Journal of Clinical InvestigationVol. 46, No. 7, 1967

Functional Characteristics of the Diluting Segment of theDog Nephron and the Effect of Extracellular Volume

Expansion on Its Reabsorptive Capacity *

GARABEDEKNOYAN,WADI N. SUKI, FLOYDC. RECTOR, JR., ANDDONALDW. SELDIN t

(From the Department of Internal Medicine, The University of Texas Southwestern MedicalSchool, Dallas, Texas)

Summary. The functional-characteristics of the ascending limb of Henle'sloop were examined during hypotonic saline infusion by measuring solute-free water clearance (CH2o) at varying rates of solute delivery. The influenceof expansion of extracellular volume was studied by comparing CH2o duringhypotonic saline diuresis in normal dogs with dogs whose extracellular vol-ume had been expanded acutely by saline infusions or chronically by theadministration of deoxycorticosterone acetate and salt.

In normal animals hypotonic saline infusions greatly increased urine flow(V) and CH2O without appreciably augmenting osmolar clearance (Cosm).CH2Owas, therefore, analyzed as a function of V, rather than Cosm, since V wasthe best estimate of delivery of filtrate to the diluting segment. CH90 in-creased as a linear function of V without any evidence of saturation.

The validity of interpreting increases in CH2o and V as indications of in-creased sodium reabsorption and delivery was reinforced by tissue studiesthat disclosed a rise in papillary osmolality with rising urine flows. Theobserved increase in CE20 and V could not, therefore, be due to a decreasein back diffusion of solute-free water as a'result of a diminished osmotic driv-ing force, but probably represented increased formation consequent to aug-mented delivery as a result of decreased fractional reabsorption in the proxi-mal tubule.

In animals whose extracellular volume was acutely or chronically over-expanded before the infusion of hypotonic saline, sodium excretion wasgreater, and CH20 less, at any given V. Although the curve relating CH20 toV was flatter than in the control group, no tubular maximum was observed.The diminished CHO2 in this group was interpreted to mean that massive ex-pansion of extracellular volume inhibits sodium reabsorption in the ascendinglimb of Henle's loop.

Introduction of urine is the reabsorption of sodium in a water-

It is now well established that the principal impermeable segment of the nephron. In conse-process mediating the concentration and dilution quence, the clearance of solute-free water (CH2O)

and the reabsorption of solute-free water (TCH9o)* Submitted for publication September 13, 1966; ac- r ipthe magnitudeof so dium reab-

cepted March 22, 1967. reflect principally the magnitude of sodium reab-Supported in part by grants 5-SOI FR-5426-05 and sorption in the ascending limb. It is possible,

5-TI HE-5469 from the National Institutes of Health therefore, that measurement of these parametersand in part by a grant from the Dallas Heart Association.

t Address requests for reprints to Dr. Donald W. Southwestern Medical School, 5223 Harry Hines Blvd.,Seldin, Dept. of Internal Medicine, University of Texas Dallas, Texas 75235.

1178

FUNCTIONAL CHARACTERISTICS OF THE DILUTING SEGMENT

may permit a characterization of sodium reabsorp-tion in this segment.

The use of TeC2o as an index of sodium reab-sorption in the ascending limb has proved haz-ardous. Previous studies in this laboratory (1)and by Goodman, Cohen, Levitt, and Kahn (2)have indicated that with increasing solute clear-ance in the dog TCHSO, instead of rising or remain-ing unchanged, actually diminishes and finally be-comes negative. This is doubtless a result of thefact that distal tubular fluid in the dog is alwayshypotonic (3). As solute diuresis mounts, there-fore, delivery of hypotonic fluid into the medullaryportion of the collecting duct results in a calculatedvalue for TCH2Othat is falsely low and does not re-flect the full extent of water reabsorption (1).TCH2O will thus tend to underestimate to an un-predictable extent the magnitude of sodium reab-sorption in the ascending limb during rapid rates ofsolute diuresis.

CH2O is a better index of sodium reabsorption inthe ascending limb, since the low water perme-ability in the distal tubule is actually an advan-tage. It has been established in several studies(4, 5) that CH2O depends principally on the load

of sodium delivered to the diluting segment. How-ever, as sodium delivery is increased, usually bymannitol infusion, a maximum for CH2O is reachedbeyond which no further free water can be gen-erated despite increasing solute excretion (6, 7).These studies, however, suffer from the disad-vantage that CH20 was measured over a compara-tively narrow range of solute clearance (COsm) .Moreover, the use of mannitol has a serious dis-advantage. Mannitol will depress the concentra-tion of sodium delivered to the diluting segment,thereby placing a concentration limit on the amountof sodium that can be reabsorbed in the ascendinglimb independent of the intrinsic capacity. Inother studies (8) the use of diuretics to assessCH20 is especially unreliable, since the diureticsmay themselves impair sodium transport in thediluting segment. For these reasons the charac-terization of sodium transport in the ascendinglimb using CH20 has not been entirely satisfactory.

We have attempted to circumvent these diffi-culties by using saline diuresis as a means of in-creasing delivery of sodium to the ascending limb.Control studies were first performed in which CHSOwas examined over a wide range of sodium de-

livery, which was accomplished by hypotonic sa-line infusions. The influences of a variety offactors that might affect CH2O independent of de-livery-hyponatremia, medullary blood flow, ef-fective extracellular fluid (ECF) volume-couldthen be identified.

