Glucose6Phosphate Dehydrogenase Deficiency Decreases Vascular Superoxide and Atherosclerotic Lesions in Apolipoprotein E/ Mice

Glucose-6-Phosphate Dehydrogenase Deficiency DecreasesVascular Superoxide and Atherosclerotic Lesions in

Apolipoprotein E�/� MiceReiko Matsui, Shanqin Xu, Karlene A. Maitland, Roberto Mastroianni, Jane A. Leopold,

Diane E. Handy, Joseph Loscalzo, Richard A. Cohen

Objective—Glucose-6-phosphate dehydrogenase (G6PD) is a key enzyme in the pentose phosphate pathway that is a majorsource of cellular NADPH. The purpose of this study was to examine whether G6PD deficiency affects vascularoxidants and atherosclerosis in high-fat fed apolipoprotein (apo) E�/� mice.

Methods and Results—G6PD-mutant mice whose G6PD activity was 20% of normal were crossbred with apoE�/� mice.Among male apoE�/� mice that were fed a western-type diet for 11 weeks, G6PD wild-type (E-WT), and G6PDhemizygous (E-Hemi) mice were compared. Basal blood pressure was significantly higher in E-Hemi. However,superoxide anion release, nitrotyrosine, vascular cell adhesion molecule (VCAM)-1, and inducible nitric oxide synthaseimmunohistochemical staining were less in E-Hemi compared with E-WT aorta. Serum cholesterol level was lower inE-Hemi, but aortic lesion area was decreased in E-Hemi even after adjusting for serum cholesterol.

Conclusions—Lower NADPH production in G6PD deficiency may result in lower NADPH oxidase-derived superoxideanion, and thus lower aortic lesion growth. The association of higher blood pressure with lower serum cholesterol levelsin this mouse model is indicative of the complex effects that G6PD deficiency may have on vascular disease.(Arterioscler Thromb Vasc Biol. 2006;26:910-916.)

Key Words: atherosclerosis � genetically altered mice � reactive oxygen species � NADPH

Enhanced vascular superoxide anion production is associ-ated with hypercholesterolemia and may contribute to

the initiation and progression of atherosclerosis.1 Vascularcell-derived superoxide anion mediates oxidative modifica-tion of low-density lipoprotein (LDL)2,3 and oxidized LDLfurther promotes superoxide production and foam cell forma-tion.4 Also, reactive oxygen species promote various pro-cesses including endothelial dysfunction, smooth muscle cellgrowth and migration, and induction of adhesion molecules.5

A major source of superoxide anion in vascular cells isNADPH oxidase, the expression of which is enhanced inatherosclerotic lesions.5–7 Pharmacological inhibition ofNADPH oxidase decreased aortic superoxide anion produc-tion and atherosclerotic lesions in apoE�/� mice,8 and geneticdeficiency in the p47phox subunit of NADPH oxidase attenu-ates the inflammatory response to hypercholesterolemia9 andreduces aortic lesions in apoE�/� mice.10

Glucose-6-phosphate dehydrogenase (G6PD), a key en-zyme in the pentose phosphate pathway, provides NADPHfor various cellular reactions including glutathione (GSH)recycling, superoxide anion production via NADPH oxi-dase, NO synthesis, and cholesterol synthesis. Inhibition of

G6PD results in decreased production of superoxide and/orNO in granulocytes11,12 and other cell types includingendothelial cells.13–15 In addition, recent studies suggestthat the level of pentose phosphate pathway-derivedNADPH may regulate vascular superoxide production.16

Consistent with these studies, we found that G6PD-deficient mice had lower aortic superoxide production andless hypertrophy in response to angiotensin II infusion.17

This is a rather paradoxical result, because G6PD isgenerally considered to be an antioxidant enzyme. G6PDnull embryonic stem cells are extremely sensitive tooxidative stress.18 In various conditions G6PD activity israpidly upregulated in response to oxidative stress, pre-sumably to maintain GSH in its reduced form.19 –22 HumanG6PD deficiency, the most common genetic enzymopathy,is reported to either enhance or decrease the risk ofcardiovascular disease,23,24 but the mechanisms by whichrisk might be affected are not known.

In the present study, we examined whether G6PD defi-ciency affects the development of atherosclerosis by modifi-cation of oxidant production because of an altered supply ofNADPH. We crossbred mice whose G6PD activity was

Original received July 15, 2005; final version accepted January 13, 2006.From the Vascular Biology Unit (R.M., S.X., K.A.M., R.M., R.A.C.), Whitaker Cardiovascular Institute (J.A.L., D.E.H., J.L.), Evans Department of

Medicine Boston University School of Medicine, Boston, Mass.Correspondence to Reiko Matsui, MD, Vascular Biology Unit, Boston University School of Medicine, X707, 650 Albany St, Boston, MA 02118.

E-mail [email protected]© 2006 American Heart Association, Inc.

Arterioscler Thromb Vasc Biol. is available at http://www.atvbaha.org DOI: 10.1161/01.ATV.0000205850.49390.3b

910 by guest on July 26, 2015http://atvb.ahajournals.org/Downloaded from by guest on July 26, 2015http://atvb.ahajournals.org/Downloaded from by guest on July 26, 2015http://atvb.ahajournals.org/Downloaded from by guest on July 26, 2015http://atvb.ahajournals.org/Downloaded from by guest on July 26, 2015http://atvb.ahajournals.org/Downloaded from by guest on July 26, 2015http://atvb.ahajournals.org/Downloaded from by guest on July 26, 2015http://atvb.ahajournals.org/Downloaded from by guest on July 26, 2015http://atvb.ahajournals.org/Downloaded from by guest on July 26, 2015http://atvb.ahajournals.org/Downloaded from by guest on July 26, 2015http://atvb.ahajournals.org/Downloaded from by guest on July 26, 2015http://atvb.ahajournals.org/Downloaded from

�20% of normal with apoE�/� mice. We found that loweractivity of G6PD is associated with higher basal bloodpressure, but lower superoxide, serum cholesterol, and ath-erosclerotic aortic lesions.

