Thorax 1991;46:124-130

Cysticfibrosis - 4 Series editor:DJShale

Abnormalities of airway epithelial functionand the implications of the discovery of the cysticfibrosis gene

AlanW Cuthbert

Of all the functional abnormalities present incystic fibrosis, those affecting the lung are themost serious. In 1989 research workers inCanada and the United States reported thestructure of the gene responsible for 70% ofcystic fibrosis among white people.`3 I shallexplore what is known about abnormal lungfunction in cystic fibrosis at the cellular andmolecular level and discuss how the recentlydiscovered genetic information may be relatedto the cellular defects. Although it is too earlyto speak of a treatment for cystic fibrosis,the increased understanding of underlyingmechanisms suggests ways in which treatmentmay be approached in the future.By means of a painstaking combination of

chromosome walking and jumping andcomplementary DNA (cDNA) hybridisationover half a million base pairs on the long armof chromosome 7 band q 31 have beensequenced.`3 The putative protein coded forby the cDNA for the cystic fibrosis locusconsists of 1480 amino acids. The tri-nucleotide codon for the 508th amino acidfrom the N terminus (phenylalanine) wasfound to be deleted in 70% of patients withcystic fibrosis. On the basis of analogy andhomology a tentative structure for the proteincoded for by the gene has been deduced (figure1). The protein has been named the cysticfibrosis transmembrane regulator.This protein has two related motifs, each

consisting of six membrane spanning regionsjoined by a long polypeptide chain. It isproposed that very little (5%) of the protein isexposed externally and that some 80% is inthe cytosol. Two features of the cytosolicdomain are important. The first is a highlycharged R domain with 16 potential sites forphosphorylation and the second twonucleotide binding folds, which may bind oreven hydrolyse ATP. The phenylalanine dele-tion (AF508) occurs in the first nucleotidebinding fold.Other proteins with similar structures and a

fair degree of homology are already known,principally the P glycoproteins or multipledrug resistance (MDR) proteins. Theseproteins are responsible for the energy depen-dent efflux of cytotoxic drugs from cells.Increased expression of multiple drug resis-tance proteins after exposure to cytotoxic

Figure 1 Hypothetical structure of the cysticfibrosistransmembrane regulator. Two sets of six membranespanning domains arejoined by a long polypeptide chain.Two nucleotide bindingfolds (NBF) bind A TP. Incysticfibrosis the missing phenylalanine is in the nucleotidebindingfold nearest to the N terminus. The R domaincontains 16 potential phosphorylation sites. Note thatvery little of the protein is exposed to the external(upper) surface of the cell.

drugs may be a major cause of drug resistancein cancer chemotherapy.4

Transepithelial ion transport in therespiratory systemWhat goes awry in the respiratory system incystic fibrosis and how is this related to thepresence of the cystic fibrosis transmembraneregulator protein with phenylalanine deletedat position 508? It is obvious, but worthrestating, that biological systems have no wayof pumping water. Water can be made tomove across biological barriers only byosmosis-provided, of course, that thebarriers are water permeable. Movement offluid therefore depends on solute transport,usually of inorganic ions. In the lungs surfaceliquid elaborated in the alveoli and small air-ways is moved by ciliary action to moreproximal regions. As fluid does not accumulatein the proximal regions even though there is a4000 fold reduction in total volume and acomparable reduction in surface area, fluidmust be reabsorbed as a consequence of iontransport.5 These same processes are vitallyimportant at birth for converting fluid filledlungs to respiratory organs.How do cells lining the respiratory system,

UniversityDepartment ofPharmacology,Cambridge CB2 1QJAW CuthbertReprint requests to:Professor Cuthbert

124

on October 2, 2021 by guest. P

rotected by copyright.http://thorax.bm

j.com/

Thorax: first published as 10.1136/thx.46.2.124 on 1 F

ebruary 1991. Dow

nloaded from

Abnormalities ofairway epithelialfunction in cysticfibrosis

Figure 2 Models of electrogenic sodium absorption (a) and electrogenic chloridesecretion (b). Cells are joined at the apical surface by tight junctions. In (a) sodium ionsenter the cell through apical sodium channels and are pumped out of the basolateralsurface by the sodium pump (P). Potassium ions cycle via the sodium pump andbasolateral K channels. In (b) chloride ions enter the cell by a basolaterally locatedcarrier (C) along with potassium and sodium ions. The latter two are dealt with by thepump and potassium channels as before. Chloride ions leave through the apical surfacevia chloride channels sensitive to cyclic AMP. Counterions, chloride in (a) and sodiumin (b), move through paracellular pathways. A-amiloride.

