Kidney International, Vol. 45 (1994), pp. 1028—1 036

Epidermal growth factor antagonizes vasopressinin vivo and in vitro

PADDY A. PHILLIPS, SHARON L. GRANT, A. FI0NA DAVIDSON, JOHN RIsvANIs,JAYNE STEPHENSON, and CHRISTINE B. Gow

Department of Medicine, University of Melbourne, Austin Hospital, Heidelberg and School of Agriculture,La Trobe University, Bundoora, Victoria, Australia

Epidermal growth factor antagonizes vasopressin in vivo and in vitro.Since EGF causes diuresis through a renal action and may antagonizethe hydroosmotic effect of AVP in vitro we investigated the antagonisticaction of EGF with AVP in vivo and the mechanism of the antagonism inVitro. Conscious ewes received i.m. injections of a selective AVP V2-receptor agonist (1-desamino, D-Arg8 vasopressin acetate, DDAVP) every12 hours for days 5 to 16. All ewes received an i.v. isotonic salineinfusion (100 mI/day) for days ito 8 and days 13 to 16, and i.v. EGF in100 ml saline/day at doses of 0 (N = 8) or 10 (N = 8) pg/hr for days 9to 12. DDAVP reduced both urine volume and water intake, andincreased urine osmolality. In contrast, simultaneous infusion of EGFreversed the DDAVP-induced responses, resulting in a transient nega-tive fluid balance, kaliuresis and a transient natriuresis (all P < 0.05).When EGF treatment ceased, the effects of DDAVP treatment alonegradually became apparent. From the in vitro studies, the AVP-relatedpeptides displaced specific AVP VI- and V2-receptor antagonist radio-ligands from rat renal inner medullary membranes, whereas EGF hadno effect. However, EGF antagonized AVP V2-stimulated cAMPproduction in a dose-dependent way (IC50 = 2 x 107M). Therefore, thediuretic effect of EGF is not via direct antagonism of the antidiureticAVP V2-receptor but seems mediated by inhibition of the antidiureticAVP V2-receptor second messenger system.

Although it is known that epidermal growth factor (EGF) ismitogenic for renal tubule cells in vitro [1, 2] and in vivo [3], itis increasingly recognized that EGF may also influence renalfunction causing altered sodium transport and water permeabil-ity in vitro [4—6] and in vivo [7—91. In isolated cortical collectingtubules, EGF antagonized arginine vasopressin (AVP)-stimu-lated water transport [4]. Recently, we reported that chroniclong-term intravenous (i.v.) infusion of EGF into conscioussheep caused diuresis, natriuresis and polydipsia [7—9]. Thedoses of EGF reported to alter fluid and electrolyte balance insheep include 0.5 mg/day [7], 5 or 10 pjg/kg live weight per day[8] and 10 or 20 pjg/hr [9]. These effects seemed primarilycaused by a direct renal action of EGF in that i.v. EGF infusioninto animals with fixed water intake caused diuresis, natriuresisand negative fluid balance with compensatory polydipsia and

Received for publication August 23, 1993and in revised form November 5, 1993Accepted for publication November 8, 1993

© 1994 by the International Society of Nephrology

sodium retention upon cessation of EGF infusion and ad libitumwater access [10]. The mechanism of these effects was unclear.

In the kidney AVP produces antidiuresis by activating theAVP V2 (antidiuretic)-receptor subtype and its second messen-ger system, adenylyl cyclase and cAMP generation. The aim ofthe in vivo experiment reported here was to determine whetherEGF (10 jzg/hr) could antagonize the antidiuretic effects of aselective AVP V2-receptor agonist (l-desamino, D-Arg8 vaso-pressin acetate, DDAVP) in conscious ewes and so causediuresis. The in vitro study was undertaken to investigatefurther the effects of EGF on the binding of selective AVP V1(vascular)- and V2-receptor antagonist ligands to their respec-tive receptors, and on renal AVP V2-receptor-stimulated cAMPproduction in rat renal inner medullary membranes.

Methods

In vivo studies

Experimental animals. Eight 2 to 2.5 year-old crossbred(Border-Leicester x Merino) ewes (39 to 47 kg live weight) wereselected from a single flock. All ewes had been ovariectomizedunder general anesthesia three months before the experiment inorder to exclude any influence of reproductive hormones onfluid and electrolyte balance. General anesthesia was inducedusing 800 mg sodium thiopentane (i.v. injection, Pentothal,Boehnnger and Ingelheim, Melbourne, Australia) and main-tained with a mixture of halothane (Fluothane, Id, Melbourne,Australia) and oxygen. The crossover experiment was con-ducted in four trials, separated by eight weeks to allow theanimals to recover from the previous experimental procedures.Immediately before the first part of the experiment, the eweswere shorn, weighed and randomly allocated to a group as-signed to receive an i.v. infusion of isotonic saline alone(control treatment) or recombinant hEGF (Auspep Pty. Ltd.,Melbourne, Australia) in isotonic saline (EGF treatment). In thesecond part of the crossover experiment, the infusions werereversed such that all ewes received both types of infusion bythe end of the experiment (N = 8 per group). Animals werehoused with continuous light in a temperature-controlled (18 to22°C) room.

Each ewe was kept in an individual metabolism cage thatallowed the separation and collection of urine and feces. Waterwas available ad libitum and each ewe was offered 60:40

1028

Phillips et al: Antagonism between EGF and AVP 1029

lucerne: rolled barley (9.8 MJ metabolizable energy/kg, air dry).The ration (0.6 to 0.8 kg) was offered in two halves at 1000 and2115 hours daily. During the three weeks before each trial, theewes became accustomed to handling. All procedures on sheepwere carried out with the prior approval of the La TrobeUniversity Animal Experimentation Ethics Committee.