Methods

Studies were performed on female mongrel dogsweighing 10 to 18 kg fed regular commercial diets.The same general procedure was followed in all studies.

On the day of study an indwelling urethral catheterwas inserted under mild thiopental sodium anesthesia;while still anesthetized all dogs were given 500 ml ofwater by stomach tube; thereafter the dogs were al-lowed to awaken. Water diuresis was maintained bythe infusion of 0.45% sodium chloride, which was startedat rates of 6 ml per minute and increased thereafter torates up to 40 to 50 ml per minute. The rate of infu-sion was such that it always exceeded urine flow rate by5 ml per minute. Since the experiments lasted 3 to 4hours the positive balance of hypotonic saline at the endof the experiment was 900 to 1,200 ml. Four dogs wereinfused, in addition, with 2.5% glucose solution at 4 to 6ml per minute throughout the entire experiment to en-sure the development of pronounced hyponatremia. Uri-nary collection periods were 15 minutes, and arterialblood samples were drawn from an indwelling Cournandneedle in the femoral artery at the midpoint of each col-lection period. After an appropriate loading dose, a sus-taining infusion of inulin and para-aminohippurate (PAH)in normal saline was given at about 2 ml per minute.Enough potassium chloride was added to this infusion todeliver 90 ,Eq per minute of potassium, in order to com-pensate for the potassium loss sustained during the mas-sive diuresis that ensued.

In the first series of experiments CH20 was studied in4 groups of animals during hypotonic saline diuresis.Group I consisted of 4 normal dogs. Group II consistedof 14 dogs pretreated for 7 days with an im injectionof deoxycorticosterone acetate (DOCA), 1 to 2 mg perkg body weight daily. The diet was supplemented with5 to 10 g potassium chloride daily to prevent hypokalemia.Group III consisted of 3 unilaterally nephrectomized dogs,similarly treated with DOCA. Group IV consisted of 3normal dogs infused with 2 to 3 L of isotonic saline overa 5-hour period immediately preceding water administra-tion and hypotonic saline infusion.

In the second series of experiments renal hemodynamicswere studied in normal and DOCA-treated animals. Theanimals were prepared 2 to 3 weeks before the study.The left kidney and renal vein were exposed through anabdominal incision, and the left ovarian vein was ligatedand cut. A Teflon catheter was then introduced intothe left renal vein through the inferior vena cava andconnected with a silastic catheter filled with heparinizedsaline. The free end of the silastic catheter was tied

1179

EKNOYAN, SUKI, RECTOR, AND SELDIN

with umbilical tape and implanted subcutaneously in thesubcostal abdominal wall. On the day of study, afterinduction of a water diuresis, the free end of the silasticcatheter was freed, flushed with saline, and used for ob-taining samples of left renal venous blood. Under lightpentobarbital anesthesia the right ureter was exposedthrough a flank incision and cannulated with polyethylenetubing. Urine samples from the left kidney were ob-tained from the bladder. Arterial and venous bloodsamples were collected simultaneously into heparinizedsyringes, transferred to ice-cooled tubes, and immediatelycentrifuged at 4° C in a refrigerated Sorvall centrifugeto prevent leakage of PAH from red blood cells.

In the third series of experiments the effect of waterdiuresis and hypotonic saline diuresis on medullary tonicitywas studied by tissue analysis. After induction of waterdiuresis or hypotonic saline diuresis, the dogs were anes-

thetized with pentobarbital, and a minimum of two con-

secutive collection periods with stable urine flow and os-

molality was obtained before the kidneys were removed.The kidneys were sectioned transversely; the cortex,outer medulla, and inner medulla were identified by gross

anatomic characteristics and random pieces of each takenfor analysis. The tissue slices were blotted on filterpaper and weighed. They were then dried under vacuum

at 60° C for 36 hours and weighed again. The differencein weight was considered the water content. Each driedtissue sample was ground, made up to a volume of 10 mlwith demineralized water, allowed to stand overnight,and then autoclaved at 120° C under a pressure of 15pounds per square inch for 2 hours. The samples were

centrifuged and portions of the supernate taken for sodiumand potassium analysis. Urea was initially analyzedboth on fresh and homogenized samples, but was negligi-ble, and was therefore not determined in later studies.

Inulin was analyzed by the method of Walser, David-son, and Orloff (9), PAH by the method of Smith andco-workers (10), as adapted for the Technicon Auto-analyzer. Serum and urinary sodium and potassium were

determined by internal standard flame photometry as

modified for the Autoanalyzer. Plasma and urinary os-

molality were measured on the Fiske osmometer.

Cosm and CH2o were calculated in the usual fashion andexpressed per 100 ml glomerular filtrate for the purpose

of comparing results from different experiments. Totalrenal plasma flow (RPF) was calculated from the for-mula of Wolf (11): RPF= V (U - R)/(A - R), whereV is urine flow and U, R, and A, the PAH concentrationof the urinary, renal venous, and arterial samples, re-

spectively. Clearance of PAH (CPAH) was assumed toapproximate renal cortical plasma flow, and noncorticalplasma flow (NCPF) calculated from the difference be-tween RPF and CPAEI. Noncortical blood flow has beenwidely used as an estimate of medullary blood flow (12,13). Papillary osmolality was calculated as twice thesum of the concentrations of sodium and potassium intissue water. Concentration of urea, in all experimentsin which it was measured, was found to be negligible andtherefore not included in these calculations.

