Endothelial Mechanism in the Vascular Actionof Hydralazine

ERIC G. SPOKAS, P H . D . , GIANCARLO FOLCO, P H . D . , JOHN QUILLEY, P H . D . ,

PRAVEEN CHANDER, M.D., AND JOHN C. MCGIFF, M.D.

SUMMARY Relaxation of precontracted isolated chains of aortic rings with intact endothelium andin those with the endothelium removed was studied in response to various antihypertensive vasodila-tor drugs. Of the drugs tested — nitroprusside, nitroglycerin, prazosin, minoxidil, diazoxide andhydralazine — only the vascular relaxant effects of hydralazine were found to be dependent, in part,on the presence of intact endothelium. The endothelial component of the hydralazine responserepresented a major contribution to the net relaxant effect on the vascular smooth muscle, particularlyat lower concentrations, 90 nM to 1 /xM, which are also clinically relevant.(Hypertension 5 (supp I): I-107-M11, 1983)

KEY WORDSprostacyclin

endotheliumvasod ilatation

hydralazine • antihypertensive drugs

OF the antihypertensive drugs classified as "va-sodilators", hydralazine (1-hydrazinophthal-azine) is one of the most widely used for the

long-term management of hypertension.' The primarysite of action of hydralazine is generally accepted to bethe vascular wall, as low doses administered arteriallycause local vasodilation of gradual onset.2 Althoughthe basis for the direct vasodilator effect of hydralazinehas been studied extensively for three decades, themechanism has not been established. Part of the diffi-culty associated with efforts to define the underlyingbiochemical changes has been the variable in vitroactivity of the drug.3"7 In studies with isolated bloodvessels, millimolar concentrations may be required toevoke relaxation, even when superimposed on a strongcontraction induced by norepinephrine (NE).

The vascular endothelium has been shown by Furch-gott and Zawadzki8-9 to be crucial for demonstration ofrelaxation of blood vessels in response to acetylcholine(ACh) and some other naturally occurring vasodilatorsubstances. More to the point, the failure of ACh tocause relaxation of isolated blood vessels in earlierstudies was an artifact arising from unintentional re-

From the Department of Pharmacology, New York Medical Col-lege, Valhalla, New York, and the Cattedra Di Farmacologia E DiFarmacognosia, Universita Di Milano, Milano, Italy.

Supported by the US-Italy Scientific Exchange, sponsored byInternational Programs, National Heart, Lung, and Blood Institute,and through NIH Grants HL-25394, HL-25406 and CA-27893 fromthe National Institutes of Health. Dr. Spokas is a recipient of a NewInvestigator Research Award, HL-27967, from the National Heart,Lung, and Blood Institute.

Address for reprints: Dr. Eric Spokas, Department of Pharmacol-ogy, New York Medical College, Valhalla, New York 10595.

moval of the intima attendant to preparation of vascu-lar rings or strips for tension recording. Besides resolv-ing a long-standing controversy as to the anomalousvascular effect of ACh in vitro, the results from Furch-gott and Zawadzki's laboratory suggested to us thepossibility of an endothelial component to the vasodi-lator effect of hydralazine, which might explain thelarge doses required to relax isolated blood vessels.We report here the effects of the removal of endothelialcells on the mechanical responses of aortic rings tohydralazine and to other vasodilator/antihypertensiveagents. The effectiveness of the denuding proceduredesigned to remove the endothelium was assessedfunctionally from the response to ACh as well as histo-logically, using light microscopy.

MethodsTransverse rings were obtained from the descending

thoracic aorta of male New Zealand white rabbits.Endotheliurn was removed by apposition of the intimalsurfaces and gentle rubbing between the fingers forapproximately 30 seconds. Four to six rings werejoined together with surgical silk (3.0) to form a chainand placed in a 15 ml glass organ bath containingKrebs' Ringer bicarbonate solution at 37°C, aeratedwith a mixture of 95% O2 and 5% CO2. For eachexperiment, four chains of rings from the same aorta,consisting of two rubbed and two unrubbed chainswere studied. Basal tension (2 g) was applied follow-ing an equilibration period of at least 2 hours, and thechanges in isotonic contraction were monitored withModel 356 Harvard transducers and a model 1244 Sol-tec recorder. Responses to various vasodilator agentswere studied during enhancement of vascular tone with

1-107

by guest on February 12, 2018http://hyper.ahajournals.org/

Dow

nloaded from

1-108 1982 BLOOD PRESSURE COUNCIL SUPP I, HYPERTENSION, VOL 5, No 2, MARCH-APRIL 1983

a submaximal concentration of NE (10 or 20 nM)added in the presence of 35 juM ascorbic acid.