Materials and MethodsAnimal ModelThe G6PD-deficient mouse model in the C3H strain was bred at ourinstitution from frozen embryos obtained from the Medical ResearchCouncil (Harwell, UK).25 The G6PD deficient mouse was originallycreated by Pretsch and Charles26 and showed decreased translation ofthe protein caused by a single mutation in the untranslated region inthe splice site of the X-linked G6PD gene.27 Male hemizygotes(XmY) with a C3H background were bred with female apoE�/� micewith C57BL/6J background obtained Jackson Laboratories (BarHarbor, Me), and female offspring (F1) (apoE�/�, G6PD XmX) werebred with male apoE�/� mice. Among offspring (F1�, F2) of thisbreeding, male apo E�/� mice littermates were selected with X or Xm

G6PD genotype which was ascertained by polymerase chain reaction(PCR) as previously described.22 Therefore, both apoE�/� control(E-WT) and apoE�/� G6PD mutant (E-Hemi) mice used in this studywere littermates and had �25% C3H and 75% C57BL/6 geneticbackground (Figure I, available online at http://atvb.ahajournal-s.org). Further analysis of their genetic background was done basedon analyzing Ath-1, the principal genetic locus that determinesatherosclerosis susceptibility in the C57BL/6 strain28 and tnfsf4 thegene within that locus to which that susceptibility has been attribut-ed29 (Supplement, please see http://atvb.ahajournals.org.). MaleE-WT and E-Hemi mice were fed a western diet (21% fat, 0.15%cholesterol w/w, TD 88137; Harlan Teklad, Madison, Wis) betweenage 10 and 21 weeks. Systolic blood pressure was determined by tailcuff plethysmography (Visitech Systems, Apex, NC) as describedpreviously.30,31 This method includes training sessions, a quiet,warm, and dark environment, and provides values similar to thoseobtaining by intra-arterial measurements in conscious mice.29 Theprotocol was approved by the Boston University Medical CenterInstitutional Animal Care and Use Committee.

Measurement of Serum CholesterolAt 21 weeks of age, mice were anesthetized with isoflurane andblood, aorta, heart, and liver were removed. Serum cholesterol wasmeasured enzymatically using a kit from Sigma Diagnostics.8

Detection of Aortic Superoxide Anion byLucigenin ChemiluminescenceMeasurement of superoxide anion from intact mouse aorta wasperformed according to the method published previously.8,30 Briefly,the aorta was isolated under a dissecting microscope and incubatedin a tube containing 1 mL of physiological buffer with lucigenin(5 �mol/L). This lower concentration of lucigenin was demonstratednot to be involved in redox cycling.32 The tube was placed in aluminometer (model 20e; Turner Design, Mountainview, Calif) inwhich the light chamber was maintained at 37°C. The luminometerwas set to report arbitrary units of emitted light; after a 15-minuteequilibration, repeated measurements were integrated every 30seconds, and an average value was obtained over a 5- to 10-minuteperiod. Tiron (10 mmol/L), a cell-permeable nonenzymatic scaven-ger of superoxide anion, was then added to quench all superoxideanion-dependent chemiluminescence. Tiron-quenchable chemilumi-nescence was normalized to aortic wet weight.

Determination of G6PD ActivityFreshly isolated aorta was homogenized in 20 mmol/L Tris bufferwith 0.35 mol/L sucrose and centrifuged at 12 000g for 5 minutes.The supernatant was analyzed for protein concentration, and enzy-matic activity of G6PD was assayed according to the methoddescribed elsewhere.17,33

Quantification of Aortic AtheroscleroticLesion AreaAtherosclerotic lesions were quantified by planimetry of SudanIV-stained lesions on the aortic intima as described previously.8 Theentire thoracic and abdominal aorta was cut open longitudinallythrough its ventral side under a dissecting microscope, and immersedin Sudan IV (Fisher). Quantification of stained lesion area wasperformed on the digitized images using Scion Image and NIHImage software.

Immunohistochemistry of Aortic SectionsThe aortic arch was placed in 4% formalin overnight, dehydrated,and embedded in paraffin. Tissue sections (5 �m) were obtainedfrom the descending thoracic aorta, 3 mm distal to the left subclavianartery, and processed as previously described in detail.17,30 Specific-ity of anti-3-nitrotyrosine antibody was confirmed as previouslydescribed.17 Polyclonal anti-VCAM-1 antibody was purchased fromSanta Cruz Biotechnology (Santa Cruz, Calif). Polyclonal anti-iNOSantibody was from Biomol (Plymouth Meeting, Pa). To assurespecific staining, staining with all antibodies was routinely per-formed with comparisons made to a nonspecific IgG control anti-body as shown in Figure 4a.

Semi-Quantitative Analysisof ImmunohistochemistryScoring of nitrotyrosine and iNOS in aorta was performed based onthe method previously used in our laboratory.30 Photographs ofimmuno-stained mouse aorta were taken under microscope (�100magnification) and randomly shown to observers without identifica-tion of samples. Score (grade 0 to 4) was given to aortic endothelium,media, and adventitia, respectively, according to the intensity ofstaining. The average of scores from 3 observers for each elementwas taken as scores for the sample.

Western BlotA part of the thoracic aorta was homogenized in lysis buffer (1%NP-40, 0.25% deoxycholic acid, 50 mmol/L Tris, pH 7.4, 1 mmol/LEDTA, 150 mmol/L NaCl, 1 mmol/L NaF, 1 mmol/L sodiumorthovanadate, 1 mmol/L phenylmethylsulfonyl floride, 10 �g/mLleupeptin). Protein was analyzed by immunoblot as previouslydescribed.17 Polyclonal anti-G6PD antibody was obtained fromBethyl Laboratories (Montgomery, Tex). Monoclonal anti-�-actinantibody was from Sigma (St. Louis, Mo), and polyclonal mouseVCAM-1 antibody was from R&D systems (Minneapolis, Minn).

Measurement of Tissue GlutathioneA piece of frozen tissue (heart, aorta) was homogenized in 0.35 Nperchloric acid and centrifuged at 3000g for 10 minutes. Thesupernatant was used to assay glutathione (GSH) according to thecolorimetric method provided by a kit (GSH-400; Oxis, Portland,Ore). The precipitate was re-suspended in 0.2 N NaOH and used forprotein assay (Bio-Rad). Tissue GSH content was expressed asnmol/mg protein.

Data AnalysisData are expressed as mean�SE. Statistical comparisons wereperformed by ANOVA and Student t test. Significance was acceptedwhen P�0.05.