particularly in the more proximal regions,effect fluid absorption? Two types of transport-ing mechanism are depicted in figure 2-namely, electrogenic sodium absorption andelectrogenic chloride secretion. Both proces-ses are powered by the sodium pump, a

sodium-potassium activated, magnesiumdependent ATPase. The transporting epi-thelial cells are asymmetrical-that is, theapical and basolateral membrane domains dif-fer in their properties in such a way that themolecular apparatus of transport, ion chan-nels, ion pumps, and exchange mechanismsare exclusively assigned to one membranedomain or the other. These molecular entitiesare prevented from passing from one domainto the other by the tight junctions betweenindividual cells. How an individual cell is ableto insert these protein molecules into thecorrect membrane domain remains an intri-guing and largely unsolved problem.Sodium absorption is dependent on the

presence of selective sodium channels in theapical membrane (fig 2a), which can bespecifically blocked by drugs such as amilorideand triamterene.6 Although sodium channelsare present there will be net sodium influx intothe cell only if a favourable electrochemicalgradient exists. Sodium ions are expelled fromthe cell across the basolateral membrane viathe sodium pump, which has the stoichio-metry of 3Na+ to 2K+ ions. The transfer ofthe sodium cation across the epithelial barriercreates a favourable electrical gradient for themovement of the chloride counteranion,which moves passively across the tissue,probably by a paracellular route. Potassiumions establish electrochemical equilibriumacross the basolateral face via potassium chan-nels, which in many systems are sensitive tothe concentration of intracellular calcium.Electrogenic sodium absorption can beprevented by inhibitors of the sodium pump,such as ouabain and other cardiac glycosides.There appear to be few endogenous substan-ces that can regulate this sodium absorptiveprocess in respiratory epithelium, though

adrenaline is known to stimulate the processin the perinatal period.7Normally electrogenic chloride secretion is

not present in respiratory epithelium, but itcan be activated by beta agonists and otheragents that increase intracellular cyclic AMP,prostaglandins, and bradykinin. The first stepin chloride secretion requires that chloride isaccumulated in the cell to a level greater thanexpected from the elctrochemical gradient.The cell does this by using a sodium-potas-sium chloride cotransporter with the stoi-chiometry Na+: K+: Cl- 1:1:2 (fig 2b). Theenergy stored in the electrochemical gradientfor sodium between the outside of the cell andthe interior provides the energy source fordriving potassium and chloride uphill into thecell. The sodium and potassium ions thatenter the cell along with chloride are dealtwith by the sodium pump and by equilibra-tion through basolateral K+ channels respec-tively. Chloride is then free to move out of thecell, across the apical membrane, and down itselectrochemical gradient, provided that thechloride channels are open. Agents thatincrease cyclic AMP increase the open stateprobability of these channels. Thus theprocesses of chloride secretion lead to a trans-fer of charge (anions) across the tissue suchthat a favourable gradient for accompanyingcountercation movement is generated. Themain counterions will be sodium ions movingmainly through paracellular routes.Both the absorptive and secretory processes

that have been described will influence watermovement, inwards and outwards respec-tively, across the epithelium. This is possiblebecause respiratory epithelium is water per-meable. This is not so for epithelia with lowwater permeability (for example, sweat duct),where salts can be absorbed without con-sequent fluid movement. The presence in thepulmonary system of these transport proces-ses, with a predominant, unregulated absorp-tive process plus a secretory process that canbe activated and regulated by autacoids andtransmitters, provides the opportunity for finetuning of the thickness of the fluid (sol) phaselining the proximal airways. The control ofthe thickness of the fluid lining has not beendefinitely demonstrated, but it is known thatboth these processes are present in epitheliumlining the proximal airways and, more impor-tantly, how they are modified in cystic fibrosis.

Abnormalities of epithelial ion transportin cystic fibrosisThe first insightful observation in relation tothe transport function in cystic fibrosis airwayepithelium was that the potential differenceacross the nasal mucosa and bronchi of patientswith cystic fibrosis was around twice that ofsubjects not suffering from cystic fibrosis.8 Onits own this result is difficult to interpretbecause either increased sodium absorption orreduced chloride conductance at the apicalsurface would give this result.When amiloride is superfused into the apical

surface of nasal epithelia in vivo the potential

125

on October 2, 2021 by guest. P

rotected by copyright.http://thorax.bm

j.com/

Thorax: first published as 10.1136/thx.46.2.124 on 1 F

ebruary 1991. Dow

nloaded from

Cuthbert

Normal CF

Na Na

Cl

cAMP]

