Nerve growth factor in the urinary bladder of the adult regulates neuronal form and function

Nerve growth factor in the urinary bladder of the

adult regulates neuronal form and function.

W D Steers, … , D Creedon, J B Tuttle

J Clin Invest.

1991;

88(5)

:1709-1715.

https://doi.org/10.1172/JCI115488

.

Urethral obstruction produces increased voiding frequency (0.7 +/- 0.06 to 1.1 +/- 0.08 h-1)

and hypertrophy of the urinary bladder (89 +/- 1.7 to 708 +/- 40 mg) with profound

increments in the dimensions of afferent (4, 6) and efferent neurons (299 4.7 to 573

+/-8.6 microns2) supplying this organ in the rat. We discovered that hypertrophied bladders of

rat and human contain significantly more nerve growth factor (NGF) per milligram wet

weight, protein, and DNA than normal bladders. The temporal correlation between NGF

content, neuronal hypertrophy, and bladder weight was consistent with a role for this growth

factor in the neurotrophic effects associated with obstruction. Autoimmunity to NGF

abolished the hypertrophy of NGF-sensitive bladder neurons in the pelvic ganglion after

obstruction. Relief of urethral obstruction reduced bladder size (349 +/- 78 mg), but neuronal

hypertrophy (460.2 +/- 10.2 microns2) and elevated NGF levels were only partially reversed.

Bladder hypertrophy (133 +/- 4.3 mg) induced by osmotic diuresis slightly increased

ganglion cell area (365.2 +/- 6.1 microns2) and only doubled NGF content of the bladder.

These findings provide important new evidence that parenchymal cells in the hypertrophied

bladder can synthesize NGF and possibly other molecular messengers that act to alter the

size and function of neurons in adult animals and man.

Research Article

Nerve Growth

Factor in the

Urinary

Bladder

of

the Adult

Regulates

Neuronal Form and Function

W. D.Steers,*S.Kolbeck,*D. Creedon,*andJ. B. Tuttle

Departmentsof*Urology,tPhysiology, and0Neuroscience, University of Virginia Health Sciences Center, Charlottesville, Virginia 22908

Abstract

Urethral obstruction produces increased voiding frequency

(0.7±0.06 to 1.1±0.08 h-') and hypertrophy of the urinary

bladder(89±1.7to708±40 mg)withprofound increments in thedimensions of afferent (4, 6)andefferent neurons(299±4.7

to573±8.6

gm2)

supplyingthis organ in the rat. We discovered thathypertrophiedbladdersof rat and human contain

signifi-cantly more nerve growth factor (NGF) per milligram wet weight,protein,and DNA than normal bladders. The temporal

correlation betweenNGF content, neuronalhypertrophy,and bladderweightwasconsistentwith a rolefor thisgrowth factor intheneurotrophiceffectsassociatedwithobstruction. Autoim-munity to NGF abolished the hypertrophyofNGF-sensitive

bladder neurons in thepelvic ganglionafter obstruction.Relief

of urethral obstruction reduced bladder size(349±78mg), but neuronal hypertrophy (460.2±10.2

Mm2)

and elevated NGF lev-els were onlypartiallyreversed. Bladderhypertrophy (133±43

mg)induced by osmotic diuresis slightly increased ganglion cell area (365.2±6.1

gm2)

and only doubled NGF content of the bladder. These findings provide importantnew evidence that

parenchymalcells in the hypertrophied bladder cansynthesize NGFandpossiblyothermolecular messengers that act to alter thesizeandfunction of neurons in adult animals and man. (J.

Clin.

Inwest.

1991.88:1709-1715.)Key words:

hypertrophy.

outletobstruction * neural plasticity* smooth muscle-major

pelvicganglion

Introduction

Urinary

bladder hypertrophy is produced by anatomical or

functional obstruction ofthebladder outlet. The most com-mon cause ofobstruction in man is benign prostatic

hyper-trophy

(BPH)',

the treatmentof whichrepresents the second

costliest expenditure of Medicare funds in theUnited States

Partof thiswork appearedinabstractform (Creedon,D., W.Steers, and J.Tuttle. 1989.Abstr. Soc. Neurosci. 16:301).

Address reprint requests to J. B. Tuttle, Ph.D., Department of Neuroscience,Box230,University of Virginia Health Sciences Center,

Charlottesville,VA22908.

Receivedforpublication 13 February 1991 and in revisedform 27 June 1991.

1.Abbreviationsused in this paper:Al, autoimmunization;BPH, be-nign prostatic hypertrophy;CONT, control; CPRG,chlorophenol

red-beta-D-galactopyranoside; DELIG,deligation; DIUR, diuresis, DRG, dorsalrootganglia; FG,Fluoro-Gold;NGF, nerve growthfactor,OBS, obstruction; TH-IR,tyrosine hydroxylase.

J.Clin. Invest.

©TheAmericanSocietyfor ClinicalInvestigation,Inc.

0021-9738/91/11/1709/07 $2.00

Volume88,November 1991, 1709-1715

(1).

Amanifestation of bladderdysfunctionseenin60-75% of

patients with BPH is bladder hyperactivity (2). Because ob-structionof the bladder leadstochangesin neuronal

morphol-ogyandfunction, these alterations in neural pathways could contribute to bladder dysfunction and alter treatment

out-come(3-7).

Ithasbeen proposed that dynamic interactions occur be-tweenatarget tissue and itsinnervationin the adult (8, 9).In support of this idea experimental manipulations ofa target

alter the geometry of innervating neurons. For example, creat-ing large targets results in more morphologicalcomplexityin neural pathways, whereas reductions in target size produce

smaller, less complex neurons (8, 9). This interaction betweena

tissue and itsnervesupplyhas been attributed to the action of

asyetunidentified molecular messengers. One candidate that fits criteria for an intercellular signal molecule between the target organ and nerve is nervegrowthfactor(NGF) (10). How-ever, aspecific role for NGF in the regulation of neural

plastic-ity in vivo or in a disease process in the adult has not been

established.