ResultsGenetic Characterization of Mice in the StudyAmong mice used in the study, the percentage of C57BL/6background was tested by analysis of microsatellites on theAth-1 locus on chromosome-1 that determines susceptibilityto atherosclerosis.28,29 Of the 10 mice analyzed in each group5 of the mice were homozygous C57BL/6 at this locus in theE-WT group and 5 were homozygous in the E-Hemi group

Matsui et al G6PD Deficiency and Atherosclerosis 911

by guest on July 26, 2015http://atvb.ahajournals.org/Downloaded from

(Figure II, available online at http://atvb.ahajournals.org).The remaining 5 mice in each group analyzed were heterozy-gous C57BL/6 at this locus, indicating that the geneticbackground of Ath-1 locus was similar between E-WT andE-Hemi. In addition, the cardiac mRNA expression of tnfsf4that is the critical gene within Ath-1 locus that determinessusceptibility of the C57BL/6 strain to atherosclerosis29 wasnot significantly different between E-WT and E-Hemi mice(Figure III, available online at http://atvb.ahajournals.org).Thus, differences in the 2 groups could not be ascribed todifferences in genetic background of Ath-1 locus.

G6PD-Deficient apoE �/� Mice (E-Hemi)Demonstrate Higher Blood Pressure and LowerSerum Cholesterol Than apoE�/� Mice (E-WT)After administration of western diet, blood pressure measuredby tail-cuff was significantly higher in E-Hemi mice (E-WT92�2 versus E-Hemi 100�3, mm Hg, P�0.05; Figure 1).When mice were euthanized at 21 weeks of age, E-Hemi micedemonstrated the same body weight, heart weight, and liverweight as those of E-WT mice. However, serum cholesterollevel was significantly lower (E-WT 1756�82 versusE-Hemi 1567�57, mg/dL; P�0.05; Table I, available onlineat http://atvb.ahajournals.org). After administration of west-ern diet, G6PD activity in E-Hemi mouse aorta was 23% ofthat in E-WT mouse aorta. Immuno-blot also confirmed thatG6PD protein expression was lower in E-Hemi aorta com-pared with E-WT (Figure IV, available online athttp://atvb.ahajournals.org).

Aortic Lesion Area Is Significantly Less in E-HemiCompared With E-WTSudan IV-stained lesions were concentrated in the aortic archand spinal artery branches in both E-WT and E-Hemi mice(Figure 2). The average atherosclerotic lesion area wassignificantly less in E-Hemi mice (E-WT 1653�214 versusE-Hemi 938�131, �103 �m2, n�15 to 16, P�0.01; Figure2). When mice in the 2 groups were selected to match serumcholesterol levels within the same range (E-WT 1478�87versus E-Hemi 1533�58, mg/dL, n�12 to 13, not signifi-

cant), the lesion area was still significantly less in E-Hemimice (E-WT 1715�260 versus E-Hemi 1025�133,x103 �m2, P�0.05). The significant difference in lesion areaalso was found in mice that were homozygous C57BL/6 atthe Ath-1 locus (P�0.01, n�5, Figure II).

Less Aortic Lesion in E-Hemi Mice WasAssociated With Lower Expression of VCAM-1VCAM-1 expression as a marker of vascular inflammationwas studied by immunohistochemistry and immunoblot.VCAM-1 staining was localized on atheromatous plaques butalso observed in nonlesion endothelium in E-WT mouseaorta. E-Hemi mouse aorta showed less staining comparedwith E-WT mouse aorta (Figure 4a). Immunoblot of aortichomogenate also showed lower expression of VCAM-1 inE-Hemi mice (Figure 3).

G6PD-Deficient Mice Generate Lower SuperoxideAnion and Demonstrate Lower 3-o-NitrotyrosineStaining in the AortaAortic NADPH content was decreased �50% in the G6PDmutant mice,17 and therefore superoxide anion generation wasmeasured in aorta of apoE�/� mice to examine whether lowerNADPH associated with G6PD deficiency might contributeto lower superoxide production via vascular NADPH oxidase.Superoxide anion production detected by lucigenin wassignificantly lower in E-Hemi mouse aorta (E-WT 9.3�1.1versus E-Hemi 6.2�0.9 mU/min/mg aorta, n�12 to 15,

Figure 1. Blood pressure (BP) of western diet-fed apoE �/�

mice. Each point represents an average of more than ten read-ings of tail systolic BP in each mouse. The average (the horizon-tal line) of BP in E-Hemi was significantly higher than that inE-WT (*P�0.05).

Figure 2. Aortic lesions in western diet-fed apo E�/� mouse. a,En face aortic lesion stained with Sudan IV. Shown are stainedlesions (black areas) on the aortic intima of representative aor-tas. b, Lesions were quantified by planimetry using digitizedimages. Each symbol represents the lesion area (�103 �m2) on1 mouse aorta.

912 Arterioscler Thromb Vasc Biol. April 2006

by guest on July 26, 2015http://atvb.ahajournals.org/Downloaded from

P�0.05). In addition, nitrotyrosine staining was less inE-Hemi mice (Figure 4a). Semi-quantitative analysis in 7 to 8mice per group showed significantly less staining in E-Hemimice aorta, consistent with lower production of superoxideanion (Figure 4b). Nitrotyrosine staining was less in lesions(Figure 4a), and significantly less in the media and adventitiathat are not involved with lesions (Figure 4b).

Inducible Nitric Oxide Synthase Is Less in E-HemiCompared With E-WT Mouse AortaInducible NO synthase (iNOS) is induced by inflammatorycytokines in atherosclerotic lesions and may be responsiblefor producing reactive oxygen/nitrogen species includingnitric oxide and peroxynitrite. Immunohistochemistry dem-onstrated iNOS expression in endothelium, media, and ad-ventitia of E-WT mouse aortas, but significantly less expres-sion was observed in all aortic cell layers in E-Hemi mice(Figure 5).

GSH Content Was Not Decreased in E-Hemi MiceFed Western DietBecause GSH can be decreased by acute oxidative stress inG6PD-deficient tissues,22 GSH content was measured in heartand aorta of E-WT and E-Hemi mice. Although G6PDactivity of E-Hemi heart was 20% of E-WT heart (data notshown), GSH in the heart (E-WT 7.3�0.2 versus E-Hemi7.3�0.5 nmol/mg protein, n�4) and in the aorta (E-WT39�12 versus E-Hemi 33�13 nmol/mg protein, n�4 to 6)were not significantly different.