Results

Effect of hypotonic saline infusion on electro-lyte and water excretion. The first series of stud-ies concerns the effects of hypotonic saline infu-sion in normal dogs (group I). The protocol ofa representative experiment is presented in TableI (dog M). With infusion of hypotonic saline, Vrose, urinary osmolality (Uosm) fell slightly, Cosmrose only slightly, and CH2o rose sharply. Theglomerular filtration rate (GFR) was relativelystable. Sodium excretion increased from 10 to167 ,uEq per minute.

The results of the group as a whole are chartedin Figure 1. CH2O is plotted as a function of COsmand V. As V mounted to a maximum of 17.8 mlper minute per 100 ml GFR, CH2o rose sharplyfrom 6.3 to 15.0 ml per minute per 100 ml GFR.This rise in CH2Owas associated with only a mini-mal increase in Cosm from 1.3 to 3.1 ml per minuteper 100 ml GFR. Despite continued high ratesof infusion it was impossible to increase Cosm be-yond 3.1% of GFR. No maximum for CEI2ocould be detected, although the range of Cosm ad-mittedly was small.

The infusion of hypotonic saline into normalanimals, though associated with a marked rise in Vand CH2o, produces only a slight absolute increasein Na excretion, although the relative rise is great(Table I). The increase in V cannot reasonablybe attributed to further suppression of antidiu-retic hormone secretion. During the initial period,the concentration of serum sodium was low, ECFvolume was expanded, and the urine was maxi-

14

12

WO00d 10o *

8y-~E Ocosm

W 90 6 0 Urine Flow

0 4 8 12 16 20 24 28ml/min/IOOmi GFR

FIG. 1. THE FRACTIONAL CLEARANCEOF SOLUTE-FREEWATER (CH2o) PLOTTED AGAINST FRACTIONAL OSMOLARCLEARANCE(Co.m) AND AGAINST FRACTIONAL URINE FLOWIN UNTREATEDCONTROLDOGS.

1180

FUNCTIONALCHARACTERISTICSOF THE DILUTING SEGMENT 1181

0%a aa ) ot i 00 0C414 cM OOOono OOO~ncmamoj4 C *

CL,~~~~~~~~~~~~~~~~~~~~~~~~~~CCa B WOOOncd I- 00 -tin -os n n sn-ommO omCr-400X oZ cqtoeq qen v14i 1tmen C14m n en~~m m

X C- elno 0 C 0 Cs-,o --tm 0~r- 0 --t00 no Z9.10 r oE-m_ 6) oo O5 e6 4 ui N ci 4 rZ c; O5 o6 Xoco

81 -N- -l-N n m-M 4)-- _ ,,_____o =O~ _ d 0

Z t

Ira X tO00-40t f'Qc1

i > ---4

a)0- 4-4 -

0~~~~~~~~~~~~~~

?>~~~~~ ~ ~ -t-~+ W - ee o e o o_ ~ ; 8x

U,~~~~~~~~~~~~Uc

t~~~~~~~~~e 00 _N _ WcK W -- r- mf<7 mac~ No E3 0

t E $ O_0 0 _ 0-- e + o

U W~~~~~~~~~~~~~~~~~U

*S~~~~~~~~~~~~o -t ~oM -4 to M Cq Cq Oeq+ -O0\O- =0

.v U

X - o W

qc42 --mm c a)

o C11 C1 -11 r- We#,) --o -4oe n o)

It O obXXXo ;*

0~~~~~ II a) a))WoU oly . t wo ooo c

I, 0 ,

0 -.--e4 e ici --c cd +L

OUM),4 \0 r- -Rit U-) UX

4mmm 5 \ o -m--d' in C c

-o~~~~~~~~~~~~~~~~~ca) '0 ~0 04cq c

00~~~~~~~~~~~~~

0>t.u~ ca

EKNOYAN, SUKI, RECTOR, AND SELDIN

mally hypotonic. It is, therefore, much more rea-

sonable to attribute the rise in V to increased freewater formation resulting from suppression ofproximal reabsorption because of expansion of ex-

tracellular volume with hypotonic saline. Theamount of saline given is in excess of that shownby micropuncture studies to reduce the ratio oftubular fluid to plasma inulin at the end of theproximal tubule (14). The failure of sodiumexcretion to rise greatly as urine flow mounts isprobably the result of hyponatremia. Ample evi-dence exists that the concentration of serum so-

dium per se influences sodium reabsorption (15).The present studies constitute strong evidence thatthe site in the nephron where hyponatremia stimu-lates Na reabsorption is the ascending limb ofHenle's loop.

Because of the difficulty in the preceding experi-ments in achieving sufficiently high rates of Cosmand V to evaluate CH2o adequately, the studieswere repeated in animals treated chronically withDOCA(group II). Animals treated in this fash-ion are known to have an enhanced sodium excre-

tion in response to saline infusion. The resultsof such an experiment are presented in Table I(dog S-5). The infusion of hypotonic saline intothe DOCA-treated dog resulted in much higherfractional urine flows (V/GFR X 100) and frac-tional Cosm (Cosm/GFR X 100) than were achievedin the normal dog (dog M).