The accessibility ofperipheral autonomic neuronsprovides

anopportunity to perform a detailed analysis of the relation-ships between neuronal geometry and the function or size of the target innervated. One target organ whose function requires autonomicinput is the urinary bladder. Thus, bladder hyper-trophy induced by partial obstruction may offer a clinically

relevant model thatprovidesinsightintoneurotrophic interac-tions in vivo.

Indeed,partial urethral ligation inthe rat,which produces

bladderenlargementandsmoothmusclehypertrophy,is asso-ciated with morphological and physiological alterations in the neural pathways regulating micturition. Morphological

changes include hypertrophy ofretrogradelylabeled afferent (5-7) and efferent (4, 7) neurons supplying the bladder, as well as increased labeling of afferentprojections from the bladder

surrounding preganglionicneurons in the sacralspinalcord (6, 7). Physiological alterationsare seen in multiunit recordings

from bladder nerves (3). Enhanced spinal reflex discharges ap-pear afterobstruction whicharelinkedto anincrease in

void-ingfrequency(3).Decentralizationofthe bladderfailsto pre-vent thesizeincreasesofneurons induced by obstruction (4).

Taken together these findings provide strong evidence that

neurotrophic interactions occur between the bladder and its

innervation and that these interactions are altered after ob-struction.

Itisproposed that obstruction stimulates parenchymal cells in the bladder to produce a factor which regulates size of the neurons supplying this organ. Thishypothesis was tested by

assaying

normal andhypertrophiedbladders forNGFtoassess the potential involvement of this

single

trophic factor on

pre-vented the hypertrophy ofa population of efferent nerves

in-nervatingthe bladder.

These findingshave profound investigational andclinical ramifications. They provide important new evidence for

dy-namic interactions betweenparenchymal target cells andtheir innervation inthe adult. It can be speculated thatalterationsin NGF or othertrophic factors participate in the sensory distur-bances and bladder hyperactivity observed in patients with BPH,suggestingnovel avenues fortherapy.

Methods

Experimental

groups

Obstruction(OBS).Amodification of the technique described by Mat-tiasson and Uvelius (3, 4, 11) was used to create partial bladder outlet

obstructionin13 Wistar rats (222-250 g). Under halothane anesthesia, theskinwasshaved, prepped with an iodine/alcohol mixture, and a low midline incision made. The urethral diameter was then reduced to 1 mmbytying two 4-0 nylon sutures around the urethra and an extralu-minally-positioned(1 mm O.D.) polyethylene tube. The tubing was removed and theabdominalincision reapproximated using two layers of interrupted nylon sutures. 12 control (CONT) rats underwent sham surgery in which the urethrawascircumferentiallydissected, but not

ligated.Occasionally, aligatedanimal developed urinary retention or wasincontinent as monitored by voidingfrequency,presumably due to

overflowincontinence. These animals weresacrificedimmediately and notincluded in the study.

Diuresis (DIUR). Bladderhyperactivitywithout obstruction was produced ineightratsby inducingosmoticdiuresis with 5% (wt/vol) sucroseintheir ad libwatersupply.

Deligation(DELIG).6 wk after urethral ligation, six rats underwent

surgical removal of urethralligatures under halothane anesthesiato

examine reversibilityof the changes invoiding frequency,neuronsize,

and levelsofNGF.

Autoimmunization (AI).30 femalerats wereimmunized with 100 ,ug murine NGF incompleteFreund'sadjuvantfollowedby1i5-g

in-jectionsinincompleteadjuvantat4-5-wk intervals(12).Successful Al was evaluated in each animal bytesting for circulatinganti-NGF activ-ity usingamodifiedneurite outgrowth bioassay. Dorsal rootganglia (DRG)from 8-9-d-old chick embryoswerecultured inequal20-Ml

portions ofchickenplasma, thrombin (I mg/ml)inmodifiedEagles mediumand dilutedrattailveinseruminculture medium with 15 ug/ml murine NGF. Positive controls were included with titers of known, purifiedanti-NGF. Under theseconditions fiberoutgrowth

wasprofusewhennoantiserumwasadded. Animalswereconsidered

autoimmunetoNGF if1/100to 1/1,000dilution ofserumprevented

neurite outgrowth. Neuriteoutgrowthwasevaluated usingaNikon

Diaphot inverted microscope and phasecontrast optics. Fiber

out-growth waseasily distinguishedfromoutgrowthofexplantsupporting cells.

Autoimmune+obstructed (AJOBS).Six Al animals(see

Autoim-munization)wereobstructed(see Obstruction)todeterminewhether

endogenous antibodytoNGFpreventedthe neuronalgrowthand

uro-dynamicchangesaccompanyingobstruction.

Neuronal labeling.Bladderneurons werelabeledby retrograde

ax-onal transport of Fluoro-Gold(FG; Fluorochrome, Inc., Englewood,

CO)becauseganglioncells in themajor pelvic ganglionsupplyseveral organs.Eightseparateinjections (3-4

l)

of4%FG(wt/volindistilled water) were made into the bladder wallusingaI0-MlHamiltonsyringe witha30 G needle. The needlewasleft inplacefor 30sand the site rinsed with saline then swabbed withacotton-tipped applicatorto

min-imizeleakageoftracer. FGinjectionwasperformed48 hbefore

re-movalofganglia. Injectedbladderswerenotused for NGF assays. Humansubjects. 2-g specimensof smooth muscle from the anterior bladder wallweretaken for NGFanalysisfrompatients (n=₃₎

under-goingsuprapubicprostatectomy forobstructionduetoprostatismor

urinarydiversion for functional bladder outlet obstruction

resulting

from spinal cord injury. Bladders from obstructed subjects were grossly thickened and trabeculated. Voiding flow rates were<10ml/s consis-tentwith obstruction. Obstructive symptoms such as diminishedurine flow,nocturia, and hesitancy and irritative symptoms of urinary ur-gency and frequency were present for at least 1 yr. Control (n = ₃₎

specimenswereobtained from patients undergoing urinary diversion for incontinenceorbladdercancer.Specimens were not obtained from areaswhichcontainedmalignancy.Control subjects did not have ob-structive symptomsorflow rates<20 ml/s. Tissues were obtained in accordance with the University ofVirginia Human Investigation Com-mittee.

