THE YALE JOURNAL OF BIOLOGY AND MEDICINE 53 (1980), 133-148

Tamm-Horsfall Protein Antibody In Patients withEnd-Stage Kidney Disease

JEFFREY WORK, M.D., AND VINCENT T. ANDRIOLE, M.D.

Department of Internal Medicine, Yale University School of Medicine,New Haven, Connecticut

Received November 16, 1979

Circulating antibody to Tamm-Horsfall protein (THP) was measured using a radioimmu-noassay in forty-five patients on maintenance hemodialysis and compared to levels of antibodytiters measured in sera from ten healthy controls. The etiology of the end-stage kidney disease inthe patient population was polycystic kidney disease in thirteen, glomerulonephritis in fourteen,diabetic nephropathy in nine, interstitial nephritis and chronic pyelonephritis in three each,multiple myeloma in two, and urinary tract obstruction in one. Four patients had significantlyelevated titers of antibody to THP but shared no other unifying characteristics. The results alsoindicate that none of the groups studied had mean antibody titers significantly different fromcontrols. Furthermore, no general trend was apparent between levels of antibody to THP andnumber of months on dialysis.

Observations made during the study revealed that heparinized samples of blood had lowertiters of antibody to THP than did non-heparinized samples from the same patient. This findingwas repeated when other anti-coagulants, i.e., ethylenediaminetetraacetate (EDTA) and sodiumcitrate, were used. Titers returned toward normal when CaC12 was added back to samples anti-coagulated with EDTA and sodium citrate. This suggests that clotting factors, probablyfibrinogen, interfered with the measurement of antibody titers. Therefore, only serum should beused in further investigations of THP antibody using this assay.

INTRODUCTION

Since its discovery and characterization by Tamm and Horsfall in 1950 [1], theimportance of the 80,000 molecular weight urinary glycoprotein, known as Tamm-Horsfall protein (THP), remains unknown. Only the cells lining the ascending loop ofHenle and distal convoluted tubule synthesize this protein, which is secreted into thetubular lumen [2,3]. No physiological role for the protein is known although THP isknown to inhibit hemagglutination caused by influenza, mumps, and Newcastledisease virus [1], but has no effect on bacteria [4]. Approximately 50 mg of THP isexcreted per day in human urine. In the presence of low pH or high salt concentrationin the urine, the protein aggregates and forms casts [4-6].

Recent work suggests that THP may play a role in some pathological processesaffecting the kidney including calculi formation [7] and other precipitates which maycause microscopic tubular obstruction [8]. Also THP may be involved in thepathogenesis of some interstitial diseases of the kidney. Intrarenal deposits of THPhave been described in medullary cystic disease [9], chronic pyelonephritis, andhydronephrosis [9,10] as well as in end-stage reflux nephropathy. Furthermore, ratsimmunized with homologous THP develop an immune complex tubulo-interstitial

133Address reprint requests to: Vincent T. Andriole, M.D., Department of Internal Medicine, Yale

University School of Medicine, New Haven, CT 06510Copyright © 1980 by the Yale Journal of Biology and Medicine, Inc.All rights of reproduction in any form reserved.

WORK AND ANDRIOLE

nephritis characterized by mononuclear cell infiltrates and immune globulin deposi-tion at the tubular location where THP is synthesized [11]. Nevertheless, the role ofTHP in the pathogenesis of these diseases has not been clearly defined.

Previous work by others in humans and experimental animals has shown thatsevere urinary tract obstruction and vesicoureteral reflux may cause reflux of urineinto the venous and lymphatic circulation as well as into the renal parenchyma[12-14]. Consequently, the potential development of a circulating antibody responseto substances found only in the urine, such as THP, might permit early non-invasivediagnosis of these conditions [12]. In fact, abnormal circulating antibodies to THPhave been detected in experimental animals with vesicoureteral reflux [12], inschoolage girls with acute symptomatic upper urinary tract infection [15], and, byearlier work from this laboratory, in adults with acute urinary tract obstruction andinfection [16]. Although the role of circulating antibodies to THP in the pathogenesisof these acute diseases remains unclear, severe vesicoureteral reflux and urinary tractobstruction can lead to silent renal destruction and end-stage renal disease in somepatients [16-20]. Except for acute pyelonephritis, urinary tract obstruction, andvesicoureteral reflux, antibody to THP has not been determined previously inpatients with other types of renal involvement, and in particular in patients with end-stage renal disease. Therefore, such a study seemed warranted in order to try to betterunderstand the role, if any, of THP antibody in the pathogenesis of these diseases.The present report describes THP antibody in patients with end-stage renal disease

on maintenance hemodialysis, with diagnoses of urinary tract obstruction, interstitialnephritis, and chronic pyelonephritis, and compares them with patients with othercauses of end-stage renal disease including polycystic kidney disease, glomerulone-phritis, and diabetic nephropathy, as well as healthy control subjects. Specifically,our purpose was to determine whether elevated titers of antibody to THP were alsopresent in patients with chronic renal disease caused by different pathologic pro-cesses, as well as whether antibody titers correlated with specific types of renal injury.

