Increased levels of cyclic adenosine 3',5' monophosphate in human polymorphonuclear leukocytes after surface stimulation

Increased levels of cyclic

adenosine-3',5'-monophosphate in human polymorphonuclear

leukocytes after surface stimulation.

J E Smolen, … , H M Korchak, G Weissmann

J Clin Invest. 1980;65(5):1077-1085. https://doi.org/10.1172/JCI109760.

Levels of cyclic AMP (cAMP) (but not cyclic GMP) in suspensions of human

polymorphonuclear leukocytes (PMN) increased promptly after exposure of the cells to stimuli such as the chemotactic peptide N-formyl methionyl leucyl phenylalanine, the

immune complex bovine serum albumin/anti-bovine serum albumin and calcium ionophore A23187. cAMP increased rapidly, reaching a maximum of twice the basal level 10--45 s after stimulation; after 2--5 min the amount of cAMP had subsided to basal levels.

Elevations in cAMP levels were concurrent with, or followed, membrane hyperpolarization (measured by uptake of the lipophilic cation triphenylmethyl phosphonium) and always preceded lysosomal enzyme release and superoxide anion (O2) production. Elevated

cAMP levels could be uncoupled from these later events by removal of extracellular divalent cations, replacement of extracellular Na+ with K+ or choline+, and by use of low

concentrations of stimulus; each of these conditions virtually abolished lysosomal enzyme release and O2 generation, while leaving the stimulated elevation of cAMP levels

unimpaired. Calcium ionophore A23187 did not provoke membrane hyperpolarization, thus uncoupling changes in membrane potential from changes in cAMP levels. These data suggested that cAMP is not a critical component in the earliest steps of stimulus-secretion coupling. Surface stimulation of cells pretreated with prostaglandins E1 or I2 yielded very high levels of cAMP; these high levels may be an important part of the mechanism […]

Research Article

Find the latest version:

Increased Levels of Cyclic

Adenosine-3',5'-Monophosphate

in

Human Polymorphonuclear

Leukocytes after

Surface Stimulation

JAMES E. SMOLEN, HELEN M. KORCHAK, an1d GERALD WEISSMANN,Departnenltof

Medicine,Division of Rheumatology,Newv York University School ofMedicine,

Newl York 10016

A BS T R A C T Levels ofcyelicAMP(cAMP)(butnot

cyclic GNMP) in

suspensionis

of human polymorpho-nuclear leukocytes (PMN) increased promptly after

exposuire ofthe cellstostimuli suchasthechemotactic

peptidle N-formvyl methioniyl leucyl phenylalanine,the

immnune

complex bovine

serumll

albumtini/anti-bovine serumii albuumin and calciumn ionophoreA23187. cANIP

inereased rapidly, reaching a maximnum of twice the basal level 10-45s after stimulation; after 2-5 min

theamount ofcAMPhad subsided to basal levels.

Eleva-tionsincANMPlevels wvere concurrentwith, or followed,

membrane hyperpolarization (measured by uptake of the lipophilic cation triphenylmethyl phosphonium) an(d alwvays preceded lysosomal enzyme release and

superoxide anion

(0'7)

production. Elevated cAMP

levels

couldl

be uncoupled from these later events by

removal of extracellular divalentcations, replacement ofextracellularNa+ withK+ orcholine+, and byuseof low concentrations of stimulus; each ofthese

condi-tionsvirtually

abolished

lysosomnal

enzyme

release

and(

2 generation, while leaving the stimulated elevation ofcAMPlevelsunimpaired. Calcium ionophoreA23187 did notprovokememnbrane hyperpolarization,thus

un-coupling changesin

memnbrane

potentialfrom changes in cAMPlevels.These data suggested that cAMP is not acritical component in the earliest steps of

stimulus-secretion coupling. Surface

stimnulation

of cells pre-treated with prostaglandins E1 or12 yielded veryhigh levelsofcAMP;these high levelsmaybe animportant

part of the

mechanismii

bywhich stable prostaglandins

inhibit

_lysosomal

enzyme release and °2 generation.

Dr. Smolen and Dr. Korchak are Arthritis Foundation

Fellows.

Receivedfor publication 23July1979andinrevised form

26December1979.

INTRODUCTION

Surface stimulation of human polymorphonuclear

leukocytes(PMN)1 provokesavariety ofresponses, of whichIysosomal enzyme release andsuperoxideanion radical (°2) generation are particularly pertinent to

theirbiological funetion. To elucidatethe biochemical andphysiological pathways responsiblefor the transmis-sioInof stimulatory signals from receptors on the plasma

membrane and the translation of these signals into

cellular responses, wehaveemployedmethods which

allow a detailed examination of several processes

during the first minute ofPMN stimulation. The first

measurable responses of cells exposed to particulate

and soluble stimuli were changes in membrane

potential (1): hyperpolarization was observed within

10 s. A distinct stimulus-dependent lag period of

20-60 s elapsed before theinitiation of°2 generation (1, 2). Wehave alsoreported that theinitial kinetics of lysosomal enzyme release could be timed and showed

that degranulation took place after a lag

period

of

20-60 s (3). These lag periods were also

stimulus-dependent andwere similar in duration tothoseseen

for 2 generation.

Anumberof investigators have reported that human

PMN have unchanged levels of cyclic AMP (cAMP)

afterexposure tophagocyticstimulifor several minutes

(4-8); elevated levels of cyclic GMP have been

re-ported byone group (9-11). In contrast to these find-ings, Herlin et al. (12) showed the exposure of PMN

I_{Abbreviations used in this paper: BSA/anti-BSA, an}

im-mune complex of bovine serum albumin and anti-bovine serum albumin; cAMP, cyclic AMP; cGMP, cyclic GMP; FMLP, N-formyl methionyl leucyl phenylalanine; PiCM, phosphate-buffered saline containing Ca++ and Mg"+; PMN,

polymorphonuclear leukocytes.