ACh was tested during an initial contracture evokedby NE. The bath was rinsed, and the Krebs mediumreplaced; 30 minutes later, tone was again increasedwith NE. Hydralazine was then added in a cumulativefashion at intervals of 12-15 minutes. Subsequently,the tissues were rinsed three times, and 30-45 minuteslater the relaxant response to ACh was retested duringa third NE-induced contracture. In some experimentsrabbits were pretreated with indomethacin, 5 mg/kgi.v., 30 minutes before sacrifice; indomethacin (10fiM) was also added to the Krebs medium in the organbath.

In some experiments immunoreactive PGE2 wasmeasured by radioimmunoassay (RIA). RIA of theunextracted incubates was carried out in duplicate, thelower limit of sensitivity for this assay was 60 pg/ml.PGE2 antiserum was obtained from Institut PasteurProduction, Paris, France. The cross-reactivities of theantiserum were 3.2% for PGE,, 2% for 6-keto-PGFla,0.2% for PGA2 0.15% for 13,14-dihydro-PGE2,0.11% for 13,14-dihydro-15-keto-PGE2 and less than0.1% forPGF^, PGA,, 13,14-dihydro-PGE,, 13,14-dihydro-15-ketq-PGE, and PGF]a.

The effectiveness of the denuding procedure de-signed to remove the endothelium was assessed func-tionally from the response to ACh as well as histologi-cally, using light microscopy. Chains of aortic ringswere placed in petri dishes containing Krebs' bufferand separated into individual rings. Each ring was thenopened with fine dissecting scissors, forming a stripwhich was pinned at the corners to dental wax. Controland experimental tissues, rubbed and unrubbed, weresubjected to a silver-staining technique modified fromthat described by Poole et al.,10 followed by formalinfixation and mounting on glass slides without embed-ding. Percent endothelium remaining was determinedsemi-quantitatively by en face single blind light micro-scopic examination of the entire intimal surface of therings. The rubbing procedure resulted in almost com-plete loss of endothelial cells.

The composition (mM) of the Krebs buffer was:NaCl 118, KC1 4.8, CaCl2 2.5, MgSO4 1.2, KH2PO4

1.2, NaHCOj 24, glucose 9.6; pH was 7.4. The fol-lowing drugs were used: hydralazine HC1 (Apresoline,Ciba-Geigy, Summit, New Jersey), sodium nitroprus-side (Nipride, Hoffmann-LaRoche, Nutley, New Jer-sey), nitroglycerin (Parke:Davis, Detroit, Michigan),minoxidil (Upjohn, Kalamazoo, Michigan), diazoxide(Schering, Bloomfield, New Jersey), prazosin HC1Pfizer, New York, New York), 1-norepinephrine bitar-trate (Levophed, Winthrop, New York, New York),BW 755C (Wellcome Research Laboratories, Beckeh-ham, Kent, England), 1-ascorbic acid, acetylcholinechloride, indomethacin and nordihydroguaiaretic acid(NDGA) (Sigma Chemical Company, St. Louis, Mis-souri). Drug solutions were prepared on the day of theexperiment, stored on ice, and added to the tissue bathin a volume of 120-180 fil. A stock solution of hydra-lazine HC1 (crystalline form) was prepared in distilled

water (1 mg base/ml) and diluted further with Krebssolution. A stock solution of diazoxide (43 mM) wasprepared in NaOH (0.4 M) and subsequently dilutedusing Krebs buffer. Indomethacin was solubilized in0.9% saline containing 0.1% NaCO3 (final concentra-tion indomethacin = 3.57 mg/ml, pH 8.5). This stocksolution was used undiluted for intravenous pretreat-ment of the animals (5 mg/kg) 30 minutes before sacri-fice and diluted to 10 ^iM indomethacin with Krebsmedium for treatment in the tissue bath. A stock solu-tion of 160 mM NDGA was prepared in concentratedsulfuric acid and diluted with Krebs solution, adjustingthe pH to 7.9 with a small volume of NaOH (5 M).