DiscussionIn this study, apoE�/� mice with G6PD deficiency demon-strated less atherosclerotic lesions associated with lowersuperoxide anion production and nitrotyrosine in the aorta.The aorta of the G6PD mutant mouse has 10% to 20% ofnormal G6PD activity and 50% of normal NADPH contentcompared with WT mice.17 The data presented here suggestthat a lower supply of NADPH is associated with lowersuperoxide anion production via NADPH oxidase, whichcontributes to decreased lesion formation in apoE�/� mice.

Figure 3. VCAM-1 expression in western diet-fed apoE�/�

mouse aorta. a, VCAM-1 and �-actin expression in aorta areshown by immuno-blot of aortic proteins. Representative exam-ples show lower expression in E-Hemi aorta. b, Densitometricdata for VCAM-1 expression from immuno-blots normalized to�-actin expression (n�6, P�0.05).

Figure 4. a, Representative immunohistochemistry for VCAM-1 and nitrotyrosine in aortic cross-sections. 1 and 2, VCAM-1; 3 and 4,nitrotyrosine; 5, lack of staining of E-WT aorta with nonspecific IgG. (original magnification 100�). b, Semi-quantitative analysis of nitro-tyrosine staining was performed as described in the Methods. The average�SE show lower scores in E-Hemi mouse aorta in mediaand adventitia (n�7 to 8, *P�0.05, **P�0.01).

Matsui et al G6PD Deficiency and Atherosclerosis 913

by guest on July 26, 2015http://atvb.ahajournals.org/Downloaded from

These results are compatible with NADPH oxidase having arelatively high Michaelis constant for NADPH,34 whichtherefore makes its activity likely to be influenced by aninsufficient supply of its substrate. A role of NADPH oxidasein regulating atherogenesis is also consistent with reportsshowing less atherosclerosis in p47phox-deficient mice10 orwith pharmacological inhibition of NADPH oxidase.8

There are several other factors that may have influencedatherogenesis in the apoE�/� mice in this study. First, wefound that systolic blood pressure was significantly higher inE-Hemi mice. Although apoE�/�, G6PD mutant mice showeda trend toward higher basal blood pressure in our earlierstudy,17 the statistically significant difference observed inapoE�/� mice in this study may be because of the effect thatadditional factors including hypercholesterolemia have onblood pressure. The higher blood pressure in E-Hemi mice isconsistent with a clinical report showing higher blood pres-sure in G6PD-deficient men.23 Higher basal blood pressure inE-Hemi mice might be attributed to less endothelial nitricoxide (NO) production, which may also be because of a lowersupply of NADPH. In agreement with this interpretation, NObioavailability is decreased by inhibiting G6PD activity inendothelial cells in culture.14 Decreased NO bioavailabilitywould be expected to result in enhanced atherosclerosis inE-Hemi mice because eNOS deficiency enhances atheroscle-rosis in apoE�/� mice.35 The fact that this was not observedindicates that other factors such as the decrease in NADPHoxidase-derived superoxide anion overcame that of NO in thedevelopment of atherosclerosis. Inhibiting NADPH oxidasemay reduce vascular inflammation without changing vasculartone.36

Second, serum cholesterol was significantly lower inG6PD deficient apoE�/� mice. 3-hydroxy-3-methylglutaryl(HMG)-coenzyme A (CoA) reductase, as well as severalother enzymes downstream of it that are involved in choles-terol biosynthesis, requires NADPH as a cofactor and has a

relatively high Km for the cofactor (0.08 mmol/L37). There-fore, endogenous cholesterol synthesis is NADPH-dependentand is likely impaired in G6PD-deficient mice with dimin-ished generation of NADPH. When E-Hemi mice were fedregular diet, serum cholesterol was also 20% lower thanE-WT mice (data not shown). We speculate that endogenouscholesterol synthesis is lower in G6PD-deficient animals. Infact, an epidemiological study reported that the serum levelsof total cholesterol, LDL cholesterol, and high-density li-poprotein cholesterol were significantly lower in G6PD-deficient men.38 However, in our study, the lower serumcholesterol did not likely contribute to the decreased athero-sclerosis, because the lesion area was still significantly lesswhen comparing groups of mice with similar serum choles-terol values.

Third, the G6PD mutant mouse used in this study to breedwith the apoE�/� was of the C3H strain that is atherosclero-sis-resistant compared with the C57BL/6 strain.39 In thisstudy after breeding with apoE�/� mice of a C57BL/6background, approximately one-quarter of the genomic back-ground of both E-WT and E-Hemi littermate mice were of theC3H background. Although they were littermates, a smalldeference in genes could have influenced the extent ofatherosclerosis in mice used in this study.40 The Ath-1 locuson mouse chromosome-1 has been reported to renderC57BL/6 mice more susceptible and C3H mice more resistantto diet-induced atherosclerosis.28 Recently, tnfsf4 was identi-fied within the Ath-1 locus as the major gene that influencessusceptibility to atherosclerosis.29 Therefore, we tested mic-rosatellite markers on the Ath-1 locus by PCR and tnfsf4expression in mouse hearts by quantitative PCR (Figures II,III). These data confirmed that lesser lesion in E-Hemi did notresult from any difference in genetic background of Ath-1locus or from different expression of tnfsf4.

Fourth, lower levels of iNOS were observed in E-Hemiatherosclerotic mouse aorta. Previous studies have shown that

Figure 5. a, Immunohistochemical stain-ing for iNOS in western diet-fed apoE�/�

mouse aortas (original magnification400�). b, Semi-quantitative analysis ofiNOS staining (see Methods). Theaverage�SE show lower scores inE-Hemi mice in all layers of the aorticwall (n�8 to 9, *P�0.05, **P�0.01).