The data for the entire DOCA-treated group

(group II) are plotted in Figure 2. In contrastto the normal dogs Cosm ranged from 3.2 to as

w c,uLL

'Q

cr .''-iE

w '3 E

ZI 24 32 36 40 44 48milminII00mi GFR

FIG. 2. THE FRACTIONAL CLEARANCEOF CH,0 PLOTTED

AGAINST FRACTIONAL Cosm ANDAGAINST FRACTIONAL URINE

FLOWIN DOCA-TREATEDDOGS.

c 20

E:- I

16U.0E 1400

0x"12 10

8

6

ej 4

2ui

La

0 00

0

0

0~~~~~

:. OSerumNo

*

ok

4F4

Serum No0o

I00

0O

Ma< 140 mEq/ |a> 140 mEq/l

*aI,0 4 8 12 16 20 24 28 32 36 40 44 48

URINARY FLOW(v/100lml GFR (mi/min)

FIG. 3. THE EFFECT OF SERUMCONCENTRATIONON THE

RELATION OF FRACTIONAL CH2o TO FRACTIONAL URINE FLOW

(V) IN DOCA-TREATEDDOGS.

high as 24 ml per minute per 100 ml GFR, and Vranged from 4 to 41 ml per minute per 100 mlGFR. Fractional CHO2, however, was only slightlygreater (17.2) than in the normal animals (15.0,Figure 1). Even with this massive increase inCoSm, no maximum for CH20 could be demon-strated. To determine whether the lower CHOwith DOCA-treated dogs as compared to controlanimals at any given V was the result of inade-quate hydration, we performed studies on DOCA-treated dogs infused with 2.5% dextrose solutionat the rate of 4 to 6 ml per minute for 3 to 4 hoursbefore the onset of the study. The results of one

such experiment are shown at the bottom of Ta-ble I (dog SAL-6). Despite marked hypona-tremia, Uosm was comparatively higher and CH20lower at any level of V than in the untreated con-

trol dog (dog M). Moreover, these results were

quite comparable to those of the DOCA-treatedanimal with the higher serum sodium concentra-tion (dog S-5). Indeed a comparison of theDOCA-treated dogs with concentrations of serum

sodium above and below 140 mEqper L (Figure3) discloses no influence of the concentration ofserum sodium on the relation between CH2O andV. This is strong evidence of maximal suppres-

sion of antidiuretic hormone secretion in all theDOCA-treated animals.

In an attempt to increase still further the soluteload per nephron to ascertain whether a maximumfor CH20 could be reached, 3 DOCA-treated dogs

1182

1800

160

as

14 0 * *

8-6

6<getZ~ @ * * i@Cosm80so 0*.on

n 2 at n RA 9n 9- on*Cm ADv q d lz It

FUNCTIONALCHARACTERISTICS OF THE DILUTING SEGMENT

were studied 3 to 6 weeks after unilateral nephrec-tomy to allow for the compensatory hypertrophyand increase in GFR in the remaining kidney(group III). The effect of hypotonic saline in-fusion on CH2Oper 100 ml GFRis plotted in Fig-ure 4 (open triangles), where the shaded area istaken from the results in Figure 2 comparingCH2o and V. All the points from the unilaterallynephrectomized dogs fell within the shaded area,indicating that despite increased GFRper nephron(16), the relation between CHO and V wasunchanged.

It is apparent from Figures 1 and 2 that CH2O,at any given Corm or V, is higher in the normalthan in the DOCA-treated dogs. In Figure 5 thegreater CH2o at any given V in the control animals(crosshatched) than in the DOCA-treated animals(stippled) is more clearly illustrated. To deter-mine whether the difference between the normaland DOCA-treated group was an effect of DOCAor of volume expansion, we studied CH2O in 3 non-DOCAtreated dogs (group IV) immediately aftervolume expansion by isotonic saline (Figure 5,closed triangles). The response of all 3 animalsdiffered from the normal group and was the sameas in the DOCA-treated group. This is strongevidence that the effects of chronic DOCAtreat-ment on CH20 are attributable to extracellular vol-ume expansion, not to DOCAper se. Furtherevidence that DOCAper se has no effect on thediluting segment was obtained in experiments (not

C 20-E

E 4SO:

Q 16

o 14

a- I2

U 10 L

6_

L:Lu 4

4] 2

E &

a a

a~~~~~

. aa2A a~~

a

IN

a

- - --*

4 8 12 16 20 24 28 32 36 40 44 48 52URINARY FLOW (v)/lO ml GFR (ml/min)

FIG. 4. THE EFFECT OF UNILATERAL NEPHRECTOMYONTHE RELATION OF FRACTIONAL CH2o TO FRACTIONALV. Thestippled area represents the relation of CH20 to V in intactDOCA-treated dogs (Figure 2).

c 20-

E 18-

L.-J 6-

E

So\

LNa

'-,

pi

la~

cr

LW 4

A

A

A A

4 8 12 16 20 24 28 32 36 40 44 48 52URINARY FLOW(v)/100 ml G FR (ml/minI

FIG. 5. THE EFFECT OF PRIOR EXTRACELLULARVOLUME

EXPANSION WITH ISOTONIC SALINE ON THE RELATION OF

FRACTIONAL CH2O TO FRACTIONAL V IN DOGS NOT TREATED

WITH DOCA. The stippled and crosshatched areas repre-sent the relation of CH2O to V in DOCA-treated and con-trol dogs, respectively.

here presented) in which the acute administrationof DOCAto group I animals resulted in the samerelation of CH20 to V as is detailed in Figure 1.