Individual dose-response curves were linearized byplotting on semilog paper after probit transformation."For the probit analysis, responses greater than 99%loss of NE-induced tone were arbitrarily assigned avalue of 99%. The data from each chain were used forthe calculation of mean responses by fitting with indi-vidual regression lines between probits 3 and 7 (leastsquares method for log curves). Mean probits, whichwere thereby equally-weighted, were plotted at select-ed concentrations to obtain the curves shown here. Inmost experiments, three or four concentrations of hy-dralazine were tested on each chain. IDjo and ID10

values were obtained from regression lines fitted withlog curves by the least square method. Geometricmeans are presented as the statistical comparisonswere carried out on the log ID values. Students' un-paired t test was used for these comparisons.

ResultsResponses to concentrations of NE in the range of 1

nM to 10 fiM were examined in preliminary studiescarried out on six rubbed and six unrubbed chains.Rubbing did not affect the NE dose-response curve,either in terms of slope, threshold concentration ormagnitude of the peak response. ACh, tested during aninitial contracture evoked by NE, invariably relaxedthe unrubbed, but not the rubbed rings (fig. 1). Ex-pressed as percent inhibitionof NE-induced tone, re-laxation in response to 200 nM ACh averaged 57% ±4% in unrubbed chains, whereas in rubbed chains ei-ther no response or a small contraction ( - 2% ± 2%)occurred. At the completion of the experiment duringthe third NE contraction, ACh response was found tobe somewhat reduced (mean inhibition of NE-inducedtone = 41% ± 9%), but not significantly (p > 0.05)when compared to the initial response using a pairedt test.

Isolated aortic chains with intact endothelium werevery sensitive to hydralazine. From the cumulativedose-response relationships shown in figures 1 and 2,it is evident that concentrations below 300 nM wereeffective in antagonizing contractions induced by NE.The hydralazine response of unrubbed chains wascharacterized by a gradual loss of tone which, at lowconcentrations of hydralazine, was sometimes preced-ed by a latent period of several minutes. Rubbingcaused a reduction in the sensitivity of the tissue tohydralazine as indicated by the movement of the dose-

by guest on February 12, 2018http://hyper.ahajournals.org/

Dow

nloaded from

ENDOTHELIAL DEPENDENCY OF HYDRALAZINE/Spofcas et al. 1-109

response curve to the right, although the maximal re-laxant effect was comparable in rubbed and unrubbedchains (fig. 2). The IDK, defined as the concentrationof hydralazine which caused 50% inhibition of NE-induced tone, was determined for individual experi-ments where two to five concentrations were tested.Higher ID,,, values were found in the de-endothelia-lized chains (geometric mean 1.405 /xM rubbed, ver-sus 0.507 fiM unrubbed, p > 0.001). An increase inthe threshold concentration for the hydralazine re-

5mln

NITROGLYCERIN

N.E.IOnM

FIGURE 1. Representative tracings showing the effects oface-tylcholine and various antihypertensive drugs on vascular toneof rabbit aorta, having either intact endothelium (I) or endothe-lium damaged by rubbing (R). Each preparation consisted of aseries of four to 6 interconnected transverse rings, bathed withKrebs medium at 37°C. Changes in isotonic tension in responseto relaxing agents were studied during enhancement of vasculartone with 10 nM norepinephrine (NE). At times indicated by thesmall arrows, each test agent was added to two organ baths,containing either a rubbed or an intact preparation from thesame aorta. The increase in concentration in the bath is shownnext to the small arrows. Rubbing of the intimal surfaces invari-ably resulted in abolition or reversal of acetylcholine-inducedaortic relaxation. In contrast to the ineffectiveness of the rub-bing procedure in altering the aortic response to minoxidil andnitroglycerin, the relaxant response to hydralazine was muchmore pronounced in aortic rings with intact endothelium.

sponse, due to rubbing, was suggested by a significantincrease in the ID10 to 180 nM (rubbed) from 91nM (unrubbed), p < 0.05. In contrast to these resultswith hydralazine, the rubbing procedure did not alterthe response of NE-precontracted aortic chains toother vasodilator test drugs including sodium nitro-prusside, nitroglycerin, minoxidil, diazoxide and pra-zosin (table 1).