914 Arterioscler Thromb Vasc Biol. April 2006

by guest on July 26, 2015http://atvb.ahajournals.org/Downloaded from

inhibition of G6PD decreased NO production in somecells,11,13,15 and iNOS gene expression was decreased byinhibition of G6PD in glial cells.41 If reactive oxygen speciescontribute to activation of transcription factors such asNF-�B to induce the iNOS gene, decreased superoxide anionmay result in less iNOS induction. Further studies arerequired to elucidate the mechanism by which iNOS isdecreased in E-Hemi aorta. NO itself is thought to be ananti-atherogenic factor, reducing VCAM-1 expression incultured cells,42 and NOS inhibitors increase atherogenesis.43

However, peroxynitrite, a reaction product of NO and super-oxide anion, is a strong oxidant that may promote LDLoxidation.3 Also, in oxidative stress, uncoupled iNOS canproduce superoxide.44 These have been suggested as reasonswhy genetic deficiency45,46 or pharmacological inhibition ofiNOS reduces atherosclerosis.47 Thus, less iNOS in G6PDdeficient mouse aorta may have contributed to decreasedsuperoxide anion, reactive nitrogen species, nitrotyrosine, andatherosclerotic lesions. It is also possible that the effect of thedecreased iNOS derived oxidants on atherosclerosis in thismodel overcame any potential effect of decreased eNOSfunction. The decreased expression of iNOS is also consistentwith that of VCAM-1, another NF-�B–dependent gene in-volved in atherosclerosis. It is clear that the alterations iniNOS and nitrotyrosine in E-Hemi mice occurred in all layersof the aortic wall, not just in atherosclerotic lesions, consis-tent with the decrease in reactive species being a cause of thedecreased lesions, rather than the decrease in reactive speciesbeing a result of the decreased lesions.

Acute oxidative stress is often accompanied by depletionof GSH and induction of G6PD activity. In such casesinhibiting G6PD may limit GSH reductase activity whichregenerates GSH to its reduced form.19,20 However, we didnot find significant decreases in tissue GSH levels in the heartor aorta of E-Hemi mice, suggesting that other mechanisms tomaintain GSH levels are effective in chronic states of oxidantstress. GSH synthesis is induced by oxidized-LDL in macro-phages48 and known to be upregulated by oxidative stress inthe lung,49 so that it is possible that GSH level in the E-Hemimice are compensated by increased de novo synthesis.

This study together with our previous study17 implicates animportant role of NADPH derived from the pentose phos-phate pathway acting as substrate for vascular superoxidegeneration via NADPH oxidase. Also, our results support arole for vascular superoxide production in contributing to theprogression of atherosclerosis. Potentially because of theconflicting effects of G6PD deficiency cited, and the geneticcomplexity of humans with G6PD deficiency, the clinicalcardiovascular manifestations of G6PD deficiency may haveremained undetected. Our results are consistent with clinicalstudies suggesting that G6PD deficiency contributes to ahigher blood pressure,23 but a lower serum cholesterol38 andcardiovascular mortality associated with atherosclerosis.24 Inaddition to resistance to malaria,50 this report may indicateprotective aspects of G6PD deficiency for atherosclerosis fora large population with the most prevalent enzymopathy inthe world.

AcknowledgmentsThe studies were supported by National Institutes of Health grantsSCOR HL55993, and R01 HL55620, R01 AG 27080, andR03 AG19078.

References1. Miller FJ, Gutterman DD, Rios CD, Heistad DD, Davidson BL.

Superoxide production in vascular smooth muscle contributes to oxi-dative stress and impaired relaxation in atherosclerosis. Circ Res. 1998;82:1298–1305.

2. Heinecke JW, Baker L, Rosen H, Chait A. Superoxide-mediated modi-fication of low density lipoprotein by arterial smooth muscle cells. J ClinInvest. 1986;77:757–761.

3. White CR, Brock TA, Chang LY, Crapo J, Briscoe P, Ku D, Bradley WA,Gianturco SH, Gore J, Freeman BA, Tarpey MM. Superoxide and per-oxynitrite in atherosclerosis. Proc Natl Acad Sci U S A. 1994;91:1044–1048.

4. Rueckschloss U, Galle J, Holtz J, Zerkowski HR, Morawietz H. Inductionof NAD(P)H oxidase by oxidized low-density lipoprotein in humanendothelial cells: antioxidative potential of hydroxymethylglutarylcoenzyme A reductase inhibitor therapy. Circulation. 2001;104:1767–1772.

5. Griendling KK, Ushio-Fukai M. Reactive oxygen species as mediators ofangiotensin II signaling. Regul Pept. 2000;91:21–27.

6. Paravicini TM, Gulluyan LM, Dusting GJ, Drummond GR. IncreasedNADPH oxidase activity, gp91phox expression, and endothelium-dependent vasorelaxation during neointima formation in rabbits. CircRes. 2002;91:54–61.

7. Pagano PJ, Ito Y, Tornheim K, Gallop P, Cohen RA. An NADPH oxidasesuperoxide generating system in the rabbit aorta. Am J Physiol. 1995;268:H2274–H2280.

8. Cayatte AJ, Rupin A, Oliver-Krasinski J, Maitland K, Sansilvestri-MorelP, Wierzbicki M, Verbeuren TJ, Cohen RA. S17834, a new inhibitor ofcell adhesion and atherosclerosis that targets NADPH oxidase. Arte-rioscler Thromb Vasc Biol. 2001;21:1577–1584.

9. Stokes KY, Clanton EC, Russell JM, Ross CR, Granger DN. NAD(P)Hoxidase-derived superoxide mediates hypercholesterolemia-inducedleukocyte-endothelial cell adhesion. Circ Res. 2001;88:499–505.

10. Barry-Lane PA, Patterson C, van der Merwe M, Hu Z, Holland SM, YehETH, Runge MS. p47phox is required for atherosclerotic lesion pro-gression in ApoE�/� mice. J Clin Invest. 2001;108:1513–1522.

11. Tsai KJ, Hung IJ, Chow CK, Stern A, Chao SS, Chiu DTY. Impairedproduction of nitric oxide, superoxide, and hydrogen peroxide in glucose6-phosphate-dehydrogenase-deficient granulocytes. FEBS Letters. 1998;436:411–414.

12. Pascale R, Garcea R, Ruggiu ME, Daino L, Frassetto S, Vannini MG,Cozzolino P, Lenzerini L, Feo F, Schwartz AG. Decreased stimulation by12-O-tetradecanoylphorbol-13-acetate of superoxide radical production bypolymorphonuclear leukocytes carrying the Mediterranean variant ofglucose-6-phosphate dehydrogenase. Carcinogenesis. 1987;8:1567–1570.