Experimental procedures. Two polyvinyl chloride catheters(1.0 mm inner diameter by 1.5 mm outer diameter, DuralPlastics, Sydney, New South Wales, Australia) were placed inthe jugular veins of each ewe four to five days before infusionsbegan, One catheter was used for infusion of saline or EGF, theother for the collection of blood. In each trial, all ewes received4 jig DDAVP (1 ml, Minirin, Ferring AB, Malmo, Sweden)intramuscularly (i.m.) twice daily (at 0920 and 2100 hours) fordays 5 to 16. The ewes received a continuous i.v. infusion of 100ml isotonic saline/day for eight days (days I to 8), followed by0 (control treatment) or 10 jig EGF/hr (EGF treatment) in 100ml isotonic saline for four days (days 9 to 12). Animals werepaired such that one control animal was examined wheneverone EGF-treated animal was studied. Subsequently all ewesreceived 100 ml isotonic saline/day i.v. for four days (days 13 to16). Infusates were kept in an ice-water bath and deliveredthrough a filter (0.22 jim, Millipore Corp., Bedford, Massachu-setts, USA) using a peristaltic pump (Gilson, Minipuls 2, JohnMorris Scientific Pty. Ltd., Balwyn, Victoria, Australia).

Blood samples (30 ml) were collected each day at 0900 hours(before DDAVP administration) and water intake and urinevolume were measured [4] daily at 0930 hours. Feces werecollected and weighed every day and duplicate samples (—'90 g)were dried overnight at 65°C for determination of the output offecal dry matter.

Blood measurements. Free-flowing blood was obtained formeasurement of microhematocrit (PCV), plasma sodium andpotassium concentrations (Instrumentation Laboratories, 501ion specific electrode), osmolality (Wescor 5lOOC vapor pres-sure osmometer) and hormone concentrations. Blood forplasma AVP concentrations [11] was taken into chilled lithiumheparin tubes, and for plasma atrial natnuretic peptide (ANP)concentrations [121 was taken into chilled sequesterene tubescontaining 500 kallikrein inhibitor units of aprotinin (Trasylol,Bayer). All blood samples were placed on ice, centrifuged at4°C and plasma stored at 4°C until assay for osmolality, sodiumor potassium or at —20°C for all other assays. Both radioimmu-noassays had intra- and interassay coefficients of variation ofless than 8%. In a separate set of in vitro experiments it wasdemonstrated that there was zero cross reactivity of the plasmaAVP assay to DDAVP (Minirin).

Urine measurements. Urine samples were centrifuged andaliquots stored at 4°C until assayed for sodium and potassiumconcentrations and for osmolality.

Statistical analysis. Since the effects of DDAVP were well-known and the data of interest were those comparing EGFtreatment versus control on days 9 to 12 and the subsequentperiod after this (days 13 to 16), data for these days wereanalyzed by paired t-tests. Statistical significance was taken atthe P < 0.05 level. Data are presented as means SEM.

In vitro studies

Experimental animals. Female Sprague-Dawley rats (approx-imately 250 g live weight) were obtained from the Biological

Research Laboratories, Austin Hospital. After sacrifice bydecapitation the kidneys were removed and renal inner medul-lary membranes prepared according to the method of Trinder etal [13]. All procedures on rats were carried out with the priorapproval of the Austin Hospital Animal Ethics Committee.

V2-receptor displacement assay. Renal inner medullarymembranes (250 jig) were incubated with binding buffer, theAVP V2-receptor selective antagonist radioligand ([3H]-desGlyNH29d(CH 2)5[D-Ile2Ile 4]AVP, 0.2 x 109M) as well asAVP, DDAVP, a selective AVP V1-receptor antagonist (d(CH2)5[Sar7]AVP, dSAVP), a selective AVP V2-receptor antagonist(d(CH2)5[D-Ile2Ile4]AVP), oxytocin, angiotensin II (AuspepPty. Ltd.) or EGF at varying concentrations (l0— to l0' M).The peptides were obtained from Peninsula Laboratories(Parkville, Victoria, Australia), unless indicated otherwise. Thebinding buffer consisted of 0.1 M Tris (pH 7.4), 0.01 M MgC12,0.5 mg/mi bacitracin (Sigma, St. Louis, Missouri, USA), 100lU/mi aprotinin (Sigma) and 0.1% bovine serum albumin (BSA,CSL Ltd., Parkville, Victoria, Australia). The radioligand (Du-pont, North Sydney, Australia) had a specific activity of 50.7CiJmmol. Following a two hour incubation the samples werefiltered through Whatman GF/B glass fiber filters in a Brandelfiltration apparatus. The filters were rinsed three times withwash buffer (0.1 M Tris pH 7.4, 0.01 M MgCI2 and 0.1% BSA).Measurement of beta radiation was determined using a PackardTricarb 4530 for two minutes/sample with 60% efficiency.Specific binding was measured as total binding minus nonspe-cific binding determined by an excess of unlabeled AVP (10_6M) and is shown as the percentage of bound/unbound in thepresence of AVP, related peptides, angiotensin II or EGF.