1077 J. Clin. Invest. © TheAmerican SocietyforClinicalInvestigation,Inc. * 0021-9738/80/05/1077/09 $1.00

tolatex particles provoked a rapid twofold increase in

cAMP levels. This increment was prompt (maximal

withini 15 s) and brief. Thef£actthat basalcANIP levels wererestored within 1-2 min explained why previous investigators, who had assayed the cell suspensions

after 2-5 min, were unable to detect this response. Herlin et al. suggested that the brief increment in

cAMPP might have a regulatory role in glycogen

metabolism (12, 13).

Since the reported increment in cAMP levels

ap-peared to coincide with membrane hyperpolarization and to precede lysosomal enzyme release and02 genera-tion, we decided to investigate the possibility that cAMPwasaclmediatorof these other events. Insummary,

wefound that prompt, brief elevations in cAMP levels were concurrent with or followed memnbrane

hyper-polarizationi, dependinguponthe stimulus employed. These cAMPincrements always preceded the onset of lysosomal enzyme release and O° generation and could be uncoupled from these latereventsbyseveral

meth-ods; therefore, it seems unlikely that cAMP is an im-port-antmediator of stimltulus-secretion coupling. Finally,

pretreatment of PMN with prostaglandin(PG) E1 or

PGI2 produced modest elevations of cAMP content

which were greatlyaugmiientedbysubse(quentexposure to a surface stimnulus.

XlETHODS

Materials. Cytochalasini B was purchased from Aldrich Chemiical Co. Inc., Milwaukee, Wis. Coneanavalin A and

theophylline were obtained fromii Sigmal Chemiiical Co., St. Louis, NMo. anid N-formyl methlionyl leucyl pheniylalaninie (FMILP) was purchased fromn Peniinsula Laboratories, Inc., SanCarlos, Calif.CalciumilionophoreA23187wvas agiftfrom Eli Lilly & Co., Indianiapolis, Iicl. PGE, anid PGI2 were

generous gifts from Dr. Johln Pike, the Upjohn Company, Kalamazoo, Mich.Allothermiaterials were

reagenit

grade.

Preparation of cell susnpetn.sions. Heparinized (10U/ml) venous blood wasobtainedfrom healthy adultdonors. Purified preparationisofPMINwereisolattedfrom thisbloodbymeanis ofHypaque/Ficoll gradienits (14) followed by standard

tech-ni(qtuesofdextrani sedimiientationi aind hypotonic lysisof erythro-cytes(15). This allowed studies of cell suspensionscontaining 98±+2% PMNI withfewcontaminating erythrocytes or platelets. The cells weresuspendedinabufferedsaltsolutionconsistinig

of 138 mMl NaCl, 2.7 mNI KCl, 8.1 mNI Na2HPO4, 1.5 mMN

KH2PO4, 1 mM\l MgCl2,and 0.6 mnNICaCl2, pH 7.4(PiCNI).

Htumani mononuclearcells were alsoisolatedtusingHypaque/ Ficoll gradlienits(14). Ptirifiedlplateletstuspenisionis were

pre-pared by a mninor modification of the metlhod of Hamberg

etal. (16).

Determinationt ofcAMPlevels. cANMP levels were

deter-mined bya modification of the method of Herlinetal. (12).

Ingenieral, purifiedPMN were stuspended in PiCNM (2 x 107

cells/mil) anid preinctlbated at 37°C for 5 min. An aliqtuot of this suispensioni (200 ,lA) wvas taken as a zero-timiie control samllpleandwas placedin atesttube forextractionofcyclic ntucleotides. This and all other extraction tubes contained 1 ml ofphosphate-bufferedsaline with0.5mnM theophylline

preheated and maintained at 100°C in a vigorouisly boiling

water bath. A stimutltus was added to the remilaining cells

whereupon aliquots were sequentially removed and placed inpreheatedextractiontubesat theindicatedtimes.Thetubes

were keptinthe boiling wvaterbath for 5 min. Uponremoval,

extractvolumes wvere increasedto 6mil wvith0.05 Mlsodium acetate buffer, pH 6.2. The samples werecooled in ani

ice-blath,soniicatedfor5 s(-20 W; sonicator by Kontes Co.,

Vine-land, N.J.)andcentrifuige(d at3,000 g for 15min. The suiper-niates, decanted from thepellets of denatured protein, were

assayed for cAMP content by means of an acetylated radio-immunoassay (New EnglandNuclear, Bostoni, Mass.).cGMP

wvas similarly assayed using more coneentrated extracts

derivedfromalarger numnberof PMN.

This procedure allowed samples to be taken reliably as soonas 5safterstimllulation and attime intervals as short as 5sthereafter. The extraction procedure ensured that the cells

wvere promptly dlenlature(d and that all biochemical reactions wereimmediatelytermiinlated.The extracts contained at least 95% of the cellular vcylic nucleotides, as determined by recoverv oftritiate(dstand(lards.Artifactual increasesinapparenit

cAMP levels, w.hich canl be obtained by boiling adenylate

cvclase reactioni muixtuires (17), were not found wvith this

system,asdeterminedlvbyboilingcellextractsfor timeperiods raniginig from 15sto 10mmin.