13. Hothersall JS, Gordge M, Noronha-Dutra AA. Inhibition of NADPHsupply by 6-aminonicotinamide: effect on glutathione, nitric oxide andsuperoxide in J774 cells. FEBS Letters. 1998;434:97–100.

14. Leopold JA, Cap A, Scribner AW, Stanton RC, Loscalzo J. Glucose-6-phosphate dehydrogenase deficiency promotes endothelial oxidant stressand decreases endothelial nitric oxide bioavailability. FASEB J. 2001;15:1771–1773.

15. Guo L, Zhang Z, Green K, Stanton RC. Suppression of interleukin-1 beta-induced nitric oxide production in RINm5F cells by inhibition of glucose-6-phosphate dehydrogenase. Biochemistry. 2002;41:14726–14733.

16. Gupte SA, Arshad M, Viola S, Kaminski PM, Ungvari Z, Rabbani G,Koller A, Wolin MS. Pentose phosphate pathway coordinates multipleredox-controlled relaxing mechanisms in bovine coronary arteries. Am JPhysiol Heart Circ Physiol. 2003;285:H2316–H2326.

17. Matsui R, Xu S, Maitland KA, Hayes A, Leopold JA, Handy DE,Loscalzo J, Cohen RA. Glucose-6 phosphate dehydrogenase deficiencydecreases the vascular response to angiotensin II. Circulation. 2005;112:257–263.

18. Pandolfi PP, Sonati F, Rivi R, Mason P, Grosveld F, Luzzatto L. Targeteddisruption of the housekeeping gene encoding glucose 6-phosphate dehy-drogenase (G6PD): G6PD is dispensable for pentose synthesis butessential for defense against oxidative stress. EMBO J. 1995;14:5209–5215.

Matsui et al G6PD Deficiency and Atherosclerosis 915

by guest on July 26, 2015http://atvb.ahajournals.org/Downloaded from

19. Salvemini F, Franze A, Iervolino A, Filosa S, Salzano S, Ursini MV.Enhanced glutathione levels and oxidoresistance mediated by increasedglucose-6-phosphate dehydrogenase expression. J Biol Chem. 1999;274:2750–2757.

20. Leopold JA, Loscalzo J. Cyclic strain modulates resistance to oxidantstress by increasing G-6PDH expression in smooth muscle cells. Am JPhysiol Heart Circ Physiol. 2000;279:H2477–H2485.

21. Jain M, Brenner DA, Cui L, Lim CC, Wang B, Pimentel DR, Koh S,Sawyer DB, Leopold JA, Handy DE, Loscalzo J, Apstein CS, Liao R.Glucose-6-phosphate dehydrogenase modulates cytosolic redox statusand contractile phenotype in adult cardiomyocytes. Circ Res. 2003;93:e9–e16.

22. Jain M, Cui L, Brenner DA, Wang B, Handy DE, Leopold JA, LoscalzoJ, Apstein CS, Liao R. Increased myocardial dysfunction after ischemia-reperfusion in mice lacking glucose-6-phosphate dehydrogenase. Circu-lation. 2004;109:898–903.

23. Wiesenfeld SL, Petrakis NL, Sams BJ, Collen MF, Cutler JL. Elevatedblood pressure, pulse rate and serum creatinine in Negro males deficientin glucose-6-phosphate dehydrogenase. N Engl J Med. 1970;282:1001–1002.

24. Cocco P, Todde P, Fornera S, Bonaria Manca M, Manca P, Sias AR.Mortality in a cohort of men expressing th eglucose-6-phosphate dehy-drogenase deficiency. Blood. 1998;91:706–709.

25. Leopold JA, Walker J, Scribner AW, Voetsch B, Zhang YY, Loscalzo AJ,Stanton RC, Loscalzo J. Glucose-6-phosphate dehydrogenase modulatesvascular endothelial growth factor-mediated angiogenesis. J Biol Chem.2003;278:32100–32106.

26. Pretsch W, Charles DJ, Merkle S. X-linked glucose-6-phosphate dehy-drogenase deficiency in mus musculus. Biochem Genetics. 1988;26:89–103.

27. Sanders S, Smith DP, Thomas GA, Williams ED. A glucose-6-phosphatedehydrogenase (G6PD) splice site consensus sequence mutation asso-ciated with G6PD enzyme deficiency. Mutat Res. 1997;374:79–87.

28. Paigen B, Mitchell D, Reue K, Morrow A, Lusis AJ, LeBoeuf RC. Ath-1,a gene determining atherosclerosis susceptibility and high densitylipoprotein levels in mice. Proc Natl Acad Sci U S A. 1987;84:3763–3767.

29. Wang X, Ria M, Kelmenson PM, Eriksson P, Higgins DC, Samnegard A,Petros C, Rollins J, Bennet AM, Wiman B, de Faire U, Wennberg C,Olsson PG, Ishii N, Sugamura K, Hamsten A, Forsman-Semb K, Lager-crantz J, Paigen B. Positional identification of TNFSF4, encoding OX40ligand, as a gene that influences atherosclerosis susceptibility. Nat Genet.2005;37:365–372.

30. Wang HD, Xu S, Johns DG, Du Y, Quinn MT, Cayatte AJ, Cohen RA.Role of NADPH oxidase in the vascular hypertrophic and oxidative stressresponse to angiotensin II in mice. Circ Res. 2001;88:947–953.

31. Johns C, Gavras I, Handy DE, Salomao A, Gavras H. Models of exper-imental hypertension in mice. Hypertension. 1996;28:1064–1069.

32. Li Y, Zhu H, Kuppusamy P, Roubaud V, Zweier JL, Trush MA. Vali-dation of lucigenin (Bis-N-methylacridinium) as a chemilumigenic probefor detecting superoxide anion radical production by enzymatic andcellular systems. J Biol Chem. 1998;273:2015–2023.

33. Stanton RC, Seifter JL, Boxer DC, Zimmerman E, Cantley LC. Rapidrelease of bound glucose-6-phosphate dehydrogenase by growth factors.Correlation with increased enzymatic activity. J Biol Chem. 1991;266:12442–12448.