Effect of hypotonic saline on renal hemody-namics. To investigate the role of medullary bloodflow during hypotonic saline infusion and extracel-lular volume expansion, we measured PAH ex-traction ratios and calculated renal plasma flowin 6 experiments. The results obtained in 5DOCA-treated and 1 normal dogs are summarizedin Table II, where the period of maximal urineflow in water diuresis and hypotonic saline diuresisis compared. With hypotonic saline loading, therewas an increase in PAH clearance and renalplasma flow, and a fall in the extraction of PAH.The calculated noncortical blood flow uniformlyincreased.

Effect of hypotonic saline infusions on renaltissue composition. Expansion of the extracellularfluid volume, by increasing medullary blood flow(Table II), might cause a washout of medullarytonicity (13), thereby markedly reducing theback diffusion of solute-free water out of the col-lecting duct. To evaluate the possible role ofmedullary washout on CHIo, we measured thecomposition of renal tissue at varying rates ofdiuresis in 18 experiments, 15 on DOCA-treatedand 3 on untreated control dogs. The results oftissue analysis of these 2 groups were indistin-

1183

i-

1184 EKNOYAN, SUKI, RECTOR, AND SELDIN

TABLE II

Renal hemodynamic changes during hypotonic saline diuresis in control and DOCA-treated dogs*

Dogno. V GFR CPAH RPF EPAR NCPF

ml/min ml/min ml/min ml/min % ml/minP-1 Water 5.9 78 208 311 66 103

Saline 21.3 116 290 440 64 150P-2 Water 2.4 40 95 112 85 17

Saline 6.4 37 109 145 74 36P-3 Water 4.3 32 76 105 71 29

Saline 10.6 41 139 241 56 102P-4 Water 3.4 42 112 190 58 78

Saline 7.2 46 174 368 46 194P-5 Water 3.3 26 52 69 74 17

Saline 5.9 30 75 105 69 30P-7 Water 2.2 28 54 64 84 10

Saline 6.0 30 79 103 75 24

* CPAH = clearance of para-aminohippurate; RPF = renal plasma flow; EPAH = PAH extraction ratio; NCPF= noncortical plasma flow. See Table I for other abbreviations. The values on dogs P-2 to P-7 are those obtained fromthe left kidney; only the left kidney was cannulated for the measurement of PAH extraction. In the experiment ondog P-1 only bladder urine was collected. The blood flow data were extrapolated from the extraction of PAHon the leftside. Dog P-7 did not receive DOCA; dogs P-1 through P-5 did.

guishable and were plotted together. Figure 6 in-dicates that as urine flow was increased from 3.7to 10 ml per minute, papillary osmolality rosefrom 280 to 500 mOsmper kg tissue water. Withfurther increases in urine flow from 10 to 25 mlper minute, papillary osmolality remained rela-

PAPILLARYOSMOLALITY

mOsm/KgTissue H20

PAPILLARYSODIUMmEq/Kg

Tissue H20

PAPILLARYPOTASSIUM

mEq/KgTissue H20

PAPILLARYWATER

CONTENT% of Wet Weight

500

400F

300

onn

0 0 0 000

0

00 000.I.

S200k .0I0200 0 * -

0 0 0

150 *to

000

100or *.

060 * *60 t 1b} *

40 P'

0 5 10 15 20 25URINE FLOWmIl/min

FIG. 6. THE EFFECT OF INCREASING URINE FLOWDURING

HYPOTONIC SALINE DIURESIS ON THE ELECTROLYTE AND

WATERCONTENTOF THE RENAL PAPILLA.

tively constant. The increase in papillary osmo-lality as urine flow rose to 10 ml per minute wasalmost entirely due to a rise in tissue sodium con-centration. The potassium concentration showedonly a minimal rise with increasing urine flows.Since the papillary water content remained con-stant throughout, the rise in papillary sodium con-centration represents an accumulation of sodiumin the medulla. Therefore, the increased medul-lary blood flow (Table II) did not affect a washoutof medullary sodium.

Urinary osmolality tended to rise with therise in urine flow, but not proportionately to therise in papillary osmolality, so that when the os-motic gradient between papilla and urine is plottedagainst urine flow (in the 18 experiments in Fig-ure 6) as shown in Figure 7, there is a gradualincrease in osmotic gradient with increasing urineflow, from 178 mOsmto 421 mOsmper kg water.

Z MN400

o3 c :

ei 300

° EE 200

oE1O0

) _

* 0

0*

0 0

: 0

00 0

0

0 5 10 15 20 25URINE FLOW mi/min

FIG. 7. THEEFFECT OF INCREASING URINE FLOWDURINGHYPOTONIC SALINE DIURESIS ON THE OSMOTIC GRADIENTBETWEENPAPILLARY WATERAND URINE.

FUNCTIONALCHARACTERISTICS OF THE DILUTING SEGMENT 1

Discussion

The reabsorption of sodium in the diluting seg-ment of the nephron is the principal factor re-sponsible for CH2o. CH2O, in turn, is a function ofboth delivery of substrate (sodium) and the in-trinsic reabsorptive characteristics of the dilutingsegment. The intrinsic characteristics of sodiumreabsorption in this segment, therefore, can bedelineated only as a function of delivery.