Kinietics oflIIsosooinaletnlzymerelease. Detailsconicerniing

the measuremiienitof the initial kinetics oflysosomalenzyme

releatse wsill be publ)lished elsewhere. In brief, we uised a modified flowdlialvsissvsteml,whichpermiiitted semiconitiniu-ouismnoinitorinigof enzymesecretioni.Theapparattusconsisted of a flowdialysiscell,theuipperandlower chambers separated

by a Millipore filter with a

1-Aum

pore size (NMillipore

Corp.,Bedford, Mass.).The lowerchamber,fedbyareservoir,

was continuously eluited,\vith the eluient directed toa

frac-tionlcollector. Theuipperclhamberofthedlialvsiscell contained

a soispensioinofcvtochalasin B-treatted PMN; the h\ydrostatic

pressuire of tlhe

eIltinig

system wasadjuisted stuchi thlait smlllll

amouints ofextracelltlatr miiediumil conitinuiouislytraversed the

filter into the lower chamnber.Thuis,theextracelluilar mediumiii

wascontinuously sampledas eltuenitwvasdirectedtotlle fratc-tion collector. The 6-s stepping frequency of the collector

providedla niomliinal timiie resoltutionioftlhesamiie miiagniituide. The P.MN aind the enitire apparatuis were preincbll)atedl at

37°C,aifter wvhich at boluisconitaininiigaconeenitratedi stimuitiltus

and I'4Clinulinwasatdded. Theinutlin servedasthe extracel-ltulatr space mlarker; it immediately began to cross the filter and thut.s served to miiark the m0lomenit ofstimuiilationi and to

calibrate the enitire apparattis. The appearance of['4C]inulin

in collecte(d fractionis precededthe appearance of lysosomial enzymies by 15s or mnore. Linearextrapolation oftlhe risilng portionsof thecurvestol)aselinepermittedan easy,reliable

miieains of determilining the lag period of' lysosomal enzvyme release relativetostimlall.tion(['4C]inlineltution).Thelinear

extral)olation teclhnii(quie effectively averaged togetlher six to

eighlt dataIpoints, providing a _practical resoluitioni of 1-2s

(uisingtheslamelotofcells).Thelag periods w!erenotatrtifhlcts

dutie to selective retenitioni oflvsosomal enzymies withill the apparatus, sinice no la,g period was seeni for PMIN tlhattwere

introdtuced into the upperchamber atfterstimutlationi.

All l'4C]intilin recovery cuirves (uised to markthe miioment

of'stimultusadditionandthustocalibratethe entireapparatus) atreomitted for the siakeofclarity; thezero-timiie point for all

figuires depicting lvsosomiial enzymie release corresl)on(ls to

tlhatt timiieat ws'hich l'49Cliilin wasfir-steluted.

Miscellanieous procedures. The mnembrane potential of PMN was determinede by _uptake of the lipophilic cattion

triphenylmethyl

p)hosphonium

hromide (1) and continuiouis imleasuiremilentofsuiperoxi(le ainioIn pro(lutctioin wasperformed

as _previously (lescril)ed (1).

RESULTS

Changes in cyclic nucleotides after PMN stimula-tion. Although some investigators have reportedthat

cAMP levels of PMN do not change several minutes

after surface stimulation (4-8), more recent work has indicated that significant changes do take place at short time intervals (12). Consequently, we determined whether such transient changes correlated with other early cellular responses to surface stimuli. These

re-sponses included membrane hyperpolarization,

lyso-somal enzyme release, and superoxide anion (02)

generation.

Purified resting PMN had cAMP levels of 6.7±+1.8

pmol/107 cells(n = 44)when measured byanacetylated radioimmunoassay.cGMPlevelswerefoundto be 0.32 ±0.12 pmol/107 cells (n = 16). Boiling of cell suspen-sions was used to ensure prompt termination of

bio-chemical processes; this boilingdid notproduce any

artifactualincreasesincAMPlevels(17).The extraction

procedure was efficient, since recoveries of trace quantities of tritiated cyclic nucleotides averaged 100

±+11%(n = 11)forcAMPand98±8%(n = 7)forcGMP.

When suspensions of PMN were exposed to the

chemotacticpeptide FMLP (0.1

,tM),

avery rapid in-crease incAMP levels was observed (Fig. 1).

Signifi-cantincrementsincAMPwereobservedat timeintervals as shortas5 s.The increase incAMP peakedatup to twicethe basal level in 10-15s;theaverage maximum

a)

(-0 -6 E

CL. a.

02

0

0 30 60

Time (s)

90 120

FIGURE 1 Effect of FMLP on cyclic nucleotides in sus-pensionsofPMN.Aliquots ofa suspension ofpurifiedPMN

were taken atthe indicated times after exposure to FMLP (100 nM). These samples were immediately extracted by heat-ing at 100°C and then assayed for cAMP contents as detailed in Methods. The zero-time sample was extracted before

addition of the stimulus; an appropriate amount of FMLP was added to it during the extraction process. cGMP was determined in an identical fashion. The results shown are from a single typical experiment. The average peak increment incAMPcontent was 78 +38%(SD) above basal levels (n = 17).

2.0h

0t

E __1.5

+a_CL

0-E

' 1.0

-E

c

_~~~~~~~~~~~~~~~~~~~~~~~~-o~~~~~~~~~~~~~~~~t

Cl)

ax

CD n

3> -M

0

10

a

10

-5

_-0 30 60 90 120

Time (s!

FIGURE 2 Effect ofFMLPoncAMP, transmembrane

poten-tial, lysosomal enzyme release, and 02 generation in PMN. cAMP was assayedasdescribedin thelegendto Fig. 1 and

Methods.Membranepotential

(A*)

wasmeasured bythe

intra-cellularcontentof[3H]TPMP (1). Superoxideanion(0°)was

measuredby thecontinuousassayofferricytochromec

reduc-tion (1, 2). The semicontinuous assay of 8-glucuronidase

secretion (,8-gluc) was performed using the modified flow

dialysis techniqueoutlined inMethods(3). All four responses were measured at the indicated times after exposure of the PMN to FMLP (100 nM).The results shown arefrom single

typicalexperiments.

increase, measured at 10 or 15 s, was 78±38% (SD)

(n = 17)above the basal levels. cAMPcontentdeclined rapidly thereafter and approachedbase line at 1-2 min,

a process which was complete by5 min (not shown).