Cl

Na Na[CAMP

Cl_lCl

Figure 3 Summary of the differences between airway epithelium in normal and cysticfibrosis tissues. In cystic fibrosis sodium absorption (Na) is exaggerated and chloridesecretion (Cl) is blocked, even in the presence of secretagogues that increase cyclic AMP(cAMP). In normal epithelium secretagogues increase chloride secretion whereas incystic fibrosis sodium absorption is enhanced. Excessive dehydration may remove the solphase, depositing mucus flakes with adherent bacteria on to the cilia.

difference falls. Both the absolute fall and thepercentage inhibition of the nasal potentialdifference was greater in patients with cysticfibrosis than in normal controls9 and this effectcould be mimicked by superfusing sodium freesolutions. In contrast, superfusion withchloride free solutions increased the luminalnegativity, but considerably more in normalsubjects than in patients with cystic fibrosis.9These additional findings support the notionthat both increased sodium transport andreduced apical chloride conductance are res-

ponsible for the potential difference in theairway epithelium in cystic fibrosis. Monolayersof cultured nasal epithelial cells from controlsand patients with cystic fibrosis showed similarpotential changes when subjected to amilorideor to low sodium or low chloride conditions.'0This means that cells retain their cystic fibrosischaracteristics through many generations ofcell division in culture, exactly what is expectedif abnormal characteristics are tied to a geneticdefect.

Since these early, largely in vivoexperiments, various lines of evidence haveconfirmed that in airway epithelium frompatients with cystic fibrosis sodium absorptionis enhanced and apical chloride permeability islow or absent, the latter preventing chloridesecretion in response to secretagogues that raiseintracellular cAMP. The detailed evidence forthis cannot be given here, but the centralfindings are summarised below with keyreferences.1 Isolated epithelial sheets from resectednasal turbinate material or nasal polyps were

used to measure active sodium absorption. Therate ofsodium transport in cystic fibrosis tissuewas about twice that in control tissue.'""'32 In cultured airway epithelial cells apicalchloride permeability was lower in cystic

fibrosis cells than in non-cystic fibrosis cells.'4 '53 By application of various agents thatactivate adenylate cyclase (beta agonists, for-skolin) to monolayer cultures of airway epithe-lial cells chloride secretion was stimulated innormal but not cystic fibrosis cells. In cysticfibrosis monolayers beta agonists further en-hanced sodium absorptive processes. Thesedifferences could not be explained by differencesin the generation of cAMP or in the amountof cAMP dependent protein kinases in thetissue. '"'"4 Bradykinin and A23 187, both of whichincrease intracellular calcium concentration(Ca-), are able to elicit transient chloridesecretory responses in cystic fibrosis epithelia.'9It is, however, possible that raised Cai activatesbasolateral K+ channels, leading to hyper-polarisation of the apical membrane, thusincreasing the electrochemical potential forefflux, even though the membrane has a lowchloride conductance. Figure 3 provides asummary of the comparisons in transportingproperties between cystic fibrosis and normalrespiratory epithelium, plus the effects of dif-ferent types of secretagogues.

Possible consequences of transportingabnormalities in cystic fibrosisMucociliary clearance depends on the func-tioning of cilia within the sol phase lining therespiratory passages. Not only is fluid movedtoward the proximal airways but mucus andentrapped foreign particles are moved to aposition from which they can be eliminatedfrom the lung. Excessive solute transport out ofthe airways, with consequent fluid absorption,may reduced the sol phase, depositing thick,viscid mucus on the cilia. Mucus deposits aredifficult to dislodge and provide a breedingground for Staphylococus aureus andPseudomonas aeruginosa. Although this is areasonable hypothesis there is little definitiveevidence to support it. The reason is a technicalone-it is difficult to sample fluid from theairways without evaporative losses, makingaccurate determination of composition verydifficult.The chemical composition of secreted mucus

is known to be altered in cystic fibrosis, withincreased sulphation of high molecular weightglycoconjugates.20 As cystic fibrosis is a singlegene defect it becomes increasingly difficult toexplain how properties as diverse as mucuscomposition and chloride conductance arerelated to a single abnormal form of the cysticfibrosis transmembrane regulator, thoughabnormal chloride transport may increase theconcentration of intracellular sulphate andincrease substrate driven sulphation ofglycoproteins.