Analysis

Measurements of voiding frequencies, neuronal areas, and NGF con-tentwere made in a single blinded fashion with the investigator un-awareof which group of animals were being analyzed.

Urinaryfrequency.Voiding frequencywasused as an estimate of bladderactivity in the experimental animals. Animals were housed in individual cages. Voidingfrequency wasassessedin a single blinded fashiononfour separate occasions in each animal by visualizing dis-tinct urinemarkingsonfilter paper placed beneath each cage during a 3-htestperiod (16:00-19:00h) withmonitoringof fluid intake. Ob-struction didnotalterwaterconsumption.

Morphometrics.

Pelvicgangliawereremoved in halothane-anesthe-tizedratsandplacedin4%buffered formalin for4h.Ganglia were washed in PBS andequilibratedinsuccessivePBS solutions containing 10, 20, and 30%sucroseforcryoprotection.Cryostat sections (14

gm)

werecutandmounted ongelatinizedslides, dehydrated usingaseries of alcohols, mounted inglycerin,andcoverslipped.FG-labeledganglia

werethen viewed using a NikonFXAmicroscope with UV fluores-cence(filterUV excitationat360-380nm).

Cross-sectionalareasof FG-labeled cells in every fourth section and distributed throughout thepelvic gangliaweremeasured using comput-erized image processing (Bioquant) and videomicroscopy. Video

imageswereobtained throughaDage-MTI Inc.(Wabash,MI)70 series cameraat amagnification of200. - 200 cells from eachganglionwere manually outlined forareadeterminations.Forcomparison, cross-sec-tional areas as measured using camera lucida drawings were 10%

smaller than those obtained through image analysis software. The lower limit for neuronalcross-sectionalareas was set at150

_;m2

based onpreviousmeasurementsofprincipalcells in themajor pelvic

gan-glion (4).

Tissue assays.Urinarybladders and other tissueswereremoved under halothaneanesthesia,blotteddry,weighed,frozenondry ice, andstoredat-70'C. Foranalysis,tissueswereshatteredatliquid ni-trogen temperature inametalimpactmortar,and solubilized in4vol oraminimum ofl-ml sample buffer(0.1 MNaHPO4,0.4MNaCl, 0.1%TritonX-100,0.5%BSA, 0.1mMbenzethonium-Cl,2 mM

ben-zamidine,2 mMEDTA,0.1 mMPMSF,20KIU/ml

aprotinin).

The tissuewasthenhomogenizedfor 1 min and thencentrifugedat_33,000

g for 1 h. To estimate NGF recovery, 150pg/mlofauthenticNGFwas addedto an aliquot before thecentrifugation. Aliquots of the total homogenatewerealsoassayedforprotein

(13)

and DNA

(14).

Atwo-site ELISA for NGFwasadaptedfrom the methods of West-camp and Otten

(15)

andHellweg

(16).

Maxisorb 96-wellplates (Nunc,

Copenhagen, Denmark)werecoated with 150 Ml of 0.5 ug/ml anti-NGF monoclonalantibody(BoehringerMannheim

Biochemicals,

In-dianapolis, IN)in50 mM ofsodium carbonatebuffer, pH 9.6,for 2 hat

360C.

Remainingsitesontheplatewereblocked with 200Mlofa1.0% BSA inacarbonatebufferat1 hat360C.Theplatewasthen washed 1

x 1 min, 3 x 5 min with 250 Ml of wash buffer

(0.4

MNaCland 0.1%

TritonX-100in0.1Mphosphate

buffer,

pH

7.4).

The final

sample

was

replacedwith 100Ml of eithersampleorknownamountsof NGF di-luted insample buffer.Four wellswereused for each

sample

and for each value of the standardcurve.Thesampleswereincubated

over-nightatroomtemperature inahumidifiedchamber. The

following day

10000

E

Lto N-L0

ga) .0

.0 4

1000

100.

10

10 100

NGF (pg/ml)

Figure1.Standardcurvefor NGF ELISA. Plotwaslinearto1 pg/ml of NGF. Mouse NGFwasusedto constructstandardcurve. Absor-bance obtainedat575nm.Brokenline represents three standard de-viations from blank specimen.

MannheimBiochemicals), was then added for 2 h at360C.Unbound

fl-galactosidase

was removedduring 1x1min,3x 1 h washes inbuffer with intermittent agitation. During the last buffer wash, 44 mgof chlo-rophenol red-beta-D-galactopyranoside (CPRG; Boehringer Mann-heim Biochemicals) was dissolved in 22 ml of 10mM Hepes, 150 mM

NaCl,2mMMgCI2, 0.1%(wt/vol)Na-azide, and 1%(wt/vol) BSA, pH 7.0. Theplatewasrinsedoncewith 250_,lofCPRGbuffer and 200Mlof theCPRG solutionwasaddedfor 4-6h at36°Corovernightatroom temperature. Theplatewasreadon aTitre~tekplate reader(ICN/Flow Laboratories, Inc., McLean, VA) using a 575-nm filter.

The standardcurve wascalculatedaftersubtracting the mean ab-sorbancereadingof sample buffer alone based uponaleast squaresfit (Fig. 1).Any valuefalling within three standard deviations of the zero NGFreadingwasconsidered below the limit of detection, typically

occurringat < 1 pg/ml (Fig. 1).