ThecAMPlevels of unstimulated cells didnotchange during the course ofthe experiment. Cytochalasin B

(5

,ug/ml)

didnotaffect thetime-courseor thepeak level observed (102+23% (n = 4)above the basal level). No

changes in cGMP levels were observed during this

time period with FMLP or with any other stimulus. Relationship between cyclic nucleotide changes and otherPMNresponses. Thedata from Fig.1 wereplotted along

with

other

early

responsesofPMNto

place

them all withina commontemporal framework (Fig. 2).

Stimu-lation with FMLP provoked arapid increase in

mem-branepotential

(Aifi),

asmeasuredby enhanced cellular uptake of the lipophilic tritiated triphenylmethyl

phos-phonium cation. Initiationof hyperpolarization,which wasvirtually immediatefor allstimuli,wasconcurrent withtherapidincreasein cAMPlevels.However,both of thesepromptresponsesclearly precededtheonsetof 02 generation, which had alag period of-20 s when monitoredcontinuously(1-3). Release of

/-glucuroni-dase, an enzyme found predominantly in azurophil

granules, was assayed by means ofa semicontinuous

flow dialysis system described previously (3). Like

°2

generation, release of

/3-glucuronidase

also had a lag period of -20 s. Thus, the increases in cAMP levels

clearly preceded lysosomal enzyme release and

°2

generation, but were concurrent withmembrane hyper-polarization.

When an immune complex consisting of bovine

serum albumin andanti-bovine serum albumin (BSA/

anti-BSA) was used as the stimulus, membrane hyper-polarization again appearedimmediately(Fig. 3). Both lysosomal enzyme release and °2 generation were initiated after distinct lag periods of -40 s. Unlike the casewith FMLP,theincrementincAMPlevelinduced

by an optimum secretory concentration of

BSA/anti-BSA (150 ug/rml) had a lag period. The length of this

lagperiod was variable butit wasalways distinct.-For the caseshowninFig. 3, thelag period for theincrease

in cAMP relative to membrane hyperpolarization

(LcAMP) was 35 s. Thus, since hyperpolarization

pre-ceded elevation ofcAMP content, changes in cAMP

levels areunlikelytobe the cause of changesin

mem-branepotential for this stimulus.

ThecalciumionophoreA23187

(10,uM,

theoptimum

secretory concentration) did not provokerapid changes

in membrane potential (Fig. 4), perhaps because

in-ducedcalcium fluxes bypassed the need for the usual cell surface receptor-ligandinteractions which indirectly

lead to hyperpolarization. Nonetheless, lysosomal en-zyme release and °2 generation were both elicited afterrelatively long lag periods. Increments in cAMP

2.0h

0

E

q- 1.5

C

._

0-z

E

-E

.,

1.0 H

cz

c

0 0

a>

cx

0 30 60 90 120

Time (s)

FIGURE 3 Effectof BSA/anti-BSAoncAMP,transmembrane potential, lysosomal enzyme release, and 0° generation in

human PMN.cAMP,membrane potential,°2generation, and

B-glucuronidase

(fl-gluc)

releaseweremeasuredasdescribed

inthe legends toFigs. 1 and2and inMethods. All four

re-sponses weremeasured atthe indicatedtimesfollowing

ex-posureofPMNtoBSA/anti-BSAimmunecomplex(150ZgIml)

(18). The results shownarefromasingle typicalexperiment.

Theaverage peakincrement in cAMPinducedby

BSA/anti-BSA was 47±4% (SD) above basal levels (n =_{3). LcAMP,}

lagperiod forincreaseincAMPrelativetomembrane hyper-polarization.

.15_{°- 1.5}

a-CL

F- 1.0

E )W

c.

'.

0

'7

0 30 60

Time (s)

,

R

c

c

CL-ID CD

90 120

FIGURE4 Effect ofA23187 on cAMP,transmembrane

poten-tial, lysosomal enzyme release, and 2generation in human

PMN. cAMP, membrane potential, 02 generation, and B-glucuronidase(,8-gluc)secretion were measuredasdescribed

inthe legendstoFigs. 1and 2 andMethods. All four responses

weremeasuredat theindicated times following exposureof

PMN to calciumil ionophore A23187 (10

_A.tM).

The results shown arefrom a single typicalexperiment.The averagepeak increment in cAMP induced by A23187 was 67+18% (SD) above basal levels (n =3).

were also observed, again with relatively long lag

periods.

Effects of mononuclear cells and platelets. Eleva-tions of cAMP levels were notdue tothe presence of contaminatingmononuclear cells. Isolated mononuclear

cells, which containsubstantial amounts ofcAMPand accumulatemorecyclic nucleotideinresponse to some stimuli (4),wereexposedtoFMLP,BSA/anti-BSA, and A23187.A quantityof these mononuclear cells equivalent to twice the normal level of contamination was stimu-lated and extracted along with the usual number of

unstimulated PMN. No systematic increases in cAMP

were observed (Fig. 5) when mononuclear cells were

exposed tothese stimuli. PMN, which were obtained

from the same blood sample from which the

mono-nuclear cells were isolated, responded fullyto FMLP (upper line). Similar experiments indicated that plate-lets at concentrations up to twofold the normal level ofcontamination could not account for the observed increases in cAMP (datanot shown).

All other stimuli tested were capable ofprovoking

increments of cAMP (but not cGMP) together with

lysosomal enzyme release and 02 generation.