Molecular correlates of abnormaltransporting propertiesPatch clamping is a relatively new techniquethat allows the functional activity of single ionchannels to be examined. Theoretically theconcept is straightforward. First devise a way

126

on October 2, 2021 by guest. P

rotected by copyright.http://thorax.bm

j.com/

Thorax: first published as 10.1136/thx.46.2.124 on 1 F

ebruary 1991. Dow

nloaded from

Abnormalities ofairway epithelialfunction in cysticfibrosis

of isolating a single ion channel so that itsactivity can be monitored, without interferencefrom the thousands of other ion channelspresent in a cell membrane. Secondly, find away of detecting whether the channel is open orclosed. This is done by monitoring the flow ofions through the channel with sensitive currentamplifiers. Ion channels allow millions of ionsto pass each second, giving rise to currents inthe picoampere (10-12 A) range. By measuringthe relation between the current and the voltageacross the channel and taking account of the ionconcentration either side of the membrane youcan assess channel selectivity. An all importantfactor is the reversal potential (the potential atwhich the current through the channel reversesits direction); application of the Goldman-Hodgkin-Katz equation and comparision ofactual and theoretical values allow the channelto be designated as selective for chloride,sodium, potassium, or calcium. To isolate asingle ion channel a fire polished glass electrodewith a tip diameter of a few microns is gentlypushed against the surface of a cell. A selfsealing process occurs such that the resistancebetween the conducting fluid in the electrodeand the bathing fluid outside is very large(gigaohm seal). The patch so isolated willcontain a few ion channels, or preferably asingle one. Recordings are made ofthe currentsflowing across the patch either in the cellattached mode or by pulling the patch off thecell.

Figure 4 shows a patch clamp record from achloride ion channel from a human cysticfibrosis sweat gland cell. Diagrammaticrepresentation of major findings in cysticfibrosis and airway epithelial cells is given infigure 5, which shows idealised diagrams ofpatch clamp recordings. Figures 5a and 5bshows the usual features seen with patch clam-ping. The current flowing through a particulartype of ion channel is constant at a givenvoltage-that is, the channel conductance ischaracteristic of the channel, and channels areeither open or closed. If two channels arepresent in a patch and are open simultaneouslythen the current flowing is twice that of a singlechannel. By contrast, the duration of the openand closed periods varies even for a singlechannel. In the simplest systems analysis ofpatch clamp records shows that open andclosed time histograms are fitted by singleexponentials, suggesting that a single process isresponsible for transitions from open to closed.Two groups have reported their findings on

chloride channels in airway epithelial cellsusing the patch clamp technique.2 22 In the cellattached configuration chloride channels wererarely seen unless the cell was activated byagents that increased cAMP. After stimulationthere was an increased probability of chloridechannel opening in normal cells (fig 5c) but notin cystic fibrosis cells (fig 5d). Again with the cell

attached configuration, the calcium ionophoreA23187 was able to activate chloride channelsin both normal and cystic fibrosis cells. Isolatedpatches from cells activated by cAMP generat-ing secretagogues showed chloride channelactivity (fig 5ei), whereas patches from similarly

I

4 pA10 ms

Figure 4 Patch clamp recordingfrom an apicalmembrane patch showing the activity of a single chlorideselective channelfrom a cysticfibrosis sweat gland cell.Channel opening is in the upward direction, indicatingchloride ionflowfrom outside the apical surface to inside.The solution bathing both sides of the patch containedpotassium (140 mmol/l), sodium (7 mmol/l). Theholding potential was 70 mv with the outside apicalsurface negative.

treated cystic fibrosis cells were electricallysilent (fig 5eii). Chloride channels could beactivated in cystic fibrosis patches either bystrong depolarisation or by exposure of thecytosolic face to calcium ions (fig 5eiii). Theeffects of these last two manoeuvres are not

(a) c

(b) c

(c) c

(d) c

(ei)cL(eii) c

(eiii) c

(fi) c

(fii) c

Figure S Theoretical patch clamp recordsfrom chloridechannels. c indicates the current level at which thechannel is closed. The patch may contain one (a) or morethan one (b) channel. Activation of a respiratoryepithelial cell with cyclic AMP (cAMP) increases theprobability that the chloride channels are open. In normalcells channel activation can be seen in the cell attachedconfiguration (c) but not with cystic fibrosis cells (d).When the patch is pulledfrom activated cells channelactivity continues in patchesfrom normal cells (ei) but isstill absent in patchesfrom cysticfibrosis cells (eii).Patchesfrom cysticfibrosis cells can, however, beactivated by strong depolarisation (eiii). When a patch isheld at a potential that prevents chloride channelactivation opening can be induced by a cocktail ofA TP,cAMP, and the catalytic subunit ofprotein kinase A(added at 0) in patchesfrom normal (fi) but not cysticfibrosis cells (fii).