The mouse NGF usedfor the standard curve was the generous gift ofDr.Eugene M. Johnson (St. Louis, MO), and wasfreshly diluted in

200

--- freQ

--A* area

160

120

80 2300

-'A- /gm wwt -*- /100 mg

protein

, 1150

0 2 4 6

Weeks of Obstruction

sample buffer for each assay. Controls for the assayalwaysincluded

bindingnonimmunemouseIgGtotheplateand thezeroNGFpoint

onthestandardcurve. Specificity for antibody bindingtoNGF was tested withinsulin,astructurallysimilar molecule ofasimilar size. The standard curve for NGFtypically was linear to 600pg/ml, beyond whichsaturation occurred. However, 1

gg/ml

insulin was below the levelof detection. The absorbance values for the standards varied, but theslope of the standard curve was consistent. The specificity of the assay for NGF comparedtootherneurotrophinsin thesamefamily (BDNF andNT3) couldnotbeassayed duetolack ofsufficiently char-acterized reagents. Thus, the tissue samples may have contained some levelsof cross-reactive factors structurally similartoNGF.

Human tissueswereassayed by W. Wong and G. Bennett at Genen-tech, Inc.(SouthSanFrancisco, CA) usingasimilar two-site ELISA but with antiserum specific for human NGF. Astandardcurve was ob-tainedusingrecombinant human NGF andanhorseradish

peroxidase-linked secondary antibody.

Statistics. Inallstudiesmeancross-sectionalareasfor FG-labeled neurons werecomparedusing an ANOVA whilevoidingfrequencies, bladderweights,andNGFcontent wereanalyzed witha nonparamet-ricMann/Whitney U test using SPSS PC+ software. Differences in the distributions of neuronalareas wereanalyzedusingtheSmirnofftest. Data presented as mean±SE.

Results

Bladderweight increased from 89±1.7 mg in the control group

to708±40 mg after 6 wk of obstruction(OBS) (Fig. 2 A). The averagefrequency of urinary voiding increased from 0.7±0.06 to1.1±0.08h-'afterobstruction confirming the development ofbladderhyperactivity (Fig. 2 A). The mean cross-sectional areaofpelvic ganglioncellsinnervating the obstructed bladder rosesignificantlyreaching 573.0±8.6

,m2

at 6 wk (P < 0.001) (Fig. 2 A). Neuronal hypertrophy was detected after 1 wk, lag-ging behind the rapid rise in bladder weight.

Bladderenlargement was accompanied by a rise in NGF

content which preceded increases in voiding frequency and neuronal area (Fig. 2). Likewise, the amount ofNGF in grossly

600

400 1

200

0 Figure 2. Time course for alterationsin bladder

3.50 weight, function, and neuronal sizeafter obstruction. --- . CXA Bladderweight in milligrams(-),whereasvoiding

frequencies(+) and neuronal areas (A) are expressed aspercent control(0.7±0.05h-',300±4.9 Am2for 750

cells).

Neuronalsizewasmeasuredas

cross-sec-1.75 tional area of retrogradely labeled pelvic ganglioncell

somata atIwk (663 cells), 3 wk (575cells), and 6 wk (750 cells). Bladder weightrisesbefore theincreases in voiding frequency and neuron size. (B) Bladder tissue NGF. NGFis expressed as

(.)

nanograms per100mg protein of the whole homogenate; (A) nanograms per

0.00 _{gram wetweight;and (+) picograms per microgram}

DNA.NGF expressed per wet weight or protein peaks beforechangesin neuronal area or voiding frequency.

I

hypertrophiedhuman bladders (7.8 pg/mg pro, 1400 pg/g wet weight) exceeded that innonhypertrophied samples(2.6 pg/mg pro, 400 pg/g wet weight) from the patientsundergoingurinary diversion forincontinence.

Tissue levels of NGF were assayed in organs other than the

urinarybladderfromnormal andobstructedanimals to

ascer-tain whethertheincreasesin bladder NGF were specific for the

urinarytract. No significantdifferenceswere seen in NGF con-tent forabdominal aorta or lung specimens (CONT, n = 5;

OBS,n=6).NGFwasnot detected in striated or cardiac

(ven-tricular)musclesamples.

IfNGF isresponsibleforone or more aspects of the neuro-nal plasticity after bladderobstructionandhypertrophy,levels ofthistrophic factorshould rise before increases in neuron size andvoiding frequency. Indeed, theamountofNGF peaked 1 wkafterobstructionwhether the data areexpressedperwhole organ,milligram protein,orgram wet weight of bladder tissue

(Fig. 2 B). Over the ensuing weeks NGF levels dropped below

peak values but remained fourto six timeshigher than control

values. WhennormalizedtoDNA, NGF contentcontinuedto

riseafter3 wk(Fig. 2B).

After deligation,neuronal size(460.2±10.2

'Um2)

and NGF

levels failed to return to control values (Fig. 2), despite the

significant(P <0.05)decreasein bladder weight(349±78mg) and return of voiding frequency to CONT levels (0.73

±0.08

h-1).

In ratswith bladder hypertrophy inducedbydiuresis, blad-derweight (133±4.3 mg), voiding frequency (1.0±0.09

h-'),

neuronal area(365.2±6.1

'Um2),

and NGF levels were elevated compared to control. Yet,weights,areas, and NGF levelsfrom thediuresisgroup weresignificantly(P <0.01to P < 0.001) less

than those intheOBSgroup.

Thetemporalcorrelation between NGF, neuronal hyper-trophy, and bladder weight doesnotestablish NGFasthe

caus-ativefactor for neurotrophic effects. Thus,wesoughttoobtain

more conclusive evidence by using endogenous antibody againstNGFwhich is effectivein

disrupting

theaccessof neu-rons tothenutritive supply of this

trophic

substance (10, 12). Bladderweights in the Algroup were slightlygreater(1 16±4 mg)thancontrol.Despite this slight increaseinbladderweight,

the mean areaforneurons

innervating

thebladder from

ani-malswith endogenous

antibody

against NGF

(Al) (Fig.

3)was

less(279.3±3.4

jim2,

P<0.05)than age and

weight-matched

control (300.2±4.9

gm2).

