generated by the Nox or CYP2C8-expressing cells while eNOS resulted in the highest signal (Fig.6G). Thus, NADPH-stimulated membrane assays fail to detect Nox activity but are reasonable indicators of eNOS activity. The latter effect is a potential explanation why so many studies detected differences in the lucigenin assay which they erroneously attributed to Nox: the expression of eNOS is redox-sensitive and H2O2, depending on the concentration applied, can induce or impair eNOS expression. Similarly, stimuli which can activate Nox2, like TNFα, also activate eNOS and lead to an increase of the inducible NOS. That NOS is a special oxidoreductase and well established due to its diaphorase activity – the ability to oxidize reduced NADH or NADPH. Even in the absence of NADPH, NOS can reduce electron acceptors like formazan (5) and thus it is conceivable that compounds like lucigenin, which also act as electron acceptors are directly reduced by NOS in the present of NADPH. Similar, also cytochrome P450 monoxygenases are oxidoreductases and also these have been shown to directly reduce lucigenin (2).

6

Page 7 of 23


7 In kidney homogenates, loss of Nox enzymes resulted in attenuated NADPH-stimulated lucigenin chemiluminescence without affecting, however, any other assay. Moreover, overexpression of Nox4 in HEK293 cells had no significant effect on NADPH-stimulated lucigenin chemiluminescence. Given that Nox4 is through to predominantly produce H2O2, we are currently unable to reconcile this finding. Probably, additional, methodological independent NADPH oxidase assays have to be applied to address this aspect. However, NADPH interferes with most ROS-detectors, even cytochrome C, although chemiluminescence assays are particular sensitive to this artefact. Thus, ROS-independent measures of Nox activity have to be applied. NADPH consumption in homogenates is high and basically Nox-independent (R.P.B., unpublished observation). In leukocyte homogenates, in the classic Nox assay, oxygen consumption is recorded. In other cell types, the impact of Nox enzymes on oxygen consumption was too low to yield reliable measurement (R.P.B., unpublished observation). A potential explanation for our failure to detect Nox activity in membrane preparations is that these may lack a constituent of Nox enzymes. For example, it was suggested to add FAD to membrane assays (8) and also the “traditional” cell-free assays of leukocytes in which oxygen consumption is measured, requires addition of FAD, FMN and the cytosolic cofactors. Addition of FAD or using crude homogenates rather than isolated membrane, however, also did not yield a Nox-attributable chemiluminescence signal in our hands (data not shown). It is, however, possible that Nox enzymes degrade more easily than eNOS or CYP450 enzymes due to their dependency on p22phox or through irreversible oxidation. It has been shown that diamide and thiol oxidation lowers Nox activity (6). In fact, we did not include thiol protection by DTT in our assays, as this does not belong to the standard protocols for Nox measurement in membranes. In conclusion: We generated and characterization Nox1-Nox2-Nox4 triple knockout mice, which develop normally and have a phenotype with aspects of knockout of the individual Nox homologues. Importantly, NADPH-stimulated chemiluminescence is grossly unaffected by the triple knockout. Thus, the present study casts serious doubt on the validity of “Nox membrane assays”. Scientists are advised to measure ROS formation in intact tissue rather than boosting their signal with NADPH or using homogenates when addressing ROS production to Noxes.

Notes Nox-/- animals and animal procedure Triple knockout mice (3N-/-) for Nox1, Nox2 and Nox4 were generated by crossing Nox1 and Nox2 mice, as both genes are on the X chromosome. The Nox1/2 double knockout animals obtained were further crossed with Nox4-/- animals. All mice were previously backcrossed at least for 10 generations into the C57/Bl6J background. Nox2-/- mice were obtained from Jackson labs, Nox1-/- mice were kindly provided by K.H. Krause (University of Geneva). Nox4-/- mice were generated by the authors. All experiments performed with animals were in accordance with German animal protection law and were carried out after approval by the local authorities. Animals were housed in groups with free access to chow and water in a specified pathogen free facility with a 12/12 hours day/night cycle. Given the impact of gender on ROS production, only male animals aged 8-16 weeks were used in this study. 7