V1-receptor displacement assays. The AVP V1-receptor an-tagonist radioligand [1251]-dSAVP was prepared by the chiora-mine T method of iodination and HPLC purification [14] andincubated (1 x,409 M) with renal inner medullary membranes(250 jig), binding buffer (as above), and the same peptides asabove (10—s to 10- " M) for one hour at room temperature. Thebound and unbound ligands were separated as above andgamma radiation measured (LKB Wallace 1260 Multigammacounter) for one minute at an efficiency of 80%. Calculationswere as for the AVP V2-receptor binding assays.

cAMP determination. Adenylyl cyclase activity was mea-sured using the method of Trinder et al [15]. Renal innermedullary membranes were incubated in a medium containingguanosine triphosphate (1 x l0- M, Sigma), iso-butyl-methyl-xanthine (2 x i0 M, Sigma) in the presence or absence ofAVP. Stimulation of the reaction was initiated through theaddition of ATP (1 X 10 M, Sigma) and terminated after 10minutes (37°C) through the addition of 0.15 M acetic acid. ThecAMP levels produced were measured by radioimmunoassay ofthe supernatant (pH 7.4) after centrifugation using a commercialkit (Amersham cyclic AMP [3H] Assay system TRK 432,Buckinghamshire, UK). Sodium fluoride (NaF, 0.01 M), apotent stimulator of G-proteins, was used as a positive control.The levels of cAMP were measured as pmollmg protein per 10minutes and expressed as a percent increase in cAMP frombasal levels. For an assay to be classified as successful, an apriori requirement was for at least a 40% increase on basalunstimulated cAMP production. To determine any inhibitoryeffect of EGF on cAMP production, EGF (10k to 10' M) wasadded to tubes containing AVP or NaF. Results are presented

1030 Phillips et a!: Antagonism between EGF and AVP

as the reduction in cAMP production after stimulation withAVP or NaF.

Results

In vivo studies

At the end of the experiment, the mean live weight of theewes with the control treatment (N = 8) and EGF treatment(N = 8) was 42.4 1.4 kg and 42.4 1.3 kg, respectively.

Appetite response. All ewes ate most of the ration offeredthroughout the experiment and the mean daily feed intake onday 4 was 0.66 0.03 kg for the control ewes and 0.68 0.04kg for ewes given the EGF treatment. There was no effect ofDDAVP injection nor EGF infusion on daily feed intake nor onfecal dry matter output. The mean output of fecal dry matter onday 4 was 0.16 0.02 kg and 0.16 0.02 kg for the ewes giventhe control and EGF treatments, respectively.

Fluid balance. The mean daily urine volume (Fig. lA) wasrelatively constant during days one to four (saline infusiononly), being 0.61 0.14 to 0.78 0.21 liters and 0.62 0.22 to0.76 0.28 liters for the ewes given the control and EGFtreatments, respectively. After the twice-daily injections ofDDAVP commenced, mean daily urine volume during days fiveto eight was reduced by 48% to 0.35 0.02 liter for the controlewes and 0.36 0.01 liter for the EGF-treated ewes. Thecontrol ewes which received DDAVP from day 5 onwardsmaintained this pattern of reduced urine volume for the rest ofthe experiment. By contrast, the simultaneous infusion of EGFduring DDAVP treatment induced an immediate and sustainedincrease in urine volume (days 9 to 12, P < 0.05) of approxi-mately 128% (to 0.82 0.05 liter) over DDAVP-inducedchanges, approaching values recorded during the pretreatmentperiod (days 1 to 4). The urine volume of the EGF-treated ewesfell slowly after the EGF infusions ceased and remained higherthan that of the control ewes on day 13 (P < 0.05) and day 16(P < 0.01).

The daily water intake (Fig. 1 B) was also constant during thepretreatment period (days 1 to 4), with mean values of 1.750.15 to 1.91 0.26 liters and 1.81 0.26 to 2.01 0.30 litersbeing recorded for ewes in the control and EGF treatments,respectively. As for urine volume, treatment with DDAVPreduced water intake such that the mean daily water intakeduring days five to eight was reduced by 25% (to 1.39 0.03liter) for the control ewes and by 29% (to 1.37 0.04 liter) forthe EGF-treated ewes. The ewes given the control treatmentmaintained this pattern of reduced water intake for the rest ofthe experiment. By contrast, during the period of simultaneousEGF infusion on days 9 to 12, water consumption increasedslowly towards pretreatment values. Water intake of the ewesgiven EGF treatment was greater than that of the control eweson days 10 (P < 0.01) and 11 (P < 0.05) and the mean value ofthe EGF-treated ewes for days 9 to 12 had increased to 1.790.10 liter (+ 30%). The daily water intake of the EGF-treatedewes fell immediately after the EGF infusions ceased, reachinglevels recorded during the first period of DDAVP injections byday 13.

Daily fluid balance (intake minus output in urine and feces)was 0.94 0.12 to 1.12 0.15 liter and 0.98 0.16 to 1.220.12 liter for ewes given the control and EGF treatments,respectively, throughout days one to four of the experiment(Fig. 1C). There was no effect of DDAVP treatment on meanfluid balance because of the spontaneous reduction in waterintake in parallel with DDA VP-induced antidiuresis. On the first

A

, ***

DDAVP 4tg ni. 12 hourly

160

0.0C

U)

CU

C

U)a

1.4

1.2

1.0

0.8

0.6

0.4

0.2

0.0

2.6

2.2

1.8

1.4

1.0

0.6

B

**

DDAVP 4tgi.rn. 12 hourly

0 2 4 6 8 10 12 14 16

C

1.4

1.0

U)

U)0Caa.00U-

DDAVP 4pgi.m. 12 hourly

0.6

0.20 2 4 6 8 10 12 14 16

Time, daysFig. 1. Daily (A) urine volume, (B) water intake and (C) fluid balance ofcross bred ewes before (days 1—4) and during (days 5—16) treatment(i.m. injections every 12 hr) with 4 ,sg DDAVP. Ewes also received acontinuous infusion (100 mLfday) of isotonic saline for 8 days (days 1—8),then either 0 (O—O, N = 8) or 10 (D—D, N = 8) sg EGF/hour insaline (100 mI/day) i.v. for 4 days (days 9—12), followed by saline (100mi/day) for 4 days (days 13—16). Closed symbols refer to the period ofconcomitant EGF infusion during DDAVP treatment. Plotted pointsrepresent mean values and selected standard errors are shown asvertical bars. *P < 0.05; a 0.01; *** a 0.001 for data during andafter EGF infusion versus control data on the same experimental day.