Thepercent ofneuronswithareas > 500

jim2

was

substantially

reduced when obstruction was created in theAl group

(Fig.

4,

AIOBS-OBS).

This inhibition

accounted for the reduced mean area for the AMOBS group

(475.7±6.5

gm2)

comparedtoOBSanimals withno

circulating

anti-NGF (573±8.6

gm2,

P <

0.001).

Bladder

weights

in the

AI-OBSgroup(800±65 mg)weresimilartoOBST.

In contrast toareas,

voiding

frequency

inAIrats was ele-vated(1.14±0.08

h-1,

P<

0.05)

compared with control. This increase in

voiding

frequency

may have accounted for the

slight increasein bladder

weights

inthe AI group.Obstruction ofAlanimalselicited afurther increasein

voiding

frequency

(1.8±0.1) whichexceededthatof the OBSgroup

(P

<

0.05).

Discussion

Our results support the concept that the geometry of

auto-nomic neurons intheadultis influenced by ongoing

interac-A

0 C:

a)

:3

a)

B

25

20

15

10

5

02

25

° 20

ol

>y 15

U a) 3 10 1.

U)

L.

LL 5

o

25

C

NO 20

ol

>V 15 0

a) 3 10 a)

0

CONT OBS

150I

5001000

50 20

150 500 1000 1500 2000

LIIICONT Al

150 500

150 500

1000 1500 2000

1000 1500 2000

Neuronal Area (urn2)

Figure3. Area distributionsfor retrogradelylabeledbladder neurons in thepelvic gangliainnormal(CONT), obstructed(OBS),

autoim-mune(AI), andAI +OBSanimals. (A) 6 wk afterpartialurethral li-gationanincreaseinmean area(573±8.6jum2 for 750 cells)wasdue

to arightwardshift in distribution for cell sizes. Note that theentire population shiftswith the emergenceofcellswithareas >1,000Am2.

(B) Autoimmunizationproducesalossoflabeledneurons(611cells)

with areas > 800jm2.(C)After6 wkof obstructionofAIrats(715 cells), theoccurrenceof cells in the largest size categories is reduced comparedtoAand neuronsrarelyhave areas > 1,000 jm2.

tions with the tissues they innervate (8, 9). Consistent with this

hypothesis, obstruction-inducedhypertrophyof the small in-testine stimulates growth of neurons in themyentericplexus

(17, 18). Likewise, Voyvodichas shown hypertrophy of the submandibularglandincreases somal cross-sectionalareasand creates more complex dendritic arborization of innervating neurons(9).

Similarly theurinary bladderundergoesanatomical,

AIOBS-OBS

.y

.9 0

*

6

IZI OBS-CONT

500 1000 1500 2000

Nefonal Area (un2)

Figure4.Differences in the frequencies of bladder neuron size in the

pelvicganglion with obstruction and NGF autoimmunity. The dif-ferences in percentneuronsin each 50 um2 bin are plotted for ob-structed(OBS) minus control(CONT) frequencies(openbars)and NGF autoimmune obstructed (AIOBS) minus obstructed (OBS)

fre-quencies (solid bars).Apositivevalueindicates an increase in the proportion of cells in that bin, whereas a negative value represents a decrease.Obstruction (OBS-CONT) increases theproportion oflarge neurons with areas>500ym2.However,AIOBSanimals have a rel-ative deficiency in the largest sizes (AIOBS-OBS) (P<0.001). This implies that antibody prevented hypertrophy of NGF-responsive neurons.

possibility

that the bladderusesNGFtoinitiate ormaintain

these alterations in neural

pathways.

Whereas thecurrentdata

only address size changesintheMPG,adultratDRGneurons

havebeenshowntorespondtoNGF

(12).

Additional studies

areunderwaytoexamine the effect of increasedbladder NGF oncentral afferent

projections

andelectrophysiological

alter-ations after obstruction of the bladderoutlet.

NGFcontentin the bladderpeaked 1 wkafter

obstruction,

before

hypertrophy

of bladderneurons. When normalizedto

DNAcontent,NGF continuedtoriseafter 3wkofobstruction.

Thesefindingssuggestthat the initial rise in NGFwas

outpaced

by bladder

hypertrophy.

Thenumber ofneurons

innervating

the obstructed bladder isconstant

(4),

but thisenlargedtarget organcould

provide

anexcess ofa

trophic

moleculeto each

ganglion

cell. Thus, the obstructed bladdermaydeliverupto 50timesmoreNGFtoeach

ganglion

cell thananormal

blad-derbaseduponNGFcontentin thetotalorganand

ignoring

possible

differences in access andtransport (20). Neuron

hy-pertrophy

could occurinresponsetothis

rapid

increasein the

peripheral

supply of NGF.

The temporal correlation between bladder

weight,

NGF

content,neuronal

hypertrophy,

andchangesinbladder

func-tion areconsistent withalink betweenthistrophic molecule and

adaptive

changes in neural

pathways

with

obstruction.

However, the strongestevidence that NGF is atleastin part

responsible for these

alterations

wasthe

finding

that

endoge-nous

antibody

toNGF reduced the

proportion

of cells in the

uppermost size

categories

forneurons

supplying

the bladder after obstruction.

Indissociated cell culture,

pelvic ganglion

cells from therat immunoreactive for

tyrosine

hydroxylase

(TH-IR)

are

elimi-natedby NGF antibody (2

1).

Neurons in the

pelvic ganglion

innervating

thebladder withareas >500

,m2

areTH-IR

(22).

Furthermore,

administration

of NGF

antibody

to neonatal mice eliminates

catecholamine-fluorescent

varicositiesin

pel-vic ganglia (23). Theseobservationsimply that NGF antibody

in vivoprobablyaffects adrenergicneuronsinthe major pelvic

ganglion.Itisunlikelythat theeffects of obstructionare merely

limitedtoadrenergicneurons because ashiftin area occurs for

the entire population ofbladder neurons (4) and TH-immuno-reactiveneurons account for - 25% of the cells in the pelvic

ganglion (22).