127

on October 2, 2021 by guest. P

rotected by copyright.http://thorax.bm

j.com/

Thorax: first published as 10.1136/thx.46.2.124 on 1 F

ebruary 1991. Dow

nloaded from

Cuthbert

immediate or necessarily reversible and it is notclear how activation is achieved. Kunzelmannet al2' have shown that if isolated patches areformed at 37°C rather than at room tem-perature there is immediate activation ofchloride channels in cystic fibrosis cells withoutthe need to apply strong depolarisation. He hasargued that regulation of the chloride channelis defective rather than the channel itself.Possibly the cytosol in cystic fibrosis contains atonic inhibitor of the chloride channel that isremoved once an inside out patch is formed.Finally, as shown in figure 5f, when apicalmembrane patches are held at room tem-perature and at a voltage inhibiting channelactivation, addition of ATP and the catalyticsubunit of protein kinase A causes channelactivation in normal but not cystic fibrosismembranes. Thus there appears to be a phos-phorylation defect at a stage later than eithercAMP generation or the activation of proteinkinase A.24"2 There are two other findings with abearing on this problem. Firstly, the patchclamp findings have now been confirmed in animmortalised, virally transformed epithelialline developed from cystic fibrosis airway cells.This establishes that the distinctive cysticfibrosis phenotype is preserved in cells aftermany generations of culture in vitro,26 andtherefore cannot be related to the infectedconditions in cystic fibrosis airways. Secondly,apical membrane chloride channels incor-porated into lipid bilayers show propertiessimilar to those in normal membranes, butundergo a rapid run down. Activity can berestored by addition of the phosphorylating"cocktail" including the catalytic subunit ofprotein kinase A, indicating that a phos-phorylation event is concerned in the function-ing of airway chloride channels.27 Protein kin-ase C can also activate chloride channels inairway epithelial cells, but again regulation bythis kinase is defective in cystic fibrosis.28Whatever method has been used to activatecystic fibrosis chloride channels in vitro theirproperties have proved to be identical to thoseof normal channels-that is, conductance ofaround 50 pS at 0 mv, an outwardly rectifyingcurrent-voltage relationship, and a Cl- to Na+selectivity ratio of 10:1.Far less is known about sodium channel

mechanisms in airway cells. The mean channellifetime for sodium channels in cystic fibrosiscells is twice that ofnormal channels, providinga correlate for the increased sodium reabsorp-tion in cystic fibrosis.29 It has also been shownthat the macroscopic Kd for amiloride isincreased and there is an altered relation be-tween sodium concentration and sodium trans-port in cystic fibrosis epithelium.'0 These latterstudies were made with sweat duct epithelium,however, not airway epithelium. Nevertheless,a picture is emerging of altered molecularcharacteristics affecting both chloride andsodium channels in cystic fibrosis. From thepatch clamp data it is difficult to argue that thesodium channel effects are a result of aberrantchloride transport, yet both alterations in chan-nel properties are a consequence of a singlegene defect.

Future therapeutic strategiesAn experiment that many must now beattempting is to transfect various cells with thecDNA for cystic fibrosis transmembraneregulator. This approach may show whether ornot the cystic fibrosis transmembrane regulatoris the chloride channel. This seems a littleunlikely as altered sodium channel characteris-tics are retained in isolated patches. Moreprobably it will be shown that transfection ofcystic fibrosis cells with the cystic fibrosis trans-membrane regulator cDNA will correct the ionchannel abnormalities. In the long term trans-fection of airway epithelium in patients withcDNA in a retrovirus vector may be possible.Insufflation of the appropriate material into therespiratory tract may be easily arranged andconsiderable improvement might be envisagedeven from transfection of only a fraction of theaffected cells. Major obstacles need to be over-come before ethical permission is likely to begiven, not the least ofwhich is the possibility ofswitching on cellular oncogenes.An alternative approach is to correct the

transport irregularities with agents active fromthe apical surface. Nebulisation provides aroute of administration that is likely to confinethe pharmacological effect to the airways.Amiloride, an epithelial sodium channelblocker, is an obvious candidate and has beentried.3' Fourteen volunteer adults with cysticfibrosis took part in a study in which amiloridesolution or control vehicle was administered bynebuliser in a double blind crossover studywith two 25 week treatment periods. The mean(SEM) loss in forced vital capacity was reducedsignificantly from 3-4 (1 1) ml/day with vehicleto 1-4 (07) ml/day with amiloride. Further-more, the rheological properties of the secre-tions returned to virtually normal values dur-ing amiloride administration, though bacterialcontamination was unmodified. This trial raisesthe possibility that prophylatic treatment foryoung patients with cystic fibrosis before acuterespiratory symptoms have developed may beworthwhile. Analogues with greater potency ora longer duration-for example, benzamil32-might be more useful.As we have seen, evidence suggests that, at

least in airway epithelia, chloride secretion failsin response to secretagogues working viacAMP whereas agents that increase Cai are stilleffective. Simultaneous inhibition of sodiumabsorption and stimulation of chloride secre-tion may prove even more effective than theformer alone. There is an added incentive todevelop this strategy because when amilorideblocks the apical sodium channels in airwayepithelia a profound hyperpolarisation of theapical membrane results, exactly what isrequired to promote chloride exit from the cellprovided that apical chloride channels canopen. Thus a small increase in chloride con-ductance may have an amplified effect onchloride secretion in the presence of amiloride.Several ways of increasing chloride conduct-ance may be envisaged, via novel agents thatraise Ca, or development of chloride channelopeners working independently of normalregulatory mechanisms. Agents of the former