Phillips et a!: Antagonism between EGF and A VP

Q)

0EEazccE(I)a0

1031

Fig. 2. Daily (A) PCV, (B) plasma osmolality and plasma concentrations of(C) sodium and (D) potassium of crossbred ewes before (days 1—4)and during (days 5—16) treatment (i.m. injections every 12 hr) with 4 pg DDA VP. Ewes also received a continuous infusion (100 mllday) of isotonicsaline for 8 days (days 1—8), then either 0 (O—O, N = 8) or 10 (D—D, N 8) pg EGF/hr in saline (100 mI/day) i.v. for 4 days (days 9—12),followed by saline (100 mI/day) for 4 days (days 13—16). Closed symbols refer to the period of concomitant EGF infusion during DDAVP treatment.Plotted points represent mean values and selected standard errors are shown as vertical bars. tP < 0.05; 0.01; *** 0.001 for data duringand alter EGF infusion versus control data on the same experimental day.

day of concurrent EGF treatment, the EGF-treated ewes wentinto negative fluid balance by 0.20 liter (P < 0.05) comparedwith the mean recorded during DDAYP treatment alone. Thisindicates that the simultaneous infusion of EGF during DDAVPtreatment had an immediate diuretic effect. Thereafter, therewas no difference in fluid balance between the two treatments,indicating that increases in water intake paralleled those ofurine volume for the rest of the experiment.

Hemazocrit. The PCV declined throughout the experiment,irrespective of DDAVP or EGF dose (Fig. 2A), probably due toblood sampling. There was a tendency for PCV to be greaterduring EGF treatment (day 11, P = 0.07), perhaps due toplasma volume contraction secondary to the diuresis and natri-uresis (see below).

Plasma electrolyte and hormone responses. There was noeffect of DDAVP nor EGF treatment on plasma osmolalityagain reflecting overall adjustment of water intake to changes inurine volume (Fig. 2B). Plasma sodium concentration (Fig. 2C)was transiently higher on the first day of EGF treatment (day 9,

P < 0.05) consistent with the effect on fluid balance (see above)but did not differ at other times reflecting overall adjustment offluid balance. In addition, the plasma potassium concentrationwas not affected by either treatment (Fig. 2D), but did increaseprogressively throughout the experiment with both treatments.

There was no effect of DDAVP nor EGF on plasma AVP(Fig. 3A) or ANP (Fig. 38), reflecting maintenance of fluidbalance.

Renal responses. The urinary excretion of sodium (Fig. 4A)and potassium (Fig. 4B) was not affected by DDAVP treatment.The profile of both parameters remained constant in ewes giventhe control treatment throughout the experiment. However,there was a significant effect of EGF infusion during DDAVPtreatment, such that urinary sodium excretion of the EGF-treated ewes was transiently higher (day 9, P < 0.05; day 10,P < 0.01) than that of the ewes given the control treatment.During the latter part of the simultaneous treatment of EGF andDDAVP (days 11 and 12), there was no difference in urinarysodium excretion between the two treatments (Fig. 3A). Upon

A C

38

34

30

26

22

300

I — I — I I —

DDAVP 4 ig i.m. 12 hourly

>0a-

E0EU)0aE0)aa-

0 2 4 6 8 10 12 14 16

B

145

143

141

139

137

135

48

4.6

4.4

4.2

4.0

3.8

3.6

DDAVP 4L 12 hourt

D

•

DDAVP 4igi.rn. 12 houiiy

295

290

285

280

275 ____________________________DDAVP 4 j.tg i.m. 12 hourly

270 • • I • • I • I • I • I

0 2 4 6 8 10 12 14 16

I.-

0EE

aE(1)cc0

0 2 4 6 8 10 12 14

Time, days

16

1032 Phillips et al: Antagonism between EGF and AVP

A A5

I

7

6

5

4

3

20

I . I

* ***

DDAVP 4 tgi.m. 12 hourly

S..

0EEC0a)C-)

a)a)zC

S..

0EEC0a)0

>,a)C

16

12

4**

*

3.2.