Muscle activity in additionto targetsizehasbeen shown to alter neuron survival and morphology. Thus, an increase in

contractile activity ofthebladder due toobstructioncould ele-vate NGF levelsindependentof smooth musclehypertrophy. This hypothesiswas tested by examining neuronal size and NGF content in bladdersofdiuresedanimals.Although

void-ing frequency in diuresed rats rose to the same extent as in

urethral-ligated animals (Fig. 5), bladder weight slightly

in-creased and neuronal areasremained 65%lessthanobstructed rats at 6 wk. NGF content of bladders fromdiuresedanimals was only2.5 timesgreater than controls compared to the

11-fold increase (per organ)seenwith obstruction. Furthermore, the increase in voiding frequency in the autoimmune group was notassociatedwith the same degree of neuronal enlarge-ment asinobstructed animals (Fig. 3). These findings indicate

that increasedbladderactivity isnottheprimary stimulus for neuralplasticityandalterations in NGF.Rather,increased

ac-tivity reflects otherchanges in thefunction ofthe bladder and

itsinnervation thatresultfrom obstruction. Noneofthe

ani-malsin thisstudy wereincontinentand theincreased voiding frequency in diuresed animalswas accountedforbyincreased water intake.

Denervation ofsmooth muscle orinflammationmay alter

NGFcontent(24).NGF contentofsmooth musclesintheiris

and vas deferens varies with qualitative differences inneural

input.NGFcontent in this situation representsthe netbalance

2I

aD

a

8

z1

850

425

2

U->1

0 LL

c

C.

S2

0

CONT OBS DELIG DIUR Al AIOBS

Freq area W NGF/bId

Figure 5. Nerve growth factor content of bladder after experimental

manipulations.Urinary frequency(solidbars;h-'),bladder neuron size(grey bars;

um2)

and NGF content(hatched

bars;

pg/organ)for

eightcontrol(CONT) animals (750 cells); six animals obstructed

(OBS)for 6 wk(750cells); sixratsobstructedfor 6 wk, thendeligated

(DELIG)for 6 wk(630cells); eight diuresed (DIUR) rats (710 cells),

eightautoimmuneanimals(Al)(611 cells), and six rats made au-toimmune then obstructed for 6 wk (AIOBS) (715 cells). NGF content and neuronalsize increased in all obstructed groups (P<0.001)

comparedtocontrol.Autoimmunityalone causedasignificant(P

betweensynthesis bytargettissuesanduptake by axons. Acute obstruction elicits a relative denervation ofthe bladder because ultrastructural studies show axonal degeneration with subse-quentreinnervation ofsmooth muscle (20). Alternatively, in-flammatory mediators including IL-l can stimulate produc-tion of NGF. However, preliminary data in our laboratory

show that surgical denervation ofthebladder or intravesical

instillation ofchemicalirritants do not raise NGF levels to the degree observed with obstruction (25). Taken together these

studiessuggestthatincreasedNGFproduction is closely linked to the mechanisms underlying smooth muscle hypertrophy.

This hypothesis is strengthened by the finding that cultured smoothmusclecellscontainmRNAcoding for NGF and

syn-thesize NGF (26).

Questions remainas tothe molecularbasis forgrowth fac-torregulation. Recentcircumstantialevidence has linked ele-vationsin intracellular calcium and subsequent expression of protooncogenes withcardiachypertrophy(27). Enhanced

ex-pressionofprotooncogenes such as

c-fos,

c-jun,

and c-mychave

also beenassociated withincreases inNGFproduction byglia and fibroblasts (28). These secondary alterations in gene

ex-pressionmaycontributeto an increased capacity of the cell for

synthesis of contractile proteinsand to theactivation of genes

responsible forthemanufacture ofa widerange ofgrowth

fac-torssuch asNGF.

Weattempted to link changesinNGFwith alterations in targetorganfunction. However,the useofvoiding frequencyto assessbladderactivity iscomplicated by the complexnatureof factors governing micturition. Theincreased residual volume that occursafterobstruction canbeassociated withoverflow incontinence,the presenceof which would reducethe

useful-ness of measuring voiding frequency. However, none ofthe

animalsinthis studywereincontinent. Episodic voiding with

periods

of continencewasobserved. TheAlratsvoidedmore frequentlythan control rats, and AIdidnot prevent therise in voiding frequency afterobstruction. Endogenous NGF anti-body could have increased urine output, enhanced smooth muscleactivityorexcitatory parasympathetic pathways

regu-lating voiding,or reduced

sympathetic

inhibition of micturi-tionpathways. These factorsmay

explain why

AIdidnot pre-vent theincrease in

voiding

frequency

seenafter obstruction.It

isalso possible that NGF is notresponsible for the

develop-mentof increased bladder

activity

after

partial

urethral

liga-tion.

Theinability of NGF

antibody

tocompletelyprevent

hy-pertrophy of

postganglionic

neuronsinduced

by

bladderoutlet

obstruction could be duetothe

dependence

of NGF-insensi-tiveneuronson other

trophic

factorsfor

growth.

Cholinergic

neuronsinthepelvic

ganglion provide

excitatory input

tothe

smoothmuscleofthe bladder

(29).

Therefore, trophic

factors whichinfluence the growthandsurvival of

peripheral

choliner-giccells suchas

ciliary

neurotrophic

factor

(30),

brain-derived

neurotrophic

factor (31), and basic fibroblast

growth

factor (32)maybe

responsible

for the

hypertrophy

of NGF-insensi-tiveneurons. Itisalso

possible

that the

antibody

identified

by

ourdorsalroot

ganglion

bioassay

technique

wasineffectiveor

its access to bladder neurons limited in vivo. However, the reductionin meansizeofMPGneuronsinAIanimalsargues

thatthe

antibody

waseffectiveanditsaccesssufficient.These datadonotestablish NGFasthe solemediator ofthe

postob-structiveneural

plasticity.