It is usual to express CH2Oas a function of C..m.However, Cosm represents the rate of excretion ofunreabsorbed solute and bears no well-defined re-lation to sodium delivery. Figure 1 discloses thatprogressive hypotonic saline infusion elicits an in-crease in CH2Oper 100 ml GFRfrom 6.3 to 15.0with virtually no increase in Cosm. This meansthat practically the entire increment in sodium de-livery to the diluting segment is reabsorbed there,with only trivial amounts escaping into the urine.COSm, therefore, does not accurately reflect in-creasing delivery and consequently cannot be re-lated to CH2O. The same data are plotted as afunction of V (Figure 1). Since very little wateris lost from the distal nephron during water diu-resis, V is a reasonable approximation of the quan-tity of sodium containing fluid reaching the di-luting segment. An increase in fractional delivery(V) from 8 to 17.8 elicits a linear increase in so-dium reabsorption (CH2o).

The use of CH2O, in this manner, as an estimateof sodium reabsorption and V as an estimate ofdelivery is in error to the extent that back diffusionof solute-free water occurs in the distal nephron.The actual rate of solute-free water formation isequal, not to CH2o, but to CHa2O plus back diffusion.Similarly, the actual rate of delivery is equal, notto V, but to V plus back diffusion. It is clear,therefore, that significant changes in back diffusionmight influence the relation between V and CH2Oindependent of the actual rate of delivery and so-dium reabsorption.

Back diffusion might change as a result of analteration in the osmotic driving force or the per-meability of the tubular epithelium to water.

Earley and Friedler have proposed (13) that sa-line diuresis elicits an increase in medullary bloodflow, thereby causing a washout of medullary so-dium. This might dissipate the osmotic force forthe back diffusion of water out of the collecting

duct. In this view the increase in CH2o as a func-tion of V would represent, not increased formationwith increasing delivery, but rather diminishedback diffusion.

To examine whether a reduction in back diffu-sion could, as a result of a diminished osmoticgradient, account for the relation of CH2O to Vduring hypotonic saline infusion, we measured thecomposition of the inner renal medulla at varyingrates of diuresis. The results of these studies areshown in Figures 6 and 7. As urine flow in-creased from 4 to 25 ml per minute, the papillaryosmolality increased from 280 mOsmper kg tis-sue water to over 500 mOsmper kg tissue water(Figure 6). This increase was entirely due to arise in tissue sodium concentration; potassium re-mained relatively constant. Although urinary os-molality rose with the rise in urine flow, this risewas not proportionate to the rise in papillaryosmolality. As a result, as shown in Figure 7, theosmotic gradient between papilla and urine in-creased. These increases in papillary osmolalityand in the osmotic gradient between the papillaand urine occurred despite a significant rise innoncortical blood flow (Table II). It is evident,therefore, that increased medullary blood flow dur-ing hypotonic saline diuresis does not effectivelyaccelerate net medullary washout, since the os-motic gradient between the papilla and urine ac-tually increases in the face of increasing non-cortical blood flow. We therefore conclude thatthe observed increase in CH20 and V cannot be dueto a diminution in the driving force for water backdiffusion in the medulla.

It is more difficult to define the role of a changein water permeability. It is possible that expan-sion of ECF diminishes the permeability of thedistal nephron to the back diffusion of water. Ifthis were so, the increased CEI20 and V during hy-potonic saline diuresis would represent diminishedback diffusion of water, not increased free waterformation. The increased papillary solute concen-tration could then be regarded as the consequenceof diminished entry of water into the medulla andreduced solute removal.

However, since extracellular volume expansionhas been shown to increase delivery of filtrate tothe diluting segment by suppressing proximal tu-bular reabsorption (14, 17) and since the increasein sodium excretion was slight (Table I, dog M),

1185

EKNOYAN, SUKI, RECTOR, AND SELDIN

it follows that almost all of the increment of sodiummust have been reabsorbed in the diluting segment.An increase in free water formation, urine flow,and medullary sodium concentration, therefore,would result from increased delivery without in-voking changes in permeability. That some of theincreased free water might have resulted fromdiminished back diffusion as a result of impairedpermeability cannot be excluded. There is, how-ever, no need to postulate a separate new effectof extracellular volume expansion on permeabilityfor which there is no independent evidence. Theincreased CH20 and V during hypotonic saline diu-resis, accordingly, appear to represent principallyincreased free water formation as delivery is aug-mented. C320 then appears to be a reasonableestimate of sodium reabsorption in, and V a rea-sonable estimate of delivery to, the dilutingsegment.

It is clear from Figure 1 that sodium reabsorp-tion in the diluting segment (CH2o) increases as alinear function of delivery (V) without any evi-dence of saturation. These results differ fromprevious studies in which a maximum for CH2,Owas demonstrated with progresive solute diuresis(6, 7, 18). Part of the disparity may be attrib-utable to the use of nonreabsorbable solute (man-nitol) as a means of increasing delivery. Whenmannitol is used to induce solute diuresis, the con-centration of sodium reaching the ascending limbis sharply reduced. This could impose a ceilingon sodium reabsorption independent of the in-trinsic reabsorptive properties of the tubule andmay explain the maximum for CH2Oachieved dur-ing mannitol diuresis.