DDAVP 4 tg i.m. 12 hourly0 • I • I • • I • • I

0 2 4 6 8 10 12 14 16

B14

12 *

10

8

6

42

DDAVP 4 tg rn. 12 hourly0 • • I I • I • I I • I • I

0 2 4 6 8 10 12 14 16

C

1750

• •

DDAVP 4tg i.rn. 12 hourly I

0 2 4 6 8 10 12 14 16

B

8

DDAVP 4 p.g i.m. 12 hourly4 • I • I • • I I • I • I •

0 2 4 6 8 10 12 14 16

Time, daysFig. 3. Plasma concentrations of (A) AVP and (B) ANP of crossbredewes before (days 1—4) and during (days 5—16) treatment (i.m. injectionsevery 12 hr) with 4 pg DDAVP. Ewes also received a continuousinfusion (100 mI/day) of isotonic saline for 8 days (days 1—8), then either0 (O—Q, N = 8) or 10 (D—D, N = 8) sg EGF/hour in saline (100mllday) i.v. for 4 days (days 9—12), followed by saline (100 mI/day) for4 days (days 13—16). Closed symbols refer to the period of concomitantEGF infusion during DDAVP treatment. Plotted points represent meanvalues and selected standard errors are shown as vertical bars. J1 <0.05; 0.01; *** 0.001 for data during and after EUF infusionversus control data on the same experimental day.

cessation of EGF infusion, the ewes previously infused withEGF excreted less sodium alter a lag of one day (day 14, P <0.02; day 15, P = 0.001) than did the ewes given the controltreatment. By day 16, there was no difference in urinary sodiumexcretion between the two treatments. Similarly, during theperiod of EGF treatment urinary excretion of potassium washigher (day 9, P < 0.005; days 10 to 12, P <0.05) than that ofthe control ewes. After cessation of EGF treatment, there wasno difference in urinary excretion of potassium between the twotreatments.

There was also an effect of DDAVP to increase urine osmo-lality above values recorded during the pretreatment period(Fig. 4C). Urine osmolality of the EGF-treated ewes throughoutEGF infusion was lower (P < 0.05) than that of the ewes given

3250

2750

E 22500ECo0a)C

1250

0 2 4 6 8 10 12 14 16

Time, daysFig. 4. Urinary excretion of (A) sodium, (B) potassium and (C) urineosmolality of crossbred ewes before (days 1—4) and during (days 5—16)treatment (i.m. injections every 12 hr) with 4 g DDAVP. Ewes alsoreceived a continuous infusion (100 mI/day) of isotonic saline for 8 days(days 1—8), then either 0 (O—O, N = 8) or 10 (D—D, N = 8) zgEGF/hour in saline (100 mI/day) i,v. for 4 days (days 9—12), followed bysaline (100 mI/day) for 4 days (days 13—16). Closed symbols refer to theperiod of concomitant EGF infusion during DDAVP treatment. Plottedpoints represent mean values and selected standard errors are shown asvertical bars. *P < 0.05; ** 0.01; *** 0.001 for data during andafter EGF infusion versus control data on the same experimental day.

i0 10-10 10 108 10 106 10

120

100

80

60

40

20

01 0-

CONC, M

Fig. 5. Displacement of the spec/ic AVP Vreceptor antagonist radi-oligand (13H1-desGlyNH29d(CH2)5[D-Ile2Ile IA VP) by the unlabeleledpeptidesAVP(D, 10 to 109M, N =3), DDA VP (I, 10_6 to 10"M,N = 2),dSA VP (A, io- to 10 M, N =2), d(CH2)5[D-11e211e4]AVP (U,1O_6 to 10" M, N = 2), oxytocin (, 10° to 10" M, N = 2),angiotensin II (•, 10 to 10° M, N = 2) or EGF (0, i0 to i0 M,N = 3) using renal inner medullary membranes from female Sprague-Dawley rats.

the control treatment, reflecting the concomitant diuresis in theEGF-treated ewes at this time. After cessation of EGF treat-ment, urine osmolality of the ewes previously treated with EGFincreased slowly towards that of the control ewes and valuesremained lower only on day 13 (P < 0.05).

In vitro studies

V1- and V2-receptor displacement assays. AVP and its re-lated peptides all displaced the specific binding of both [3H]-desGlyNH29d(CH2)5[D-11e211e4]AVP and ['251]-dSAVP to therenal inner medullary membranes in a dose-dependent mannerexpected based on the prior classification of the analogues asAVP V1- or V2-receptor selective (Figs. 5 and 6). However,neither angiotensin II nor EGF displaced either radioligand,demonstrating a lack of binding of EGF and angiotensin II toeither the renal AVP VI- or V2-receptor subtypes.

Adenylyl cyclase stimulation. EGF inhibited AVP-stimulatedcAMP production with IC50 2 x iO M but only inhibitedNaF-stimulated cAMP production at higher concentrations(Fig. 7).

Discussion

The physiological role of EGF in the kidney is unclear. Thekidneys can synthesize prepro-EGF [161 and EGF [17] andurine contains high EGF concentrations, some of which issequestered from plasma [8, 18—20]. Recently, receptors forEGF have been reported in renal tubules from the dog [21],rabbit [4, 5, 22], rat [23, 24] and sheep [20]. It is known thatEGF is mitogenic for renal tubular cells in vitro [1, 2] and in vivo[3]. In addition, a direct renal tubular effect of EGF has been

Fig. 6. Displacement of the specific AVP V1-receptor antagonist radi-oligand (['251J-dSAVP) by the unlabeleled peptides AVP (U, 10 to10"M, N = 2), DDA VP (•, 1O to 10"M, N = 2), dSA VP (A, 10to 10° M, N = 2), d(CH2)5fD-IIe2I1e4JAVP (U, iO to 10_Jo M, N =2),oxytocin (, I0 to 10"M, N = 2), angiolensin II (•, 1O to 109M,N = 2) or EGF (0, iO to 1O M, N =3) using renal inner medullarymembranes from female Sprague-Dawley rats.

suggested since microperfusion of isolated rabbit cortical col-lecting tubules with EGF reduced the transport of sodium [5, 6]and water [4].