34. Umeki S. Prostaglandin E and analogs of prostacyclin influencingsuperoxide production by the human neutrophil NADPH oxidase system.Int J Biochem. 1994;26:1003–1008.

35. Knowles JW, Reddick RL, Jennette JC, Shesely EG, Smithies O, MaedaN. Enhanced atherosclerosis and kidney dysfunction in eNOS�/�Apoe�/�

mice are ameliorated by enalapril treatment. J Clin Invest. 2000;105:451–458.

36. Liu J, Yang F, Yang XP, Jankowski M, Pagano PJ. NAD(P)H oxidasemediates angiotensin II-Induced vascular macrophage Infiltration andmedial hypertrophy. Arterioscler Thromb Vasc Biol. 2003;23:776–782.

37. Montalvetti A, Pena-Diaz J, Hurtado R, Ruiz-Perez LM, Gonzalez-Pacanowska D. Characterization and regulation of Leishmania major3-hydroxy-3-methylglutaryl-CoA reductase. Biochem J. 2000;349:27–34.

38. Muntoni S, Batetta B, Dessi S, Muntoni S, Pani P Serum lipoproteinprofile in the Mediterranean variant of glucose-6-phosphate dehydro-genase deficiency. Eur J Epidemiol. 1992;8 Suppl 1:48–53.

39. Smith JD, James D, Dansky HM, Wittkowski KM, Moore KJ, BreslowJL. In silico quantitative trait locus map for atherosclerosis susceptibilityin apolipoprotein E-deficient mice. Arterioscler Thromb Vasc Biol. 2003;23:117–122.

40. Reardon CA, Blachowicz L, Lukens J, Nissenbaum M, Getz GS. Geneticbackground selectively influences innominate artery atherosclerosis:immune system deficiency as a probe. Arterioscler Thromb Vasc Biol.2003;23:1449–1454.

41. Won JS, Im YB, Key L, Singh I, Singh AK. The involvement of glucosemetabolism in the regulation of inducible nitric oxide synthase geneexpression in glial cells: possible role of glucose-6-phosphate dehydro-genase and CCAAT/enhancing binding protein. J Neurosci. 2003;23:7470–7478.

42. Khan BV, Harrison DG, Olbrych MT, Alexander RW. Nitric oxideregulates vascular cell adhesion molecule 1 gene expression and redox-sensitive transcriptional events in human vascular endothelial cells. ProcNatl Acad Sci U S A. 1996;93:9114–9119.

43. Cayatte AJ, Palacino JJ, Horten K, Cohen RA. Chronic inhibition of nitricoxide production accelerates neointima formation and impairs endothelialfunction in hypercholesterolemic rabbits. Arterioscler Thromb. 1994;14:753–759.

44. Xia Y, Roman LJ, Masters BS, Zweier JL. Inducible nitric-oxide synthasegenerates superoxide from the reductase domain. J Biol Chem. 1998;273:22635–22639.

45. Kuhlencordt PJ, Chen J, Han F, Astern J, Huang PL. Genetic deficiencyof inducible nitric oxide synthase reduces atherosclerosis and lowersplasma lipid peroxides in apolipoprotein E-knockout mice. Circulation.2001;103:3099–3104.

46. Detmers PA, Hernandez M, Mudgett J, Hassing H, Burton C, Mundt S,Chun S, Fletcher D, Card DJ, Lisnock J, Weikel R, Bergstrom JD, ShevellDE, Hermanowski-Vosatka A, Sparrow CP, Chao YS, Rader DJ, WrightSD, Pure E. Deficiency in inducible nitric oxide synthase results inreduced atherosclerosis in apolipoprotein E-deficient mice. J Immunol.2000;165:3430–3435.

47. Behr-Roussel D, Rupin A, Simonet S, Bonhomme E, Coumailleau S,Cordi A, Serkiz B, Fabiani JN, Verbeuren TJ. Effect of chronic treatmentwith the inducible nitric oxide synthase inhibitor N-iminoethyl-L-lysineor with L-arginine on progression of coronary and aortic atherosclerosisin hypercholesterolemic rabbits. Circulation. 2000;102:1033–1038.

48. Bea F, Hudson FN, Chait A, Kavanagh TJ, Rosenfeld ME. Induction ofglutathione synthesis in macrophages by oxidized low-densitylipoproteins is mediated by consensus antioxidant response elements.Circ Res. 2003;92:386–393.

49. Rahman I, Macnee W. Oxidative stress and regulation of glutathione inlung inflammation. Eur Respir J. 2000;16:534–554.

50. Ruwende C, Khoo SC, Snow RW, Yates SN, Kwiatkowski D, Gupta S,Warn P, Allsopp CE, Gilbert SC, Peschu N. Natural selection of hemi-and heterozygotes for G6PD deficiency in Africa by resistance to severemalaria. Nature. 1995;376:246–249.

916 Arterioscler Thromb Vasc Biol. April 2006

by guest on July 26, 2015http://atvb.ahajournals.org/Downloaded from

Supplement

Methods

Background strain characterization by microsatellite markers at Ath-1 locus. Mouse

DNA was prepared from frozen mouse kidney using Clontech Nucleospin Tissue Kit.

Ath-1, a locus that carries the increased susceptibility trait for the C57Bl6 strain, is

located at 84.9 cM on chromosome 1 between markers Chr1-5 (81.6 cM) and Chr1-6

(94.2 cM). The analysis was done for these 2 markers for 10 mice in each group to

determine if the region was homozygous (100% C57BL6) or heterozygous (50%

C57BL6). Each PCR reaction (40 cycles) was analyzed on agarose gel. All procedures

were performed at Charles River Genetic Testing Services (Troy, NY).

Determination of tnfsf4 mRNA expression. Total RNA was prepared with TRIzol

reagent (Invitrogen) from frozen mouse heart. cDNA was produced using TaKaRa RNA

PCR kit with AMV reverse transcriptase XL (Takara Bio). Real-time PCR was

performed on iCycler iQ Real-Time PCR Detection System (Bio-Rad) using iQTMSYBR

Green Supermix (Bio-Rad) according to the manufacturer’s instruction. Primers for

mouse Tnfsf4 and mouse β-actin were purchased from SuperArray.