128

on October 2, 2021 by guest. P

rotected by copyright.http://thorax.bm

j.com/

Thorax: first published as 10.1136/thx.46.2.124 on 1 F

ebruary 1991. Dow

nloaded from

Abnormalities of airway epithelialfunction in cysticfibrosis

type are known (for example, thapsigargin33)but are only experimental tools. Bradykininpromotes chloride secretion in many types ofepithelia' and does not require cAMP for itsaction. Bradykinin receptors are usuallylocated basolaterally35 but in airways they arelocated apically.6 Bradykinin, however, alsocauses bronchoconst-riction, clearly undesir-able in cystic fibrosis, but is rapidly metabolisedby tissue peptidases. Some thought should begiven to developing analogues with appropriatepharmacodynamic and pharmacokineticproperties that may promote chloride secretionin airway epithelium.

In view of the similarity between the cysticfibrosis transmembrane regulator and multipledrug resistance proteins mentioned earlier wemay ask whether the cystic fibrosistransmembrane regulator performs a similarfunction, expelling unwanted endogenoussubstances from cells that otherwise alter thefunctioning of other systems, such as thebehaviour of sodium and chloride channels.The number ofendogenous materials which, inexcess, might have this effect is obviously largebut the possibility that a tonic inhibitor of thechloride channel exists is supported by theimmediate activation of these channels whenthe patches are removed from cells at 37°C.23Moreover, the volume of distribution of manydrugs is increased in cystic fibrosis, indicatingthat they can penetrate spaces from which theyare normally excluded. Is this too a manifesta-tion of an altered function of the cystic fibrosistransmembrane regulator? New informationabout the cystic fibrosis transmembraneregulator may promote questions where thereis already a body of clinically relevant informa-tion containing important clues. For example,the immunosuppressive drug cyclosporin hasbeen shown to label multiple drug resistanceproteins37 and possibly the cystic fibrosis trans-membrane regulator is also labelled; if it is, is itsfunction altered? This could be studied inpatients receiving transplants and cyclosporinto see whether they show any membranecharacteristics of cystic fibrosis and whetherthese disappear as cyclosporin dosage isreduced. Multiple drug resistance proteins areincreased in cells, particularly epithelial cells,when they are exposed to natural cytotoxicdrugs, such as vinblastine and vincristine. Ifthe cystic fibrosis transmembrane regulator canalso be induced, a search for non-toxic inducersmay lead to a useful treatment as cystic fibrosismay be a missense mutuation, where activity ofthe abnormal cystic fibrosis transmembraneregulator is not destroyed but impaired.The cystic fibrosis transmembrane regulator

with phenylalanine deleted at position 508accounts for only 70% of mutations. Otherhaplotypes, accounting for the other 30% ofmutations, are likely to be discovered soon.This information will provide further cluesabout which structural features of the cysticfibrosis transmembrane regulator are essentialfor function.

SummaryDetails of ion transporting abnormalities in

cystic fibrosis airway epithelium are nowknown. The central hypothesis, that excessivedrying of the airway surfaces is a primary eventthat leads to all the manifestations of therespiratory insufficiency in cystic fibrosis, is notproved. The hypothesis is, however, consistentwith the known transporting abnormalities andis strengthened by the modest clinical im-provement produced by a strategy designed tocorrect the transporting abnormalities. Thediscovery of the cystic fibrosis gene, togetherwith the presumed structure of the proteinproduct, provides a focal point that musteventually connect the functional abnor-malities with the genetic defect. The cellularfunction of the cystic fibrosis transmembraneregulator must now be the major target inresearch on cystic fibrosis. Strategies for treat-ment based on known cellular and molecularabnormalities are beginning to emerge but willbe undoubtedly more focused once the respon-sibility of the cystic fibrosis transmembraneregulator is known.