Previous experiments have shown that i.v. infusions of lowdoses of EGF into sheep resulted in dramatic increases in urinevolume and water intake [7—9] and that these EGF-inducedresponses were dose-related. Natriuresis was also reported inthe studies, but was not dose-related, More recently, we [10]demonstrated that the infusion of 10 g EGF/hr i.v. into water-restricted sheep had direct renal diuretic effects. Since EGFcasued diuresis when water intake was restricted [10] theEGF-induced polyuria and polydipsia demonstrated previously[7—9] must have been due to a primary renal effect rather than aprimary effect to increase thirst and water intake with second-ary polyuna. However, the molecular mechanism by whichEGF produced these renal effects was unclear. Possible mech-anisms include antagonism of the AVP V2-receptor directly, itssecond messenger system or its hydroosmotic effect.

The experiment reported here examined the first and secondof these proposals. The in vivo study showed that treatment ofsheep with the specific antidiuretic AVP V2-receptor antagonistDDAVP halved urine volume, and concomitantly increasedurine osmolality and reduced water intake. Simultaneous treat-ment of sheep with EGF dramatically reduced the DDAVP-induced antidiuresis, with a return of water intake and urineosmolality to pretreatment values. There was also an EGF-induced transient natriuresis and kaliuresis. The diuretic effectof EGF treatment was more immediate than its polydipsiceffect, and as a result there was an increase in water loss on thefirst day of EGF infusion compared to that recorded during

0

m

Phillips et a!: Antagonism between EGF and AVP 1033

0

CONC, M

1034 Phillips et a!: Antagonism between EGF and AVP

the 24 hour period of infusion [25, 26]. The differences in resultsreported by us [9] and by these workers [25, 261 may indicate adose-response effect of EGF. However, if vasodilation and afall in blood pressure had occurred with EGF and/or DDAVP,urine output would be expected to have fallen rather thanincrease as seen here.

The mechanism(s) of the antagonistic action of EGF toDDAVP was explored in the in vitro studies which demon-strated no interaction between EGF and renal AVP VI- orV2-receptors, but did show a dose-related inhibition of AVP-stimulated cAMP production by EGF. This may explain theprevious finding of EGF inhibition of AVP-stimulated watertransport in rabbit cortical collecting tubules [4]; however, itwould not explain the inhibition of 8-chlorophenythio-cAMP-stimulated water transport from the same study [4]. Thissuggests that EGF may antagonize the antidiuretic action ofAVP at more than one site.

The effect of EGF on renal cAMP production has beencontroversial. Teitelbaum [27] demonstrated that EGF inhibitedAVP-stimulated cAMP generation in epithelial cells from ratinner medullary collecting tubules. However, others have sug-gested a stimulatory effect of EGF on renal cAMP production

1 0 10_B 1 O- 1 10- 1 [22]. Our in vitro results confirm and extend those of Teitel-baum [27] in that we too demonstrated that EGF antagonizedAVP V2-receptor-stimulated cAMP production. Furthermore,we reported no direct interaction of EGF with AVP-receptors.Possible mechanisms for the effect of EGF on cAMP produc-tion include EGF-stimulated phospholipase C activity or pros-taglandin (PG) production. Phospholipase C antagonism, butnot phospholipase A antagonism or PG synthesis inhibition,blocked the antagonistic effect of EGF against AVP-stimulatedcAMP production in vitro [27], suggesting a prominent role forphospholipase C in this action of EGF.

However, it is known that EGF can stimulate PG synthesisand the release of 3H-prostaglandins and 3H-arachidonic acidfrom cultured canine kidney cells [28]. In isolated rat innermedullary collecting duct cells [23] and rat renal glomerularmesangial cells [29], EGF stimulated arachidonic acid releaseand PGE2 synthesis, with resultant cAMP accumulation [23].This EGF-stimulated PG synthesis was demonstrated to besynergistic with the activation of protein kinase C [29]. The roleof PG on renal function includes direct renal effects andantagonism of AVP at different cellular levels. In vivo treatmentof dogs with PG synthesis inhibitors enhanced the antidiureticeffects of AVP, as indicated by increased urine osmolality anddecreased free water clearance, but there was no effect onGFR, renal blood flow nor electrolyte excretion [30].

Thus EGF-induced synthesis of prostaglandins may be in-volved in the effect of EGF to oppose the antidiuretic effects ofAVP. However, other studies have suggested that the pros-taglandins may not be involved in the renal responses to EGF.Prostaglandin synthesis inhibition did not influence the antago-nistic action of EGF on AVP-stimulated water permeability [4],EGF-induced thymidine incorporation [23] or EGF-inducedinhibition of adenylyl cyclase [27]. Thus controversy remainsconcerning the interaction of EGF and the prostaglandins invivo.

The mechanism of the natriuretic effect of EGF is unclear.AVP itself has been shown to be natriuretic. This may be viatubular receptors in the medullary limb of the loop of Henle ormerely via plasma volume expansion [31]. EGF has been shown

0)

-oC0G)

ciC)C

120

100

80

60

40

20

0

CONC EGF, M

Fig. 7. Inhibition ofAVP- (0, N =3) andNaF- (•, N = 2) stimulatedcAMP production by EGF (1O— to io M) in renal inner medullarymembranes of female Sprague-Dawley rats.

b_fl

DDAVP treatment alone. However, the EGF-treated ewesremained in stable fluid balance for the rest of the experiment.Upon cessation of EGF infusion, the ewes responded withlowered urine volume and water intake, as to be expected fromthe concomitant DDAVP treatment. In addition, there wasreduced urinary excretion of sodium two to three days after thecessation of EGF treatment, presumably to compensate for theprevious EGF-induced natriuresis. Despite these changes,there was no effect of EGF treatment on plasma osmolality norplasma potassium, AVP or ANP concentrations, other than atransient increase in plasma sodium concentration on the firstday of EGF treatment. This suggests that other sodium-retain-ing hormonal systems (such as the renin-angiotensin-aldoste-rone system) were operating to maintain the electrolyte profile.