It is also possible that a maximum (Tm) forCH2o was not achieved in the present studies (Fig-ure 1) because delivery could be increased onlymoderately. By expanding ECF volume chroni-cally with DOCAplus sodium chloride, we wereable to achieve much higher rates of delivery tothe ascending limb (Figure 2). Although the re-lation of CH2Oto V is curvilinear, no evidence of aTm for CH2Owas obtained, despite an increase indelivery up to 42% of the GFR. Further supportfor the absence of an intrinsic tubular maximumfor CH20 is obtained from the experiments on dogswith unilateral nephrectomy. The adaptive in-crease in GFRper nephron of 60 to 70%o resultsin near doubling of the absolute rate of delivery

to the diluting segment at any given fractional rateof urine flow. Even at this enhanced rate of so-dium delivery no limitation on CH2O was observed(Figure 4). We, therefore, conclude that the in-trinsic mechanism for sodium reabsorption in thediluting segment exhibits no Tm even when ratesof delivery are increased to as high as 42%o of theGFR.

A comparison of the relation between CH2o andV with and without DOCA (Figure 5, shadedareas) discloses that CH20 is less at any given Vin the DOCA-treated than in the untreated con-trol animals. Three factors might explain the al-tered relation between CH20 and V in the DOCA-treated dogs: hypernatremia, DOCAper se, andECF volume expansion. Chronic treatment withDOCAtended to produce hypernatremia. Sinceseveral investigators (19, 20) have demonstratedan over-all suppression of sodium reabsorption byhypernatremia, it is conceivable that this effect oc-curs in the distal nephron and, therefore, might ac-count for the augmented sodium excretion anddiminished CH2o at any given rate of delivery inthe DOCA-treated group. This possibility wasexcluded by forcing massive amounts of water toDOCA-treated animals, thereby reducing the se-rum sodium to levels observed in the untreatedanimals; nevertheless, the same depression in CH20per unit delivery persisted (Figure 3 and Table I,dog SAL-6).

There was no evidence that hypernatremia inthe DOCA-treated animals reduced CH20 by elicit-ing the secretion of vasopressin. The reduction inthe serum sodium concentration to very loxv valuesby forcing water in the DOCA-treated group didnot alter the relation between CH20 and V. Thefailure of hypernatremia to influence CTT20 was notdue to a refractoriness of the tubule to vasopressin.In unpublished studies we have demonstrated thatinjection of minute amounts of vasopressin toDOCA-treated animals at the height of hypotonicsaline diuresis promptly results in a rise in urineconcentration. We, therefore, conclude that theoverexpansion of ECF volume in DOCA-treatedanimals given hypotonic saline suppresses vaso-pressin secretion even during hypernatremia. Thisis in accord with the demonstration by Arndt (21)that small changes in volume can override the ef-fect of the concentration of serum sodium on anti-diuretic hormone secretion.

186

FUNCTIONALCHARACTERISTICSOF THE DILUTING SEGMENT

Conceivably, DOCAper se might inhibit sodiumreabsorption, perhaps by competing with somemore potent mineralocorticoid. The fact that theacute administration of DOCAto normal dogsduring hypotonic saline diuresis did not depressCH20 is strong evidence that DOCAas such wasnot exerting an inhibitory effect on the dilutingsegment.

Finally, the reduced CH2Oat any given V in theDOCA-treated dogs might be the consequence ofvolume expansion. The results from experimentsin which ECFwas acutely expanded in the absenceof DOCAstrongly support this hypothesis. Theadministration of 3 L of isotonic saline immediatelybefore the free water study resulted in the samedepression in CH20 as was observed in animalspretreated with DOCA(Figure 5).

This depression of CH20 by chronic DOCAtreatment could result from increased back dif-fusion of water in the collecting duct or diminishedfree-water formation in the diluting segment. Aneffect on permeability seems unlikely. As men-tioned earlier volume expansion has already beendemonstrated to inhibit sodium reabsorption in theproximal tubule, and it seems reasonable to as-sume, therefore, that the same process inhibitingsodium transport may be operative in the dilutingsegment rather than to postulate a special addi-tional effect on permeability. In support of thehypothesis that volume expansion is suppressingfree water by diminishing sodium reabsorption inthe diluting segment is the observation that theDOCA-treated dogs respond to hypotonic salineinfusions with an increase in both sodium excretionand CH20. In the untreated animals hypotonicsaline infusions increased CH20 with very littlechanges in sodium excretion (Table I). The factthat volume expansion in the DOCA-treated dogssuppresses CH2O at any given V and at the sametime increases sodium excretion suggests that thesite in the nephron where volume expansion sup-presses sodium reabsorption is the diluting seg-ment.

These results suggest that with modest volumeexpansion as is produced in the untreated animalsby hypotonic saline infusion alone, sodium reab-sorption is suppressed principally in the proximaltubule so that virtually all the increment of sodiumdelivered to the diluting segment is reabsorbedthere and is reflected in increased CE2o. With

more massive volume expansion as in the DOCA-treated -animals or in the animals acutely overex-panded with 2 to 3 L of isotonic saline, there is, inaddition, suppression of sodium reabsorption in thediluting segment so that at any given V, CH2o isless and sodium excretion is higher.

The ability of massive expansion of extracellu-lar fluid volume to suppress sodium reabsorptionin the diluting segment may explain the claim ofseveral investigators (6, 7, 17) that a Tmfor CH2Oexists. It is highly likely that an infusion of hy-potonic saline, if continued long enough, would sogreatly expand ECF volume that a Tm for CH2Omight be obtained. Such a Tmwould be the ex-pression, not of any intrinsic limitation on sodiumreabsorption in the diluting segment, but ratherof progressive inhibition as a result' of volumeexpansion.