Histologic examinations (carried out single blind)confirmed the effectiveness of the rubbing procedurein removing endothelium. Rubbed rings from chainswhich did not relax in response to ACh evidencedalmost complete loss of endothelial cells (fig. 3 upperright). Estimates for percent remaining ranged from 0to 1% of the total intimal surface (n = 3 chains, 12individual strips). Estimates for percent endotheliumremaining on unrubbed chains which relaxed rangedfrom 15% to 65% of the total intimal surface (fig. 3lower left) (n = 3 chains, 12 individual strips). Fur-ther, studies of unrubbed control aortic rings notmounted for tension recording, but stained and fixedfor microscopy directly after sectioning the aorta withfine scissors, indicated that a large fraction of endothe-lium, 10% to 40% (n = 19), was routinely lost throughsurgical isolation and/or sectioning (fig. 3 upper left).

In experiments designed to investigate the possiblerole of prostaglandins, unrubbed aortic chains weretreated with the cyclooxygenase inhibitor, indometha-cin. Cyclooxygenase inhibition was verified by RIAanalysis of basal PGE2 released into the incubate over a1-hour period. Mean control levels were 104 ± 1 1 pg/ml(n = 4); after indomethacin treatment immunoreac-tive PGE2 was not detected, i.e., < 60 pg/ml. Therelaxant effects of hydralazine (60 and 600 nM) were

HYDRALAZINE

3 1 3 K) 30

FIGURE 2. Dose-response relationships for the relaxing ac-tion of hydralazine on norepinephrine-precontracted chains oftransverse rings of rabbit descending thoracic aorta. Percentinhibition of norepinephrine-induced tone is plotted as a func-tion of hydralazine concentration using a log scale for theabscissa. Various concentrations of hydralazine were tested onuntreated control rings having intact endothelium (open circles,n = 11 chains) and on rings which were rubbed on the intimalsurface to remove endothelium (closed circles, n = 14 chains).

by guest on February 12, 2018http://hyper.ahajournals.org/

Dow

nloaded from

1-110 1982 BLOOD PRESSURE COUNCIL SUPP I, HYPERTENSION, VOL 5, No 2, MARCH-APRIL 1983

TABLE 1. Relaxing Action of Drugs on Chains of Rabbit AorticRings Having Either Intact Endothelium or Endothelium Damagedby Rubbing

Dnig

nitroprusside

nitroglycerin

minoxidil

diazoxide

prazosin

Concen-tration(MM)

0.0060.0240.084

0.0060.0240.084

0.0600.6606.67

0.0600.6606.67

0.060

Intact*(% inhi-bition)

20±351±682±7

29 ±364 + 688 + 5

12 + 432+1046+15

8 + 225 + 448 + 6

80 + 6

(4)(4)(4)

(5)(5)(5)

(5)(5)(4)

(4)(4)(4)

(4)

Rubbed*(% inhi-bition)

24 + 264±789±6

30 + 368 + 489 + 2

12±332 ±843±11

6±222±145 ±3

77±6

(4)(4)(4)

(5)(5)(5)

(5)(5)(4)

(4)(4)(4)

(4)

•Relaxation is expressed as percent inhibition (mean ± SE) oftone induced by 10 or 20 nM norepinephrine and was measured 5minutes (nitroprusside, nitroglycerin, prazosin) or 20 minutes (min-oxidil, diazoxide) after addition to the bath. These intervals werethe respective cycle times between additions. All additions weremade during the sustained phase of the norepinephrine contractureand the concentrations given are cumulative. Number of chains isshown in parentheses.

not affected by indomethacin treatment. Indomethacinalso did not influence the relaxant effect of ACh onunrubbed chains.

Responses to hydralazine were tested in the pres-ence of BW 755C (n = 2) (100-400 /J.M) or NDGA(n = 2) 20 to 50 /iM, agents reported to inhibit thelipoxygenase pathway for arachidonic acid metabo-lism.12- l3 For these experiments, the effect of the inhib-itor was studied on paired preparations from the sameaorta (one rubbed chain and one with intact endotheli-um) by addition (BW 755C) after hydralazine hadcaused approximately 80% relaxation, or by pretreat-ment (NDGA) for at least 30 minutes before cumula-tive additions of hydralazine. The relaxant effect of hy-dralazine on unrubbed and rubbed chains was not re-duced by treatment with these lipoxygenase inhibitors.ACh-induced relaxation, studied using a similar ex-perimental design, also was not antagonized (n = 5).