Changes in glomerular filtration rate (GFR) cannot explainthe changes seen here, since we observed no difference in GFRwith EGF infused at similar or slightly higher doses in previousexperiments [9]. There was also no effect of chronic long-termi.v. infusion of EGF (5 to 20 p.g/hr over 4 days) on mean arterialblood pressure [9]. Previously, Scoggins et al [25] reportedhigher cardiac output at one hour, higher heart rate, lowerstroke volume and total peripheral resistance at three hours,followed by reduced mean arterial blood pressure at five hoursafter the start of an i.v. infusion of a higher dose of EGF (125pg/hr) into sheep. Carter et al [26] had similar observations aswell as vasodilation in the skin, thyroid and submaxillary glandsduring the i.v. infusion of 129 g EGF/kg live weight over 24hours into sheep. Both groups reported that heart rate remainedhigh and blood pressure had returned to normal by the end of

Phillips et al: Antagonism between EGF and AVP 1035

to influence Na/H exchange in fibroblasts and cause depo-larization of cortical collecting tubules [4]. However, this mightnot produce natriuresis. In contrast, prostaglandins have beenshown to stimulate sodium and potassium excretion [32] inaddition to their effect to antagonize AVP. Therefore EGF-stimulated PG production might also account for the natriuresisand kaliuresis seen here possibly via a tubular action.

Whether or not these effects of EGF are physiological orpharmacological remains to be clarified. Normally EGF is notdetectable in sheep plasma. Infusion of EGF at rates similar tothose used in this experiment produce plasma concentrations of0.9 to 20.7 ,ag EGF/liter [7, 8]. Since EGF is produced locally inthe kidney [16, 17], local intrarenal concentrations may approx-imate those that produce the diuretic effects seen here and inother in vivo and in vitro studies. The dramatic responses of thekidney in vivo to i.v. EGF infusion in this and in previousstudies [7—9] correspond to in vitro findings [4, 6] that EGF,perfused extraluminally but not luminally, inhibited AVP-stim-ulated water conductivity and sodium reabsorption from iso-lated rabbit cortical collecting tubules. This suggests that local-ly-formed or plasma-derived EGF might influence renalelectrolyte and water homeostasis possibly via regulation of theAVP V2-receptor second messenger system by EGF-stimulatedphospholipase C activity and/or PG production.

Acknowledgments

This work was supported by the National Health and MedicalResearch Council. The authors thank R. Fitzpatrick and A.E. Hayes fortechnical assistance.

Reprint rquests to Paddy A. Phillips, Department of Medicine,University of Melbourne, Austin Hospital, Heidelberg 3084, Victoria,Australia.

References

1. KANDA S, NOMATA K, SAHA PK, NISHIMURA N, YAMADA J,KANETAKE H, SAITO Y: Growth factor regulation of the renalcortical tubular cells by epidermal growth factor, insulin-likegrowth factor-I, acidic and basic fibroblast growth factor, andtransforming growth factor-p in serum free culture. Cell Biol mtRep 13:687—699, 1989

2. NORMAN J, BADIE-DEZFOOLY B, Noiw EP, KURTZ I, SCHLOSSERA, CHAUHDARI A, FINE LG: EGF-induced mitogenesis in proximaltubular cells: potentiation by angiotensin II. Am J Physiol 253:F299—F309, 1987

3. HUMES DH, CIE5LINSKI DA, COIMBRA TM, MESSANA JM, GAL-vo C: Epidermal growth factor enhances renal tubule cell regen-eration and repair and accelerates the recovery of renal function inpostischemic acute renal failure. J Clin In vest 84:1757—1761, 1989

4. BREYER MD, JACOBSON HR, BREYER JA: Epidermal growth factorinhibits the hydroosmotic effect of vasopressin in the isolatedperfused rabbit cortial collecting tubule. J Clin Invest 82:1313—1320,1988

5. MUTO S, FURUYA H, TABEI K, A5ANO Y: Site and mechanism ofaction of epidermal growth factor in rabbit cortical collecting duct.Am J Physiol 260:F163—Fl69, 1991

6. VEHASKARI VM, HERING-SMITH KS, MosKowlTz D, WEINER ID,HAMM LL: Effect of epidermal growth factor on sodium transportin the cortical collecting tubule. Am J Physiol 256:F803—F809, 1989

7. Gow CB, Mooit GPM: Epidermal growth factor alters milkcomposition and fluid balance of lactating ewes. J Endocrinol 132:377—385, 1992

8. Gow CB, WILKINSON M, SILVAPULLE Mi, Mooi GPM: Fluidbalance, electrolyte profiles and plasma parathyroid hormone con-centrations in ewes treated with epidermal growth factor. J Endo-crinol 135:91—101, 1992

9. 00w CB, PHILLIPS PA, CHANDLER KD, Mooi GPM: Hormonal,fluid and electrolyte responses of sheep during chronic intravenousinfusion of epidermal growth factor. Am J Physiol 265:R203—R210,1993

10. Gow CB, PHILLIPS PA: Epidermal growth factor (EGF) is diureticand natriuretic. (abstract) Proc Aust NZ Soc Nephrol 1993, p 107

11. WOODS RL, JOHNSTON CI: The maintenance of blood pressureregulation during dehydration. Am J Physiol 245:F615—F621, 1983