Wehave used CHSOas an index of sodium reab-sorption in the diluting segment without specifyingthe anatomic sites of this activity. As a first ap-proximation the diluting segment might be definedas the area in the nephron which is water im-permeable during maximal water diuresis and fromwhich sodium is reabsorbed. This might, there-fore, include the ascending limb, the distal con-volution, and the collecting duct. Micropuncturestudies in the rat (22) and the dog (3) have dem-onstrated that the osmotic pressure of the tubularfluid during water diuresis does not change fromthe beginning to the end of the distal convolution,and actually rises between the end of the distalconvolution and the renal pelvis. This means thatnet CH2O must be formed almost entirely in theascending limb. We therefore conclude that -thesuppression of CH2o by massive expansion of ECFvolume means inhibition of sodium reabsorption inthe thick ascending limb of Henle's loop.

References1. Goldsmith, C., H. K. Beasley, P. J. Whalley, F. C.

Rector, Jr., and D. W. Seldin. The effect of saltdeprivation on the urinary concentrating mecha-nism in the dog. J. clin. Invest. 1961, 40, 2043.

2. Goodman, B., J. A. Cohen, M. F. Levitt, and M. Kahn.Renal concentration in the normal dog: effect ofan acute reduction in salt excretion. Amer. J.Physiol. 1964, 206, 1123.

3. Clapp, J. R., and R. R. Robinson. Osmolality ofdistal tubular fluid in the dog. J. clin. Invest.1966, 45, 1847.

1187

EKNOYAN, SUKI, RECTOR, AND SELDIN

4. Burg, M. B., S. Papper, and J. D. Rosenbaum. Fac-tors influencing the diuretic response to ingestedwater. J. Lab. clin. Med. 1961, 57, 533.

5. Van Giesen, G., M. Reese, F. Kiil, F. C. Rector, Jr.,and D. W. Seldin. The characteristics of renalhypoperfusion in dogs with acute and chronicreductions in glomerular filtration rate as disclosedby the pattern of water and solute excretion afterhypotonic saline infusions. J. clin. Invest. 1964,43, 416.

6. Orloff, J., and M. Walser. Water and solute excre-tion in Pitressin-resistant diabetes insipidus (ab-stract). Clin. Res. 1956, 4, 136.

7. Buchborn, E., H. Edel, and S. Anastasakis. ZurEndokrinologie des distalen Nephrons. 1. Kli-nische Differenzierung und Messung der PhasenI, II, und III der Harnkonzentrierung. Klin.Wschr. 1959, 37, 347.

8. Kleeman, C. R., R. Cutler, M. H. Maxwell, L. Bern-stein, and J. T. Dowling. Effect of various diu-retic agents on maximal sustained water diuresis.J. Lab. clin. Med. 1962, 60, 224.

9. Walser, M., D. G. Davidson, and J. Orloff. The re-nal clearance of alkali-stable inulin. J. clin. In-vest. 1955, 34, 1520.

10. Smith, H. W., N. Finkelstein, L. Aliminosa, B. Craw-ford, and M. Graber. The renal clearances ofsubstituted hippuric acid derivatives and other aro-matic acids in dogs and man. J. clin. Invest. 1945,24, 388.

11. Wolf, A. V. Total renal blood flow at any urineflow or extraction fraction (abstract). Amer. J.Physiol. 1941, 133, 496.

12. Pilkington, L. A., R. Binder, J. C. M. De Haas, andR. F. Pitts. Intrarenal distribution of blood flow.Amer. J. Physiol. 1965, 208, 1107.

13. Earley, L. E., and R. M. Friedler. Observations onthe mechanism of decreased tubular reabsorption of

sodium and water during saline loading. J. clin.Invest. 1964, 43, 1928.

14. Dirks, J. H., W. J. Cirksena, and R. W. Berliner.The effect of saline infusion on sodium reabsorptionby the proximal tubule of the dog. J. clin. In-vest. 1965, 44, 1160.

15. Goldsmith, C., F. C. Rector, Jr., and D. W. Seldin.Evidence for a direct effect of serum sodium con-centration on sodium reabsorption. J. clin. Invest.1962, 41, 850.

16. Bricker, N. S., S. Klahr, and R. E. Rieselbach. Thefunctional adaptation of the diseased kidney. I.Glomerular filtration rate. J. clin. Invest. 1964, 43,1915.

17. Rector, F. C., Jr., G. Van Giesen, F. Kiil, and D. W.Seldin. Influence of expansion of extracellularvolume on tubular reabsorption of sodium inde-pendent of changes in glomerular filtration rateand aldosterone activity. J. clin. Invest. 1964, 43,341.

18. Stein, R. M., R. G. Abramson, and M. F. Levitt.Evidence for a "limit" on distal sodium transportinduced by hypotonic saline loading (abstract).Clin. Res. 1965, 13, 558.

19. Blythe, W. B., and L. G. Welt. Dissociation be-tween filtered load of sodium and its rate of ex-cretion in the urine. J. clin. Invest. 1963, 42,1491.

20. Kamm, D. E., and N. G. Levinsky. Inhibition of renaltubular sodium reabsorption by hypernatremia. J.clin. Invest. 1965, 44, 1144.

21. Arndt, J. 0. Diuresis induced by water infusion intothe carotid loop and its inhibition by small hemor-rhage. The competition of volume and osmocon-trol. Pfluigers Arch. ges. Physiol. 1965, 282, 313.

22. Gottschalk, C. W. Micropuncture studies of tubularfunction in the mammalian kidney. Physiologist1961, 4, 35.

188