DiscussionWe conclude that the rabbit aortic preparation is

sensitive to hydralazine if care is taken to preserve theendothelium. Our results contrast with those reportedin a number of previous in vitro studies of rabbit aor-ta.3"6 The high degree of responsiveness of aortic rings

*?. *v -•*

FIGURE 3. Photomicrographs of silver-stained intimalsurfaces of aortic ring preparations. Upper Left: Controlrings not mounted for tension recording, x 130 (inset X325); Upper Right: Rings rubbed to remove endothelium.X 130. Lower Left: Unrubbed rings following an organbath experiment, x 170 (inset x 325).

by guest on February 12, 2018http://hyper.ahajournals.org/

Dow

nloaded from

ENDOTHELIAL DEPENDENCY OF HYDRALAZINE/Spofaw et al. 1-111

used in our study, even when rubbed, derives fromseveral factors: 1) the use of chains of aortic ringsrather than individual rings or spiral strips; 2) record-ing isotonic rather than isometric contractions; 2) pre-contraction of the aortic rings with NE rather thanpotassium. It is noteworthy that at concentrationsthought to be attained in the plasma of patients receiv-ing hydralazine for hypertension (1 pM or less),14 theendothelial component of the hydralazine responseconstitutes a major fraction. It may be significant thatthe reduction of vascular tone produced by hydralazinein organ bath studies is not rapid in onset but character-istically requires 10 minutes or more to develop fully.A similar time course has been described for the vascu-lar resistance changes induced by hydralazine invivo," a further indication that our in vitro finding ofendothelial dependence may have relevance to the lo-cal vascular mechanism activated in the intact animal.

The contribution of endothelial cells to the aorticrelaxing action of hydralazine may be a more generalphenomenon, similar to that reported for ACh,8•' asseveral large arteries of the dog — aorta, pulmonaryand renal arteries — demonstrated blunting of hydrala-zine-induced relaxation after removal of the endotheli-um (unpublished observations). Our results do not per-mit a statement as to the nature of the endothelialcontribution to the relaxing action of hydralazine. Itdoes not appear to be prostaglandin-mediated inhibi-tion of cyclooxygenase with indomethacin did not in-terfere with the effect of the drug. Neither did BW755C or NDGA, agents which have been described toinhibit lipoxygenases.12 " It should be noted that in-hibitors of lipoxygenases have been reported to attenu-ate ACh-induced relaxation of the femoral artery,16 aneffect that may depend on the blood vessel used, theeffective concentration of the inhibitor and the sensi-tivity of the experimental preparation to agonists. Aproduct of phospholipase activity such as an arachi-donate metabolite has been suggested to account forthe endothelial contribution to vasoactive agents.1617

There are two additional considerations that bear onthe interpretation of our study. First, as active metabo-lites of hydralazine are known,' it is possible that hy-dralazine is transformed to a more active material bythe endothelium. This explanation, however, lackssupport as none of the known active metabolites ismore potent than the parent compound,5 although ametabolite may possess physiochemical properties thatfavor vascular relaxation under these experimentalconditions, e.g., facilitated access to active sites in thevasculature.

The endothelial independence of minoxidil-inducedaortic relaxation was unexpected because of chemicalsimilarities of minoxidil and hydralazine as well assimilarities in their circulatory effects and affinity forvascular tissue.1 However, the vascular smooth musclecomponents for both antihypertensive agents mayshare a common mechanism.

Finally, the endothelial component of the aortic re-sponse to hydralazine accounts for only a portion, al-beit large at low doses, of the net relaxant effect of

hydralazine on vascular smooth muscle. The parallelshift in the dose response curve for hydralazine sug-gests subtraction of a constant factor through loss ofthe endothelium. The possibility should be consideredthat elements of the vasa vasorum present in the vascu-lar rings may also act as a source of this factor. Thenature of this factor awaits definition.

AcknowledgmentsWe thank Dr. Sylvan Wallenstein, Assistant Professor of Public

Health, College of Physicians and Surgeons, Columbia University,for assistance with statistical analysis. We are grateful to PatriciaHejny, Giuseppe Rossoni, and Kathy Petnllo for their technicalassistance and to Donna Centi and Sallie McGiff for preparing themanuscript The authors also thank Dr. Eugene Weiss of PfizerLaboratories for supplying prazosin.