Zymo-san-activated serum (10%) induced alarge immediate increase incAMPlevels,which was virtually identical tothatseenfor FMLP (datanotshown).The lectin con-canavalinAproducedaweakresponsewithalag period

comparable to that of A23187. Opsonized zymosan

particles produced the smallest increments in cAMP

content (only 20+14% above basal levels, n =3); lag periods were also observed with this stimulus.

Are elevations of cAMP required for neutrophil

activation? These results allowedus toconclude that

PMN Response A Fto MLP

MononuclearCell Responsesto

A23167

s-F

CM=fLP

- BSAlanti-BS4

30 60 90 120

Time (s)

FIGURE 5 Effect of stimulioncAMPlevels ofcontaminating

mononuclear cells. A purified PMN suspension containing

3%mononuclear cellsasnormalcontaminantswasexposedto FMLP and cAMP was generated as indicated by "PMN responsetoFMLP."Anumber ofpurified mononuclear cells correspondingto6% contamination wassimilarly exposedto

FMLP (100nM), BSA/anti-BSA(150,ug/ml),and A23187 (10

,uM).These mononuclear cellswereextractedalongwith the

usual number of unstimulatedPMN(toassuresimilar

extrac-tionconditions)andassayedfor cAMP. Resultsareexpressed aspercentages of basal levels for each experiment; this was

done for the sake ofclarity since basal levelsvariedslightly duetothe addition of mononuclear cells. The results shown

arefromasingle typicalexperiment.

the observedincreasesincAMPlevelswereconcurrent

with, or followed, membrane hyperpolarization and

were thus unlikely to be the cause of this earlier

event. While the possibility remained that the early

membrane hyperpolarization could be the cause of

cAMP changes, the results of experiments in which

A23187wasthe stimulus (Fig. 4)indicated that changes

in membrane potential were notnecessary forany of

the laterresponses.

Thefollowingexperimentsshowedthatthe

stimulus-dependentincrements incAMP levels could be

simi-larly uncoupled from the two subsequent responses.

The presence of EGTA effectively blocks lysosomal enzyme release and °2 production. Yet the removal

of divalentcations didnotaffectthehyperpolarization

response (1), nor did itaffect the stimulated increase

incAMPlevels(Fig. 6);thetimecourseand peak

ampli-tudes ofthe responses were similar in the presence orabsence of EGTA.Replacement ofNa+inthe medium

witheitherK+orcholine+ inhibited lysosomalenzyme

release, 0 production, and membrane

hyperpolariza-tion,2 butdid notsubstantially affect the increment in

cAMP (Fig. 7). Finally, by useofappropriate

concen-trations of FMLP, it was possible to obtain a large

increase in cAMP levels without attendant lysosomal

2_{Korchak, H. M., and G. Weissmann. Stimulus-response}

couplingin humanneutrophil. Membrane potential changes andthe role of extracellular Na+. Submitted for publication.

FIGURE 6 Effect of EGTAon cAMPresponse to FMLP. A

sample ofPMN wasprepared and dividedintotwolots. The

first lot was washed in phosphate-buffered saline (without

divalentcations); EGTA was then addedto a concentration

of1 mM.The other lotwaswashed and otherwise prepared

normally in PiCM (control). The EGTA-treated and control

cellswereexposedtoFMLP(100 nM)and cAMPwas

deter-minedasdescribedinthe legendtoFig. 1and Methods. The

results shownarefromasingle typicalexperiment.Theaverage

peakincrement incAMPinduced byFMLPinthepresence

of EGTAwas59+28%(SD)above the basallevel(n =3).

enzymereleaseor 2generation(Fig. 8). Approximately

100 nM FMLP was required to provoke optimal

en-zyme release or °2 production; both of these later

responses werealmost undetectable when1 nMFMLP

was employed. However, a large incrementofcAMP

(observed10safterstimulation)wasinduced by1-100

nMFMLP;at1nM,theincrementofcAMPwasnormal,

while virtually no lysosomal enzyme release or

°2

generationwas found. A substantialincrease incAMP

level was obtained using as little as 0.1 nM FMLP.

20

(-)

"-cK

0

E

,.

a_

30 60 90 120 Time (s)

FIGuRE 7 Effect ofNa+-free media on cAMP response to FMLP. PMNdesignated K+wereprepared and washedin155 mM KC1, 1.3mMCaC12, 1.2mM MgCl2,and 10mM Hepes-HCI,pH 7.4.Cells designated "choline" werewashedinan

identicalmediumexceptthat choline chloride replacedKC1.

Bothlots of cellswereexposedtoFMLP(100 nM))andcAMP

wasdeterminedasoutlinedinthe legendtoFig. 1and Meth-ods.The results shownare fromasingletypicalexperiment.

Theaveragepeakincrement incAMPinduced byFMLPwas

52+23% (SD)inthepresenceof choline(n =3)and48+16%

inthepresence ofK+(n =3).

ElevatedCyclicAMPLevels in Polymorphonuclear Leukocytes

-200-z

-J

m

100-I0V)

10K

C-)

0a

E

02~

-Control

30 60

Time (s)

90 120

0 150

-J

m

g 125

0

0.01 0.1 1 10 100 1000

FMLP Concentration (nM)

v

)C

~D0

E

a 0

CD

I._C\J

0

t

FIGURE 8 Effect of FMLP concentration on inerements in cANIP, lysosomal enzyme release, and superoxide anion generation. PMN were exposed to the indicated concentra-tions ofFMLP. Theamouint of cAMP generated at 10s was

measuredand isexpressed as a percentage of thebasal (zero-time) level.,8-Glucuronidase release and O° generationwere

measuired5 min afterstimtulation tusingstandardtechniquies

(19). Theresuilts shown are fromasingle typical experiment. The average peak increments in cAMP(ahove basal levels) inducedhy FMLP were 78+38% (SD) (n 17) at 100 nM,

72±40% (n =3) at 10nM, and 63±37% (i = 4) at 1 nMI.