AddendumSince this review was completed the predictionmade under "Future therapeutic strategies"has been realised. Complementation studieshave shown that the introduction of plasmidscontaining cystic fibrosis transmembraneregulator cDNA into cultured cystic fibrosiscells does indeed restore a chloride conduc-tance response to cAMP.3839 Additionally, itnow appears that the cystic fibrosis trans-membrane regulator fails to be glycosylated incystic fibrosis cells, being retained in the golgiand not transferred to the cell membrane.40 Aplasma membrane location of the cystic fibrosistransmembrane regulator is consistent eitherwith an MDR like function or with its being adirect regulator of ion channel activity.

Results obtained in the author's laboratory were from worksupported by grants from the Cystic Fibrosis Research Trust.

1 Rommens JM, Iannuzzi MC, Kerem B, et al. Identificationof the cystic fibrosis gene: chromosomal walking andjumping. Science 1989;243:1059-65.

2 Riordan JR, Rommens JM, Kerem B, Zielenski J, Lok S,Plavsic N, Chou JL, Drumm ML, Iannuzzi MC, et al.Identification of the cystic fibrosis gene: cloning andcharacterisation of the complementary DNA. Science1989;245: 1066-72.

3 Kerem B, Rommens JM, Buchanan JA, et al. Identificationof the cystic fibrosis gene: genetic analysis. Science1989;245:1073-80.

4 Bradley G, Juranko PF, Ling V. Mechanism of multidrugresistance. Biochim Biophys Acta 1988;948:87-128.

5 Kilburn KH. A hypothesis for pulmonary clearance and itsimplications. Am Rev Respir Dis 1968;98:449-63.

6 CuthbertAW, Fanelli GM. Effects ofsome pyrazine carbox-amides on sodium transport in frog skin. Br J Pharmacol1978;63: 139-49.

7 Olver RE, Ramsden CA, Strang LB, Walters DV. The roleof amiloride-blockable sodium transport in adrenalineinduced lung liquid absorption in the foetal lamb. JPhysiol 1986;376:321-40.

8 Knowles M, Gatzy J, Boucher R. Increased bioelectricpotential difference across respiratory epithelia in cysticfibrosis. N Engi I Med 1981;305: 1489-95.

9 Knowles M, Gatzy J, Boucher R. Relative ion permability ofnormal and cystic fibrosis nasal epithelia. J Clin Invest1983;71: 1410-7.

10 Yankaskas JR, Cotton CU, Knowles MR, Gatzy JT,Boucher RC. Culture of human nasal epithelial cells oncollagen matrix supports. Am Rev Respir Dis1985;132:1281-7.

11 Knowles MR, Stutts MJ, Spock A, Fischer N, Gatzy JT,Boucher RC. Abnormal ion permeation through cystic

129

on October 2, 2021 by guest. P

rotected by copyright.http://thorax.bm

j.com/

Thorax: first published as 10.1136/thx.46.2.124 on 1 F

ebruary 1991. Dow

nloaded from

Cuthbert

fibrosis regulatory epithelium. Science 1983;221:1067-70.12 Knowles MR, Stutts MJ, Yankaskas JR., Gatzy JT, Boucher

RC. Abnormal respiratory epithelial ion transport incystic fibrosis. Clin Chest Med 1986;7:285-97.

13 Cotton CU, Stutts MJ, Knowles MR, Gatzy JT, BoucherRC. Abnormal apical cell membrane in cystic fibrosisrespiratory epithelium. An in vitro electrophysiologicalanalysis. J Clin Invest 1987;79:80-5.

14 Widdicombe JH, Welsh MJ, Finkbeiner WE. Cystic fibrosisdecreases the apical membrane chloride permeability ofmonolayers cultured from cells of tracheal epithelium.Proc Natl Acad Sci 1985;82:6167-71.

15 Stutts MJ, Cotton CU, Yankaskas JR, et al. Chloride uptakeinto cultured airway epithelial cells from cystic fibrosispatients and normal individuals. Proc Nat! Acad Sci1985;82:6677-81.

16 Boucher RC, Stutts MJ, Knowles MR, Cantley L, GatzyJT. Na+ transport in cystic fibrosis respiratory epithelia.Abnormal basal rate and response to adenylate cyclaseactivation. J Clin Invest 1986;78:1245-52.

17 Widdicombe JH. Cystic fibrosis and #-adrenergic responseof airway epithelial cell cultures. Am J Physiol1986;251:R818-22.

18 Barthelson R, Widdicome J. Cuyclic adenosine monophos-phate dependent kinase in cystic fibrosis tracheal epi-thelium. J Clin Invest 1987;80:179-802.

19 Boucher RC, Cheng EHC, Paradiso AM, Stutts MJ,Knowles MR, Earp HS. Chloride secretory response incystic fibrosis human airway epithelia. Preservation ofcalcium but not protein kinase C and A dependentmechanisms. J Clin Invest 1989;84:1424-31.