12. OGAWA K, SMITH Al, HODGSON GP, JACKSON B, WOODCOCK EA,JOHNSTON CI: Plasma atrial natriuretic peptide: concentration andcirculating forms in normal man and patients with chronic renalfailure. Glitz Exp Pharmacol Physiol 14:95—102, 1987

13. TRINDER D, STEPHENSON JM, GAO X, PHILLIPS PA, RISVANIS J,JOHNSTON CI: [3H]desGlyNH29d(CH2)5[D-Ileu2, lIeu4) AVP: AnAVP V2 receptor antagonist radioligand. Peptides 12:1195—1200,1991

14. KELLY JM, ABRAHAM JM, PHILLIPS PA, MENDELSOHN FAO,GRZONKA Z, JOHNSTON CI: ['251]-[d(CH2)5, Sar7]AVP: A selectiveradioligand for V1 vasopressin receptors. J Recept Res 9:27—41,1989

15. TRINDER D, PHILLIPS PA, RISYANIS J, STEPHENSON JM,JOHNSTON CI: Regulation of vasopressin receptors in deoxycorti-costerone acetate-salt hypertension. Hypertension 20:569—574,1992

16, RALL LB. SCOTT J, BELL 01, CRAWFORD RJ, PENSHOW JD, NIALLHD, COGLAN JP: Mouse prepro-epidermal growth factor synthesisby the kidney and other tissues. Nature 313:228—231, 1985

17. KANDA S, SAHA PK, NOMATA K, TAIDE M, NISHIMURA N, IGAWAT, YAMADA J, KANETAKE H, SArrO Y: Transient increase in renalepidermal growth factor content after unilateral nephrectomy in themouse. Acta Endocrinol 124:188—193, 1991

18. LEV-RAN A, HWANG DL, BEN-EZRA J, WILLIAMS LE: Origin ofurinary epidermal growth factor in humans: excretion of endoge-nous EGF and infused ['3111-human EGF and kidney histochemis-try. Am J Physiol 227:801—805, 1992

19. PANARETTO BA, 'MOORE GPM, ROBERTSON DM: Plasma concen-trations and urinary excretion of mouse epidermal growth factorassociated with the inhibition of food consumption and wool growthin Merino wethers. J Endocrinol 94:191—202, 1982

20. WYNN PC, MADDOCKS 1G. MooRE GPM, PANARETTO BA, DJURAP, WARD WG, FLECK E, CHAPMAN RE: Characterization andlocalization of receptors for epidermal growth factor in ovine skin.JEndocrinol 121:81—90, 1989

21, MARATOS-FLIER E, KA0 CYY, VERDIN EM, KING GL: Receptormediated vectorial transcytosis of epidermal growth by Madin-Darby canine kidney cells. J Cell Biol 105:1595—1601, 1987

22. REDHA RM, LOPEZ C, BREYER JA, JACOBSON HR, BREYER MD:Mapping of 1254 epidermal growth factor binding along the rabbitnephron. (abstract) Kidney mt 35:319, 1989

23. HARRIS RC: Response of rat inner medullary collecting duct toepidermal growth factor. Am J Physiol 256:Fl 117—Fl 124, 1989

24. TEITELBAUM I, STRASHEIM A, BERL T: Epidermal growth factor-stimulated phosphoinositide hydrolysis in cultured rat inner med-ullary collecting tubule cells: Regulation by 0 protein, calcium andprotein kinase. J Clin Invest 85:1044—1050, 1990

25. SC000INS BA, BUTKUS A, COGHLAN JP, FEI DTW, MCDOUGALLJG, NIALL HD, WALSH JR, WANG X: In vivo cardiovascular, renaland endocrine effects of epidermal growth factor in sheep, inEndocrinology: Proc 7th International Congress of EndocrinologyQuebec City, edited by LABRIE F, PROULX L, New York, ElseviersScience, 1984, pp. 573—576

26. CARTER NB, FAwCETT AA, HALES JRS, Moor GPM, PANA-RETTO BA: Circulatory effects of a depilatory dose of mouseepidermal growth factor in sheep. J Physiol 403:27—39, 1988

27. TEITELBAUM I: Cyclic adenosine monophosphate and diacylglyc-erol. J Glitz Invest 86:46—51, 1990

28. LEVINE L, HASSID AF: Epidermal growth factor stimulates pros-taglandin biosynthesis by canine kidney (MDCK) cells. BiochemBiophys Res Commun 76:1181—1187, 1977

29. MARGOLIS BL, BONVENTRE JV, KREMER SG, KUDLOW JE, SKO-RECK! KL: Epidermal growth factor is synergistic with phorbolesters and vasopressin in stimulating arachidonate release and

1036 Phillips et a!: Antagonism between EGF and AVP

prostaglandin production in renal glomerular mesangial cells. Bio-chem J 249:587—592, 1988

30. ANDERSON RI, BERL T, MCDONALD KM, SCHRIER RW: Evidencefor an in vivo antagonism between vasopressin and prostaglandin inthe mammalian kidney. J Gun Invest 56:420—426, 1975

31. COWLEY AW, MERULL DC, SMITH MJ, SKELTON MM: Role of

vasopressin in regulation of sodium excretion. Am J Med Sci 2 95:

308—313, 198832. HEBERT RL, LAMOUREUX C, SIRoIs P, BRAQUET P, PLANTE GE:

Interaction between prostaglandin E2 and leukotriene D4 on theexcretion of electrolytes by the dog kidney in vivo. Prostag!andins33:301—314, 1987