Un-certainties in °2 generation and lysosomal enzyme release

aretypically ±5%.

Thu.s, a cAMP increment at 10s is not a suifficient

prerequiisite forsubseqtuent lysosomnal enzyme release and O° generation.

Exposuire of PMN to PGE1 or PGI2 occasionally

producedbothmodest increases incelltularcAMPlevels andinhibition oflysosomal enzyme release. This

day-to-day and subject-to-subject variability was rectified

by rolitine preincubatioin of the cells with theophyl-linefor 5 min at370C.Theophylline itself (0.1-2 mnl)

had little or no effect upon the customary increment

in cAMP in response to FMLP; average peak levels

were 109±71%(SD)above basal levels (ii =3). Fig.9

shows that wheni PMN pretreated with 0.5 mM1

theo-phvlline wereexposedtoPGEI,cAMPlevelsincreasedl

two- tothreefoldoverthe spanof4min.If

PGE,-treated

cellswerestubsequentlystimulated with FMLP,alarge

increment in cAMPwas observed. This response was

promptandpeakedat alevel of 8-10-fold the normal

basal quantity. The

PGE,-treated

PMN also released

65%less

,-glucuronidase

andlysozyme in responseto FMLP (in the presence oftheophylline)thandid those cells which were not exposed tothe PG. As shown in

Fig.10, PGI2 inducedamodestincrease in cAMPlevels in theophylline-treated cells (higherconcentrations of

theophyllinewerenecessarytoobtain inhibitionof

en-zyme secretion by this prostaglandin); once again,

stimulation with FMLP produced a large additional

incrementincAMPand inhibitionoflysosomalenzyme release was observed. Decreasing theamountof PGE1 from 250to 100uMreducedtheamount ofinhibition of

Inhib,foonof

G/lucuronldose

= 50 Re/eose

Ca) U *_~~~~~~A 65%

40

E 30_ 20 10

K ~~~FMLP

1 2 3 4

Time (min)

FIGURE 9 Effect of PGE, and FMLP on cAMP levels and lysosomal enzymie release in PXiN. Puirified PMIN were pre-incuhatedinPiC\I containing 0.5 mrn theophylline for 5 min

at37C. AttimezeroPGE,(250 ,LM)was addedand samiiples were taken every minuite for fouir min (lowerecurve). After twomin,asampleofthese cellswasremnoved and stimluilated with FMILP (100 nM)asshown hy the upper curve. Lysosomcal enzyme release induced vy FNMLP (in the presence of

the-ophylline), determinied as described (19), was inhibited by pretreatmenit with PGE, (as ind(licated). The resuilts shown aIre fromiasingletypicalexperimient. The aiverage peak inerements induced by FMILP above the basal level estab-lished by PGE, +theophylline is 187±45% (SD) (n =7).

lvsosomal enzyme release without any similar redluction of either base-lineorpeak FMLP-induced cAMPlevels (data not shown); this indicates that inhibition of de-granulation is not strictly determined by the level of cAMPinduced by the PG or bv the peak level provokedl bystubsequent stimlllliationi.

DISCUSSION

Increases in the cAMP levels of humnan PMN provoked l)v surface stimtulation have not been regularly ob-served previously. Table I shows a list of selected

references, the first five of which reported no changes in cAMP. This is easily attributable to the fact that these investigators measured cyclic nucleotide levels

2 min or more after stimulation, by which time the

increments in cAMP had already subsided (12; this

work). The increment observed by Manganiello et al. (4)wasattributed tocontaminating mononuclear cells; the small number of these cells present in purified PMN preparations were not responsible for the incre-mentsreportedinthisstudy (Fig. 5).We did notobserve theincreases in cGMPwhich Ignarro andhis co-workers have reported (9-11, 20).

We found that prompt increments in cAMP levels

were provoked by each stimulus employed. These

increments are unlikely to be associated with the micro-tubuleassemblywhichgenerallyaccompanies

stimula-tion,sincemicrotubuledisassemblyhasbeen suggested asmechanismforelevatingcAMPlevels (21);however,

In

C-)

0

0

E

,a

a_

CL

ct

inhibitionof

#o_4-GAucuronidose_Releose

\ 66%

1 2 3 4

Time (min)

FIGURE 10 Effect of PGI2 and FMLP on cAMP levels and

lysosomalenzymereleaseinPMN.PurifiedPMNwere

preincu-bated with2mMtheophylline,thenexposedto 100

AtM

PGI2 (dissolvedinPiCM 30 sbeforeaddition) anidFMLP(100nM).

See legend to Fig. 9 forfurther details. The results shown

arefromasingle typical experiment. Theaveragepeak

incre-ments induced by FMLP above the basal level established byPGI2 +theophyllineis

195±91

(n =3).

it ispossible that transientdisassembly isresponsible forthe briefincrement in cAMP.

TheincrementsincAMPwere concurrentwithor fol-lowed membrane hyperpolarization and preceded lysosomal enzyme release and

02

generation. This

temporal sequence suggested that cAMP

might

be a

secondary messenger responsible for transmission of the stimulatory signal to the mechanisms

responsible

for the laterevents.However, severalexperiments

sug-gested thatthis maynotbe thecase:

lysosomal

enzyme

release andO° generation could be

virtually

abolished whereas the prompt increments in cAMP were

un-changed

(a)inthe absence of the divalentcations(Fig. 6),(b)in theabsence of extracellularNa+ (Fig. 7),and

(c) when low concentrations of FMLP (1 nM) were

used (Fig. 8).These data showedthataninerement in

cAMP is not a sufficient condition for secretion and O° generation. That elevation of cAMP contentis not necessaryfor theseevents issuggested (butnotproven) by the fact that serum-treatedzymosanproduceslittle or no change in cyclic nucleotide levels. The finding thatthe absence of Ca++ andNa+havelittleor noeffect upon the stimulated increment in cAMP is also note-worthy, in view ofthe important roles these cations play in stimulation of PMN. Degranulation of rabbit PMN is impaired by theabsence of both Na+ (22) and Ca++ (23), and enhanced influxes of both cations take placeafter stimulation (24).