Results

Background strain characterization by microsatellite markers. . Ten randomly

chosen mice from each group were used for analysis of the Ath-1 locus. Of the 10 mice

chosen from both groups, 5 mice were 100% C57BL6 and 5 mice were 50% C57BL6

background (C57Bl/C3H). When aortic lesion was compared among the mice with 100%

C57BL6 Ath-1 locus, lesion area was still significantly greater in E-WT than in E-Hemi

mice (P < 0.01, Supplemental Figure 2). These data indicate that the lesser lesion

development in E-Hemi mice did not result from any potential difference in the C57BL6

Ath-1 locus.

Determination of tnfsf4 mRNA expression. Tnfsf4 copy number was standardized by

β-actin copy number. There was no significant difference in the tnfsf4 mRNA expression

in the heart between E-WT and E-Hemi.

Table I. Body weight, tissue weight, and serum cholesterol level in apo E-/- mice.

G6PD genotype BW (g) Heart

(mg)

HW/BW Liver

(mg)

Serum Cholesterol

(mg/dl)

WT (n = 27) 48.4 + 0.9 160 +3 3.36 + 0.07 3635 + 216 1756 + 82

Hemi (n = 28) 49.2 + 0.8 153 +3 3.16 + 0.06 3876 + 199 1567 + 57 *

Table I. Body weight and tissue weight and serum cholesterol were measured when mice

were sacrificed at age 21 weeks after being fed western diet for 11 weeks. Data are

expressed as mean + SEM. * P <0.05 vs WT.

Legends

Figure I. Breeding protocol by which E-WT and E-Hemi mice were obtained. E-WT

and E-Hemi mice in the study were obtained from the F1’ and F2 generation and were

predicted to have one quarter C3H and three quarters of the C57BL/6 genetic background.

Figure II. Genetic background of Ath-1 locus and atherosclerotic lesions. Ten

randomly chosen from each group were analyzed for Ath-1 locus, and mice were divided

into two groups according to the percentage of C57BL/6 (50% or 100%). The aortic

lesion was compared between E-WT and E-Hemi in each group. In the groups having

100% C57BL/6 Ath-1 locus, the lesion was significantly less in E-Hemi (739 + 174)

compared to E-WT (2128 + 411) (X103 µm2, P < 0.01). In the group having 50%

C57BL/6 Ath-1 locus, the lesion was also less in E-Hemi (916 + 230), but not

significantly different from E-WT (1395 + 352).

Figure III. tnfsf4 mRNA expression in mouse heart. Expression of mRNA of tnfsf4

and β-actin was analyzed by quantitative PCR using cDNA from mouse heart as

described in Methods. The ratio of copy number (tnfsf4/β-actin) was not significantly

different between groups. Data are mean + SE. (E-WT 3.2 + 1.9, E-Hemi 5.4 + 2.1, n =

8)

Figure IV. G6PD activity and expression in western diet-fed apo E-/- mouse aorta.

(a) G6PD activity was expressed in arbitary units/mg protein (mean + SEM, n= 4, * P <

0.01). (b) Expression of G6PD in aorta detected by Western blot.

Figure I: Breeding protocol to obtain E-WT and E-Hemi mice

G6PD Hemi ApoE KO Male Female

(C3H) (C57BL/6J)

EE XmY ee XX

Ee XmX ee XY

ApoE KOMale

(C57BL/6J)

F0

F1

F1’ ee XmX ee XmY ee XX ee XY

ee XmY ee XY ee XmXm ee XmX (E-Hemi) (E-WT)

(E-WT)(E-Hemi)

F2

EE: apoE WT, ee: apoE KO, Xm: G6PD mutant X chromosome

Figure II. Ath-1 locus microsatellite analysis and lesions

* P < 0.01

0

1000

2000

3000

4000

50 100

Ath-1 locus (%C57Bl/6)

*aort

ic le

sion

(x10

3m

m2 )

E-WTE-Hemi

Figure III: Tnfsf4 expression in mouse heart

56

Tnf

sf4/

b -ac

tin, c

op

1234

789

E-WT E-Hemi

y nu

mbe

r ra

tioNS

Figure IV(a): G6PD activity in mouse aorta

G6P

D a

ctiv

ity (a

rb

0

200

400

600

800

1000

E-WT

U /

mg

prot

ein)

1200

*

E-Hemi

Figure IV (b): G6PD expression in mouse aorta

G6PD

α-actin

E-WT E-Hemi

59 Kd

44 Kd

E. Handy, Joseph Loscalzo and Richard A. CohenReiko Matsui, Shanqin Xu, Karlene A. Maitland, Roberto Mastroianni, Jane A. Leopold, Diane

Mice−/−Atherosclerotic Lesions in Apolipoprotein E Glucose-6-Phosphate Dehydrogenase Deficiency Decreases Vascular Superoxide and

Print ISSN: 1079-5642. Online ISSN: 1524-4636 Copyright © 2006 American Heart Association, Inc. All rights reserved.

Greenville Avenue, Dallas, TX 75231is published by the American Heart Association, 7272Arteriosclerosis, Thrombosis, and Vascular Biology

doi: 10.1161/01.ATV.0000205850.49390.3b2006;26:910-916; originally published online January 26, 2006;Arterioscler Thromb Vasc Biol. 

http://atvb.ahajournals.org/content/26/4/910World Wide Web at:

The online version of this article, along with updated information and services, is located on the

http://atvb.ahajournals.org/content/suppl/2006/01/31/01.ATV.0000205850.49390.3b.DC1.htmlData Supplement (unedited) at:

  http://atvb.ahajournals.org//subscriptions/

at: is onlineArteriosclerosis, Thrombosis, and Vascular Biology Information about subscribing to Subscriptions:

  http://www.lww.com/reprints

Information about reprints can be found online at: Reprints: 

document. Question and AnswerPermissions and Rightspage under Services. Further information about this process is available in the

which permission is being requested is located, click Request Permissions in the middle column of the WebCopyright Clearance Center, not the Editorial Office. Once the online version of the published article for

can be obtained via RightsLink, a service of theArteriosclerosis, Thrombosis, and Vascular Biologyin Requests for permissions to reproduce figures, tables, or portions of articles originally publishedPermissions:

by guest on July 26, 2015http://atvb.ahajournals.org/Downloaded from