NGF may be

operating

aloneorit

may bepermissiveforsecondary events. Theabilityof

autoim-munity to NGF to prevent some of the neural growth does not distinquish between a primary and permissive role for the neurotrophin. However, this more complex relationship lacks precedentand the moreparsimonious conclusion is a direct effect of NGF.

Another goal ofthisinvestigationwas toevaluate the revers-ibility of neurotrophic and functional alterations elicited by obstruction. Deligation of obstructed animals was associated with areturn ofvoidingfrequencyto controllevelsand a fallin

NGF.However, NGF andneuronal areas remained above con-trol values. Because bladder weights remained elevated it is possiblethatobstructionwasstillpresent. Itisalsopossible that the changes induced by obstruction would have reversed com-pletely ifexperiments were conducted beyond 6

wk,

becauseit cannot be assumed that the rateofreversalisthe same as the rateofonset. Another explanation is that neural changes which occurin response to 6 wkof obstruction are not completely reversible. It is tempting to speculate that anincomplete

rever-sal in neuralplasticity mayexplain the continued symptoms andbladderdysfunction inpatients aftersurgical relief of ob-struction.

NGFand trophic factors mayparticipatein other changes associated withbladderhypertrophy. Inadditiontoregulating

thegrowth andsurvivalofpostganglionicneuronsin the

periph-eral nervous system, NGFalsoaffectssensory neurons.

Affer-ent neuronssupplying the bladder demonstrate

significant

plas-ticityin response toobstruction (5-7). Similarto neurons in

thepelvicganglia, afferentneurons inthe L6 and S1 DRG

pro-jectingtothe ratbladderundergohypertrophy. Furthermore,a 60%increase inthe areaofWGA-HRP labeledafferent projec-tions to the intermediolateral cell column from obstructed bladdersisobserved(5-7). One mechanism for these

morpho-logic changesmaybeanincreased

availability

andretrograde

transportofNGFfromthehypertrophied bladder.

Neuralregulation of micturition involvesacomplex series ofevents

involving

centraland

peripheral

autonomic

pathways

(26).Electrophysiological

analysis

of normalratsandcatshas shown that micturition isregulated bya

supraspinal

reflex

re-layed through the pons. A second reflex

pathway,

termed a spinal

reflex,

developsin

spinal

cord-transectedanimals with

the

resumption

of micturition. Itis of interest that obstructed ratsconsistentlydemonstrate

electrophysiological

evidenceof a spinal reflexinadditiontothesupraspinalreflex

(3).

Thus, increased NGF in obstructed bladders may

produce

a func-tional

plasticity

through altered

circuitry

orpatterns of

connec-tivity

withinthe sacral

spinal

cord. These

changes

in afferent

pathwaysanda

spinal

reflexcould inturncontributeincreases invoidingfrequencyin our modelortothe sensory symptoms and bladder

hyperactivity

observedinindividualswith bladder outlet obstruction. A

positive

feedback system may exist whereby the central nervous systemrespondstobladder

hyper-trophy by

reinforcing pathways responsible

for

voiding.

It has been postulated that the functional obstruction that occurs

after spinal cord transection

produces

bladder

hypertrophy

which thereby leadstothe emergence ofa

spinal

micturition reflex (7). This reflex is

required

for the returnof

voiding

in

paraplegics

afteraninitial

period

of detrusor areflexia. Our data linkingbladder

hypertrophy,

neuronal

enlarge-mentand changesin organfunction with NGFcontentoffers an

important

newmechanismto

explain

trophic

interactions betweenatarget anditsnerve

supply

in the adult.

Although

the

from pathology exhibited in patients with lower urinary tract

obstruction, preliminaryobservations in humans raise the pos-sibility that changes in NGF may participate in clinical dis-orders associated with organ enlargement or smooth muscle hypertrophy. If neurotrophic substancesact asmolecular sig-nalsthat initiate the dynamic interactionsbetween anorgan and its innervation, the possibility exists that manipulation of growth factors or their receptors could provideauseful method of therapeutic intervention for certaindiseaseprocesses.

Insummary, the data presented suggestanovel mechanism

makingacontribution to obstruction-induced changesin effer-entpathwaystothe bladder.Anoutletobstruction causes

blad-der smooth muscletorespondto theincreasedfunctional de-mand withhypertrophy and/orhyperplasia. Concurrent with these musclegrowth responses, neurotrophic factors, including NGF, are produced in greater amounts by the bladder muscle. Neuronsinnervatingthe bladder acquire the increased

neuro-trophins and themselves respond with growth. The neural growth is postulated to affect the reflex function, perhaps

con-tributingto loweredreflexthresholds and alteredvisceromotor function.

Acknowledgments

Mouse NGFwas agenerousgift ofDr. E. M. Johnson, Washington University, St. Louis, MO. The authors would like togratefully ac-knowledge the work ofBobraSlaughter in the preparation ofthis

man-uscript.

Work supported by National Institutes of Health grants IKi

1-DK1732-04andR29NS 28566-01 and by grant-in-aid 901321 from theAmericanHeartAssociation.

References

1.Holtgrewe,H.L., W. K. Mebust, J. B. Dowd, A. Cockett, P. C. Peters, and C. Proctor. 1989. Transurethralprostatectomy: practice aspects of the dominant operation in American urology. J. Urol. 141:248-253.

2.Anderssen,J. T. 1976.Detrusorhyperreflexia in benign infravesical ob-struction.Acystometric study. J. Urol. 115:532-536.

3.Steers, W. D., and W. C. de Groat. 1988. Effect of bladder outlet obstruc-tiononmicturitionreflex pathways in the rat.J.Urol. 140:864-871.

4.Steers, W. D.,J.Ciambotti, S. Erdman, and W. C. de Groat. 1990. Morpho-logical plasticity in efferent pathways to the urinary bladder of the rat following urethralobstruction.J. Neurosci. 10:1943-1951.