Increments in cAMP and the hyperpolarization

re-sponse wereclearly uncoupled when A23187 was used

asthe stimulus(Fig. 4).Thissuggeststhat some process

other than membrane hyperpolarization, such as an increasein intracellular Ca++ (achieved either by

stimu-lated Ca++ influx or by mobilization of intracellular Ca++stores), isresponsible for elevated cAMP contents.

It iswellknown that exogenous cAMP or agents that

elevate intracellular levels ofcAMP inhibitlysosomal enzyme release (25). Indeed,itseems paradoxicalthat

stimulated cells have elevated cAMPlevels at justthe momentwhenlysosomal enzyme release and O° genera-tionisinitiated.Wewouldsuggestthata meredoubling

of cAMP levels is innocuous insofar as these later

events areconcerned.A"threshold"somewhereabove

this doubled level ofcAMP would have to bereached before secretion is inhibited. The increment of cAMP

could be eitheran accidental or vestigial response to surface stimulationorcould havesomeothermetabolic

significance (13). Consistent with this explanation, it

should be noted that the elevation in cAMP seen in PG-treated cells (Figs. 9 and 10), though often

con-sidered the mechanism by which these agents inhibit secretion, is notmuch greaterthan that obtained with

stimuli such as FMLP alone. It is far more likely that the large increments of cAMP obtained when PG-treated cells are subsequently exposedtostimulation (Figs. 9 TABLE I

Effectof Surface Stimulation otlcAMIPLevels ofHumaniPMN:ReportedValues

After Periocdof

Referenice Stimllulus Basal stimiiulationi stimilulation

p)IIIoiii0'cells

Ignarro andGeorge(10) Zymosan 0.7-1 0.7-1 2min

Manganielloetal.(4) Latex 3-5* 12-27* 5 min

Stolc(5) Latex 10 13 15min Ignarro etal. (8) Zymosan -1.0 -1.3 15 min

Zurieretal. (22) Zymosan 5 7 20 min

Herlinetal. (12) Latex 7 14 15 s

This work FMLP,

BSA/anti-BSA,

A23187 7 14 15 s *Values reportedas

pmol/mg

cellprotein.

and 10), a phenomenon also reported by Zurier et al. (25), are responsible (either directly or indirectly) for inhibition of granule discharge and °2 generation.

Thus, cellspretreated with PG are "primed," such that stimuli evoke large amounts of cAMP and ironically contribute to inhibition of secretion.

One possible explanation for this phenomenon is that the decrease in membrane microviscosity which accompanies surface stimulation (26) permits a closer coupling of PG receptors and adenyl cyclase, in a man-ner similar to that suggested by Hirata et al. (27);

stimulation of PG-treated PMN would thus lead to

greatly enhanced production of cAMP. While this is an

explanationconsistent with our data, it should be re-membered that we do not know what effect a doubling

ofcAMP levels per se has on secretion. We do not know if the intracellular/extracellular distribution of cAMP is identical with PGE1 and FMLP stimulated PMN. Furthermore, the fact that lower levels of prostaglandins

produce correspondingly less inhibition of lysosomal enzyme release, while giving rise to identical incre-ments in cAMP levels,indicatesthat these agents have some inhibitoryeffects other than the mere elevation ofcAMP. Investigations into the mechanisms and

sig-nificance ofthis phenomenon are continuing.

ACKNOWLE DM E NTS

This work was aided by grants (AM-11949, GM-23211, HL-19072, and HL-19721) from the National Institutes of

Health, TheNational Foundation-March of Dimes, The

Na-tional Science Foundation (76-05621), and the Arthritis Foundation.

REFERENCES

1. Korchak, H. M., and G. Weissmann, 1978. Changes in

membrane potentialof humangranulocytes antecedethe

metabolic responses to surface stimulation. Proc. Natl.

Acad.Sci. U. S. A. 75: 3818-3822.

2. Cohen, H. J., and M. E. Chovaniec. 1978. Superoxide

generation by digitonin-stimulated guinea pig

granulo-cytes. A basis for a continuous assay for monitoring

superoxide productionandforthestudyoftheactivation

of thegeneratingsystems.J.Clin. Invest.61:1081-1087.

3. Smolen,J. E., H. M.Korchak,and G. Weissmann. 1978.

The onsetoflysosomal enzyme releaseand superoxide productionincytochalasin B-treated human granulocytes.

J. Cell. Biol. 79: 208a.

4. Manganiello, V., W. H.Evans, T. P. Stossel, R. J. Mason, and M.Vaughan. 1971. The effect ofpolystyrenebeads

on cyclic 3',5'-adenosine monophosphate concentration

inleukocytes. J. Clin. Invest. 50: 2741-2744.

5. Stolc,V.1971.Regulationof iodinemetabolisminhuman

leukocytes by adenosine-3',5'-monophosphate. Biochem. Biophys.Acta. 264: 285-288.

6. Schell-Frederick, E., and J. Van Sande. 1974. Evidence thatcyclicAMP is notthe chemical mediator of the

meta-bolic changes accompanying phagocytosis.J.

Reticulo-endothel.Soc. 15: 139-148.