20 Cheng PW, Boat TF, Cranfill K, Yankaskas JR, Boucher RC.Increased sulphation of glycoconjugates by cultured nasalepithelial cells from patients with cystic fibrosis. J ClinInvest 1989;84:68-72.

21 Welsh MJ, Liedtke CM. Chloride and potassium channels incystic fibrosis airway epithelia. Nature 1986;322:467-70.

22 Frizell RA, Rechkemmer G, Shoemaker RL. Altered regula-tion of airway epithelial cell chloride channels in cysticfibrosis. Science 1986;233:558-60.

23 Kunzelmann K, Pavenstadt H, Greger R. Properties andregulation of chloride channels in cystic fibrosis andnormal airway cells. Pflug Arch 1989;415:172-82.

24 Li M, McCann JD, Liedtke CM, Nairn AC, Greengard P,Welsh MJ. Cyclic AMP-dependent protein kinase openschloride channels in normal but not cystic fibrosis airwayepithelium. Nature 1988;331:358-60.

25 Schoumacher RA, Shoemaker RL, Halm DR, Tallant EA,Wallace RW, Frizzell *RA. Phosphorylation fails toactivate chloride channels from cystic fibrosis airway cells.Nature 1987;330:752-4.

26 Jetten AM, Yankaskas JR, Stutts MJ, Willumsen NJ,

Boucher RC. Persistence of abnormal chloride conduc-tance regulation in transformed cystic fibrosis epithelia.Science 1989;244:1472-5.

27 Valdivia HH, Dubinsky WP, Coronado R. Reconstitutionand phosphorylation of chloride channels from airwayepithelial membranes. Science 1988;242:1441-4.

28 Hwang T-C, Lu L, Zeitlin PL, Gruenert DC, Huganir R,Guggino WB. Cl' channels in CF: lack of activation byprotein kinase C and cAMP-dependent protein kinase.Science 1989;244:1351-3.

29 Disser J, Fromter E. Properties of Na+ channels in res-piratory epithelium from CF and non-CF patients[abstract]. Pediatr Pulmonol 1989;suppl:1 15.

30 CuthbertAW, Brayden DJ, Dunne A, Smyth RL, WallworkJ. Altered sensitivity to amiloride in cystic fibrosis.Observations using cultured sweat glands. Br J ClinPharmacol 1990;29:227-34.

31 Knowles MR, Church NL, Waltner WE, et al. A pilot studyofaerosolised amiloride for the treatment of cystic fibrosislung disease. N Engl J Med (in press).

32 Aceves J, Cuthbert AW. Uptake of 'H-benzamil at differentsodium concentrations. Inferences regarding the regula-tion of sodium permeability. J Physiol 1979;295:491-504.

33 Brayden DJ, Hanley MR, Thastrup 0, Cuthbert AW.Thapsigargin, a new calcium-dependent epithelial anionsecretagogue. Br J Pharmacol 1989;98:809-16.

34 Cuthbert AW, Margolius HS. Kinins stimulate net chloridesecretion by rat colon. Br J Pharmacol 1982;75:587-98.

35 Cuthbert AW, MacVinish LJ. Diversity of kinin effects ontransporting epithelia. In: Kinns V. Part A. Abe K,Moriya H, Fujii S, eds. New York: Plenum, 1989:103-11.

36 Leikauf GD, Ueki IF, Nadel JA, Widdicombe JH. Brady-kinin stimulates chloride secretion and prostaglandin E,release by canine tracheal epithelium. Am J Physiol1985;248:F48-55.

37 Foxwell BMJ, MacKie A, Ling V, Ryffel B. Identification ofthe multidrug resistance-related P-glycoprotein as acyclosporine binding protein. Mol Pharmacol 1989;36:543-6.

38 Gregory RJ, Cheng SH, Rich DP, Marshall J, Paul S,Henirk K, et al. Expression and characterization of thecystic fibrosis transmembrane conductance regulator.Nature 1990;347:382-6.

39 Rich DP, Anderson MP, Gregory RJ, Cheng SH, Paul S,JeffersonDM,etal. Expressionofcysticfibrosistransmem-brane conductance regulator corrects defective chloridechannel regulation in cystic fibrosis airway epithelial cells.Nature 1990;347:358-63.

40 Cheng SH, Gregory RJ, Marshall J, Paul S, Souza D, WhiteGA, et al. Defective intracellular transport and processingofCFTR is the molecular basis ofmost cystic fibrosis. Cell1990;63:827-34.

130

on October 2, 2021 by guest. P

rotected by copyright.http://thorax.bm

j.com/

Thorax: first published as 10.1136/thx.46.2.124 on 1 F

ebruary 1991. Dow

nloaded from