Page 8 of 23


8 RT-qPCR Total mRNA from frozen homogenized tissue was isolated with a RNA-Mini-kit (Bio&Sell, Feucht, Germany) according to the manufacturer’s protocol. Random hexamer primers (Promega, Madison,WI, USA) and Superscript III Reverse Transcriptase (Invitrogen, Darmstadt, Germany) were used for cDNA synthesis. Semi-quantitative real-time PCR was performed with Mx3000P qPCR cycler (Agilent Technologie, Santa Clara, CA, USA) using PCR Eva Green qPCR Mix with ROX (Bio&Sell, Feucht, Germany) with appropriate primers. Relative expression of target genes were normalized to eukaryotic translation elongation factor 2 (EF2), analyzed by delta-delta-Ct method and represented as percent of control samples. Primer sequences are listed in table 3. Angiotensin treatment and blood pressure measurements Animals (n≥7 per group) were treated for two weeks with AngII (0.7mg/kg/d) delivered by a miniosmotic pump (Alzet). Tail cuff measurements were performed with a 6 channel setup (Vistech BP2000) and measurements of 5 days were averaged per mouse. Organ chamber experiments Organ chamber experiments were performed using mesenteric vessels in carbogen-aerated Krebs-Henseleit buffer containing diclofenac (10µmol/L). The phenylephrine concentration was cumulatively increased (0.03 to 0.3 µmol/L) to obtain a preconstriction level of approximately 80% of the initial KCl constriction. Subsequently, acetylcholine (ACh) was cumulatively added and dose-response curves for the relative constrictor response to phenylephrine and the dilator response to acetylcholine were constructed. ROS measurements Radical production was assessed by chemiluminescence with luminol (100 µmol/L)/horseradish peroxidase (HRP at 1 U/mL), L012 (200 µmol/L) or lucigenin (10 µmol/L) in a Berthold 6-channel luminometer (LB9505, Berthold, Wildbad, Germany). For measurements with cell homogenates or membrane fractions, 20 µg of protein were used for each sample and measurements were initiated by addition of NADPH (10 µmol/L for luminol and L012; 100 µmol/L for lucigenin). Chemiluminescence is expressed as arbitrary units. All measurements were performed in HEPES-Tyrode (HT) buffer containing in mmol/L: 137 NaCl, 2.7 KCl, 0.5 MgCl2, 1.8 CaCl2, 5 glucose, 0.36 NaH2PO4, 10 HEPES. Isolation of skin fibroblasts Murine dermal fibroblasts were isolated by collagenase Type II digestion (1000U/mL). Cells were cultured in DMEM/F12 medium containing 10% fetal calf serum and antibiotics. Amplex red measurements from cells and renal tissue Cells were cultured in 12 well plates and treated overnight with or without TGF-β (10 ng/mL). H2O2-dependent Amplex Red (50 μmol/L, Invitrogen) conversion to resorufin catalyzed by horseradish peroxidase (2 U/mL) was carried out in HT buffer. Fluorescence was determined in the supernatant at 540nm/580nm excitation/emission. Kidneys were freshly isolated and HT buffer was added to 1mL/mg tissue. Organs were then minced with a scalpel and H2O2-derived fluorescence was measured in the supernatant of 100mg tissue. 8

Page 9 of 23


9

Immunoblotting analysis for Nox4 Western blot for Nox4 was performed according to standard technique. Cells were lysed in 1% triton Tris-HCl pH7.5 buffer containing protease inhibitors. Samples were not heatdenatured and TCEP (tris(2-carboxyethyl)phosphine) was used as reducing agent. Nox4 was detected by an antibody reported by Anilkumar et al. (1). Overexpression system using HEK293 cells HEK293 cells were either stably transfected with a plasmid coding for Nox4 or transiently transfected with 1µg plasmid DNA encoding for Nox5, eNOS, CYP4502C8 using lipofectamine according to manufacturer’s instructions. Membrane fraction preparation from cells and tissue homogenates Cells and tissues were homogenized by pottering in HT buffer containing a protease inhibitor mix (antipain, aprotinin, chymostatin, leupeptin, pepstatin, trypsin-inhibitor; AppliChem), okaidaic acid and calyculin A. Cell homogenates were cleared by centrifugation (3000g, 10 minutes, 4oC) and membrane fractions were obtained by centrifugation at 100,000g (1 hour, 4oC). The membrane pellet was resuspended in HT buffer and the protein concentration was determined by Bradford assay. Twenty micrograms were used for chemiluminescence measurements. Statistics Unless otherwise indicated, data are given as means ± standard error of mean (SEM). n refers to the number of animals, where tissue / animal data were shown and to the number of independent experiments where cell culture data was reported.. Calculations were performed with Prism 5.0 or BiAS.10.12. The latter was also used to test for normal distribution and similarity of variance. In case of multiple testing, Bonferroni correction was applied. For multiple group comparisons ANOVA followed by post hoc testing was performed. Individual statistics of unpaired samples was performed by T-test and if not normal distributed by Mann-Whitney test. p-value of