5.Steers,W. D. 1990.Neuroplasticity secondary to infravesical obstruction.

Neurourol. Urodyn. 9:559-561. (Abstr.)

6.Steers,W.D.,J.Ciambotti,B.Etzel, S. Erdman, and W. C. de Groat. 1988. Neuralplasticityinafferent pathways to the urinary bladder following urethral obstruction in therat.J. Comp. Neurol.Inpress.

7. deGroat,W.D.,M.Kawatani, T. Hisamitsu, C.-L. Chenc, C.-P. Ma, K Thor,W.Steers, and J. R. Roppolo. 1990. Mechanisms underlying the recovery ofurinarybladderfunction following spinal cord injury. J. Auton. Nerv. Syst. 30:571-578.

8. Purves, D.,W.D.Snider, and J. T. Voyvodic. 1988. Trophic regulation of

nerve cellmorphologyandinnervation in theautonomicnervoussystem.Nature (Lond.). 336:123-128.

9.Voyvodic, J. F. 1989. Peripheraltargetregulation of dendritic geometry in theratsuperior cervical ganglion.J. Neurosci. 9:1997-2010.

10. Johnson, E. M., andI. A.Ferguson.1989. Nervegrowthfactor. In Nerve

Growth Factors. R.A.Rush,editor.JohnWiley&Sons,New York.213-240.

1 1.Uvelius,B., andA.Mattiasson. 1984.Collagencontentin the raturinary

bladdersubjectedtoinfravesical obstruction. J. Urol. 132:587-590.

12.Gorin,P.D., and E. M.Johnson. 1980. Effects oflong-termnervegrowth factordeprivationonthenervoussystemof the adult rat: anexperimental

au-toimmuneapproach.BrainRes. 198:27-42.

13. Lowry, 0. N.,N. J.Rosebrough,A.L.Farr,and R. J. Randall. 1951. Protein measurement withthefolinphenolreagent.J.Biol.Chem. 193:265-273. 14.Meghji,P., J.B.Tuttle, andR. Rubio. 1989. Adenosine formation and releaseby embryonicchick neuronsandgliain cell culture. J. Neurochem. 53:1852-1860.

15. Westkamp, G., and U. Otten. 1987.Anenzyme-linkedimmunoassayfor

nervegrowth factor(NGF):atoolforstudying regulatorymechanisms involved inNGFproductioninbrain and inperipheraltissues.J.Neurochem. 48:1779-1786.

16.Hellweg, R., C. Hoch, andH. D.Hartung. 1989.Animproved rapidand

highlysensitive enzymeimmunoassayfornervegrowthfactor. Tech.J.Cell. Mol. Biol. 1:43-48.

17.Earlam,R.J. 1971.Ganglion cell changes inexperimentalstenosis ofthe gut. Gut. 12:393-398.

18.Gabella, G. 1984. Size ofneuronsamongglialcellsin theintramural ganglia of thehypertrophicintestine of theguinea pig.J.Neurocytol. 13:73-84. 19.Elbadawi,A.S., S. Meyer, B.Malkowitz,A.Wein,R.Levin,and M.A.

Atta.1988.Effect of short-termpartialbladder outlet obstructiononrabbit

detru-sor:anultrastructuralstudy.Neurourol.Urodyn.8:89-116.

20.Creedon,D.J.,W. D.Steers, andJ.B.Tuttle.1989. Potential roleforNGF

in neuronalplasticity after bladder obstruction. Soc. Neurosci. Abstr. 16:307. (Abstr.)

21.Steers,W.D.,and J. B.Tuttle.1990. Dissociated cell culture of adultrat pelvic ganglion.Soc.Neurosci.Abstr. 15:300.(Abstr.)

22.Keast, J. J., and W. C. de Groat. 1989.Immunocytochemical characteriza-tionof pelvicneurons whichprojecttothebladder, colonorpenisinrats:J.

Comp. Neurol.288:287-400.

23.Bjerre,B., and E. Rosengren. 1974. Effects ofnervegrowth factor and its antiserumonaxonalregeneration ofshortadrenergicneurons in the malemouse. CellTissue Res. 150:9299-322.

24.Ebendal, T.,L.Olson, and A. Seiger. 1983. The level of nerve growth factoras afunction of innervation. Exp.CellRes. 148:311-317.

25.Albo, M., L. Nataluk, W. Steers,andJ. Tuttle. 1991.Morphological,

functional andneurotrophicalterations inpartiallydecentralized and denervated bladders.J. Urol. 145:308A.(Abstr.)

26.Creedon, D. J., W. D. Steers, and J. B. Tuttle. 1988. Cultured vascular smooth musclecellsprovideNGF-dependentneuronal survival. Soc. Neurosci.

Abstr. 14:680. (Abstr.)

27.Marban, E., and K. Yukihiro. 1990. Cell calcium, oncogenes, and hyper-trophy. Hypertension(Dallas). 15:652-658.

28. Hengerer, B., D. Lindholm, R. Heumann, U. Ruther, E. F. Wagner, and H. Thoenen. 1990.Lesion-induced increase in nerve growth factor mRNA is mediated by c-fos. Proc.Natl.Acad. Sci. USA 87:3899-3903.

29. deGroat,W.D., and W. D. Steers. 1990. Autonomic regulation of the

urinarybladder andsexorgans. InCentral Regulation of Autonomic Function. R. M.Spyer andA.Loewy,editors. Oxford University Press, Oxford. 310-333. 30. Tuttle,J.T., and G. Pilar. 1986. Target tissue influences on cholinergic

developmentof parasympathetic neurons. Dynamics of cholinergic function.

Adv.Behav. Biol. 30:879-890.

31. Barde, Y.-A., E. Edgar, and H. Thoenen. 1982. Purification of a new neurotrophic factor from mammalian brain. EMBO (Eur. Mol.Biol.Organ.) J.

1:533-549.

32. Vaca, K., S. S. Stewart, and S. H. Apel. 1989. Identification of basic