References1. Gross F: Drugs acting on artenolar smooth muscle (vasodilator

drugs). In Antihypertensive Agents, edited by Gross F. Hei-delberg: Springer-Verlag, 1977, p 397

2. Ablad B: A study of the mechanism of the hemodynamiceffects of hydralazine in man. Acta Pharmacol Toxicol 20: 7,1963

3. KirpekarSM, Lewis JJ: Pharmacological properties of hydral-lazine, dihydrallazine and some related compounds. J PharmPharmacol 9: 877, 1957

4. McLean AJ, Barren K, du Souich P, Haegele KD, McNay JL,Carrier O, Briggs A: Interaction of hydralazine and hydrazonederivatives with contractile mechanisms in rabbit aorticsmooth muscle. J Pharmacol Exp Ther 205: 418, 1978

5. Haegele KD. McLean AJ, du Souich P, Barren K, Laquer J,McNay JL, Carrier O: Identification of hydrallazine and hy-drallazine hydrazone metabolites in human body fluids andquantitative in vitro comparisons of their smooth muscle relax-ant activity. Br J Pharmacol 5: 489, 1978

6. Worcel M: Relationship between the direct inhibitory effects ofhydralazine and propildazine on arterial smooth muscle con-tractility and sympathetic innervation. J Pharmacol Exp Ther207: 320, 1978

7. Khayyal M, Gross F, Kreye VAW: Studies on the direct vaso-dilator effect of hydralazine in the isolated rabbit renal artery. JPharmacol Exp Ther 216: 390, 1981

8. Furchgott RF, Zawadzki JV: The obligatory role of endothelialcells in the relaxation of arterial smooth muscle by acetylcho-line. Nature 288: 373, 1980

9. Cherry PD, Furchgott RF, Zawadzki JV, Jothianandan D: Roleof endothelial cells in relaxation of isolated arteries by brady-kinin. Proc Natl Acad Sci 79: 2106, 1982

10. Poole JCF, Sanders AG, Florey HW: The regeneration ofaortic endothelium. J Path Bact 75: 133, 1958

11. Finney DJ: Probit Analysis, 3rd ed. Cambridge: CambridgeUniversity Press, 1971

12. Higgs GA, Copp FC, Denyer CV, Rower RJ, Tateson JE,Vane JR, Walker JMC: Reduction of leukocyte migration by acyclooxygenase and lipoxygenase inhibitor. In Proceedings ofthe Seventh International Congress of Pharmacology. Oxford:Pergamon Press, 1978, p 334

13. Tappel AL, Lundberg WO, Boyer PD: Effect of temperatureand antioxidants upon the lipoxidase-catalyzed oxidation ofsodium linoleate. Arch Biochem Biophys 42: 293, 1953

14. Israili ZH, Dayton PG: Metabolism of hydralazine. Drug Me-tabolism Reviews 6: 283, 1977

15. Spokas EG, Wang HH: Regional blood flow and cardiac re-sponses to hydralazine. J Pharmacol Exp Ther 212: 294, 1980

16. De Mey JG, Claeys M, Vanhoutte PM: Endothelium-depend-ent inhibitory effects of acetylcholine, adenosine triphosphate,thrombin and arachidonic acid in the canine femoral artery. JPharmacol Exp Ther 222: 166, 1982

17. Singer HA, Peach MJ: Calcium- and endothehal-mediated vas-cular smooth muscle relaxation in rabbit aorta. Hypertension 4(supplll): 11-19, 1982

by guest on February 12, 2018http://hyper.ahajournals.org/

Dow

nloaded from

E G Spokas, G Folco, J Quilley, P Chander and J C McGiffEndothelial mechanism in the vascular action of hydralazine.

Print ISSN: 0194-911X. Online ISSN: 1524-4563 Copyright © 1983 American Heart Association, Inc. All rights reserved.

is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231Hypertension doi: 10.1161/01.HYP.5.2_Pt_2.I107

1983;5:I107Hypertension. 

http://hyper.ahajournals.org/content/5/2_Pt_2/I107World Wide Web at:

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

  http://hyper.ahajournals.org//subscriptions/

is online at: Hypertension Information about subscribing to Subscriptions: 

http://www.lww.com/reprints Information about reprints can be found online at: Reprints:

  document. Permissions and Rights Question and Answer process is available in the

Request Permissions in the middle column of the Web page under Services. Further information about thisOffice. Once the online version of the published article for which permission is being requested is located, click

can be obtained via RightsLink, a service of the Copyright Clearance Center, not the EditorialHypertension Requests for permissions to reproduce figures, tables, or portions of articles originally published inPermissions:

by guest on February 12, 2018http://hyper.ahajournals.org/

Dow

nloaded from