7. Stolc,V.1974.Restrainedadenyl cyclaseinhuman

neutro-phils: stimulation of cyclic adenosine-3',5'-monophos-phate formation and adenylcyclase activity by phagoeyto-sis and prostaglandins. Blood. 43: 743-748.

8. Ignarro, L. J., R. J. Paddock, and W. J. George. 1974. Hormonal control of neutrophil lysosomal enzyme release: effect of epinephrine on adenosine-3',5'-monophosphate. Science (Wash. D. C.). 183: 855-857.

9. Ignarro, L. J., T. F. Lint, and W. J. George. 1974. Hormonal control of lysosomal enzyme release from humain neutro-phils. Effects of autonomic agents on enizymiie release

phagocytosis, and eyclic nucleotide levels.J. Ex1). Med. 139: 1395-1414.

10. Ignarro, L. J., and W. J. George. 1974. Hormonal control of lysosomal enzyme release from human neutrophils:

elevation of cyclic nucleotide levels by autonomic neuro-hormones. Proc. Natl. Acad. Sci. U. S. A. 71: 2027-2031. 11. Ignarro, L. J., and W. J. George. 1974. Mediation of

im-munologic discharge of lysosomal enzymes fromhnman

neutrophils by guanosine-3',5'-monophosphate. Require-ment of calcium, and inhibition by adenosine-3',5'-monophosphate.J. Exp. Med. 140: 225-238.

12. Herlin, T., C. S. Petersen, and V. Esmalnn. 1978. The role

of calcium and cyclic adenosine 3',5'-monophosphate in the regulation of glycogen metabolism in phagocytizing human polymorphonuclear leukocytes.Biochimn.Biophys. Acta. 542: 63-76.

13. Petersen, C. S., T. Herlin, and V. Esmnann. 1978. Effects of catecholamines and glucagon on glycogen metabolisii inhuman polymorphonuiclear leukocytes. Bicochirn. Bio-phys. Acta. 542: 77-87.

14. Boyum,A. 1968.Isolationofmononutclearcells and granu-locytes fromhumaanblood.Isolation of mononuclear cells by combining centrifugation and centrifugation at Ig.

Scand.J. Clin. Lab. Invest. 21: 77.

15. Zurier, R. B., S. Hoffstein, and G. Weissmann. 1973.

mechanism of lysosomal enzyme release from human leukocytes. I. Effect of cyclic nucleotides and colchicine. J. Cell Biol. 58: 27-41.

16. Hambert,NM.,J.Svensson, T. WVakabayashi, and B. Samuels-son. 1974. Isolation and structure of two prostaglandin

endoperoxidesthat causeplateletaggregation. Proc. Natl. Acad. Sci. U. S. A. 71: 345-349.

17. Larner, E. H., and C. Rutherford. 1978. An artifact pro-ducedby boiling adenyl cyclase reactioin mixturesprior tothe cyclic AMP-radioimmunoassay. Anal.Biochern.91: 684-690.

18. Ward, P. A., and N. J. Zvaifler. 1973. Quantitative phago-cytosis ly neutrophils. I. A new method with immuIILne

complexes. J. Immitnol. III: 1771-1776.

19. Goldstein,I.MI.,S.Lind, S.Hoffstein,and G.Weissmnanmm.

1977. Influence oflocal anesthetics uipon humnan

poly-morphonuclear leukocyte funictionitl vitro.Reduiction of

lysosomal enzymerelease and stiperoxide anioni prodluc-tion.J. Exl. Med. 146:483-494.

20. Smith, R. J., andl L. J. Ignarro. 1975. Bioregulatioln of lysosomal enzyme secretion fromn hlumllan nemitrophils:

roles of guanosine 3',5'-monophosphate and calcium in

stimulus-secretion coupling.Proc. Natl.Acad. Sci. U. S. A. 72: 108-112.

21. Malawista, S. E., J. M. Oliver, and S. A. Rtudolph. 1978. Microtubules andcyclicAMP inhumnan leukoevtes: On

the order of things. J. Cell Biol. 77: 881-886.

22. Showell, H. J., P. H. Naccache, R. I. Sha'afi, and E. L. Becker. 1977. The effects ofextracelltular K+, Na+, and Ca++ on lysosomal enzyme secretion from polymiorpho-nuclear leukocytes.J. Immunol. 119: 804-811.

23. Becker, E. L., and H. J. Showell. 1974. The ability of chemotactic factors to induce lysosomalenzymerelease.

II. The mechanism of release.J. Immunol. 112: 2055-2062.

24. Naccache, P. H., H. J. Showell, E. L. Becker, and R. I. Sha'afi. 1977. Changesin ionic movements acrossrabbit polymorphonuclear leukocyte membranes during lysosomal enzymerelease. Possible ionic basis forlysosomal enzyme release. J. Cell. Biol. 75:635-649.

25. Zurier, R. B., G. Weissmann, S.Hoffstein,S. Kammerman,

and H.-H. Tai. 1974. Mechanism oflysosomal enzyme releasefromhuman leukocytes.II. Effectsof cAMP and

cGMP, autonomic agonists, and agents which affect

microtubule function.J. Clin.Invest. 53: 297-309.

26. Berlin, R. D.,and J.P.Fera.1977.Changesinmembrane

microviscosity associated with phagocytosis: effects of colchicine.Proc. Natl. Acad. Sci. U. S. A. 74: 1072-1076.

27. Hirata, F., W. J. Strittmatter, and J. Axelrod. 1979. ,8-Adrenergic receptor agonists increase phospholipid

methylation, membrane fluidity, and S3-adrenergic

re-ceptor-adenylate cyclase coupling.Proc.Natl.Acad. Sci. U. S. A. 76: 368-372.