MATERIALS AND METHODS

Patient Selection

A total of forty-five chronic hemodialysis patients at Yale-New Haven Hospitaland the West Haven Veterans Administration Hospital participated in this studyafter giving informed consent. The diagnosis of renal failure in each patient wasbased on clinical, radiological, or histological evidence. In addition, informationabout urinary tract infection (UTI), current and past medications, and length of timeon dialysis was obtained from the patient's hospital record.

Patients were placed into groups based upon the etiology of their renal failure.These groups included: (1) thirteen patients with polycystic kidney disease (PCKD) inwhom the diagnosis was made by intravenous pyelography (IVP) or laparotomy; (2)nine patients with diabetic nephropathy (DN), in whom the diagnosis was made onclinical evidence of diabetes mellitus and sequellae including cataracts and retinopa-thy in eight, and in one a history of diabetes and a biopsy which showed interstitialnephritis; (3) fourteen patients with glomerulonephritis (GN), eleven of whom hadbiopsies, which showed membranoproliferative GN in three, membranous GN intwo, chronic progressive hypocomplementemic GN in one, and chronic GN (CGN) infive, and in the remaining three patients without biopsy, the diagnosis of CGN wasbased on a history of acute GN which progressed to renal failure in one, a history ofbeta hemolytic streptococcal infection followed by hematuria, albuminuria, andprogressive renal failure in another, and a history of skin infections leading to

134

ANTIBODY AND END-STAGE RENAL DISEASE

progressive renal failure in the third; (4) six patients with the diagnosis of chronicpyelonephritis (CPN) or interstitial nephritis (IN), based on a history of phenacetinabuse in one, neomycin-induced deafness and renal failure in one, biopsy in two, andby renal scan and pyelography in the other two; (5) one patient with urinary tractobstruction (UTO) diagnosed by IVP; and (6) two patients with multiple myeloma(secretory type) who developed renal failure with progression of their disease.

Blood Collection

All patients received heparin during dialysis with an initial bolus of 5000 Ufollowed by three hourly 1000 U doses until dialysis ended after four hours. Heparinantagonists were not given to any of the patients.

Blood was obtained from each patient for determination of anti-THP antibodytiters. In order to evaluate the effect of dialysis on titers of anti-THP antibody, 5 mlsamples of blood were obtained from participants pre- and post-dialysis. The samplesfrom Yale-New Haven Hospital were collected in plain vacutainers and allowed toclot at room temperature for one hour before serum was removed. The samples fromthe West Haven Veterans Administration Hospital were collected in heparinizedvacutainers (approximately 120U Hep) and the plasma removed immediately.Sodium azide was added to each specimen for bacteriostasis.

Urine

Clean catch urine specimens were obtained for culture from all patients who werenot anuric. Quantitative urine cultures were performed by the standard pour platetechnique. Bacteria recovered were identified by the scheme of Schaub and Foley[21].

THP Antibody

Serum antibody titers to THP were determined by a radioimmunoassay initiallydeveloped in this laboratory and described in detail in previous publications [16,22].The assay depends on binding specific antibody to a solid phase antigen (THPadsorbed to a microtiter plate). The subsequent identification of bound antibody wasachieved by adding 125I labeled protein A which binds to the Fc region of IgG [23].

Serial dilutions of rabbit sera containing known amounts of specific antibody tohuman THP were assayed simultaneously with the patient samples to obtain astandard curve for conversion of counts per minute to ng antibody/ml.

Heparin Effect

Preliminary results indicated that the heparinized samples consistently containedlower anti-THP titers than unheparinized samples. Further studies were done toevaluate the effect of heparin on the assay.

Heparinized and non-heparinized samples from medical student volunteers wereobtained to evaluate the effect of heparin on freshly collected samples. The non-heparinized samples were allowed to clot for one hour and the serum was thenremoved. The heparinized samples were immediately centrifuged and plasma re-moved. All samples received sodium azide as a preservative.To evaluate the effect heparin might have directly on antigen-antibody binding in

the assay, heparin, in amounts approximately equal to 5, 10, 50, and 100 times thecalculated concentration of heparin in patients' serum immediately following dialysis,was added to serum from a rabbit immunized to human THP. The amount ofheparin added was calculated by approximating the heparin T1/2 in serum at 60

135

WORK AND ANDRIOLE

minutes [24] for the amounts used. Samples containing rabbit serum and heparinwere then evaluated for anti-THP activity and compared to serum from the samerabbit to which no heparin was added.

In order to determine whether heparin exerted a specific effect, or if the inhibitionobserved was a result of some other plasma component, i.e., fibrinogen, bloodsamples anticoagulated with ethylenediaminetetraacetate (EDTA) and sodium citratewere compared to samples which were heparinized and to samples which were notanticoagulated. Four samples of blood were taken from each of ten volunteers; onesample without anticoagulant, one with heparin, one with EDTA, and one withsodium citrate. As a further study to evaluate the effect of fibrinogen on the assay,samples collected with EDTA and citrate were defibrinated by adding back CaCl2.This was done by incubating samples and CaCl2 (final concentration in samples0.025M) for one hour at 370C and then at 40C for twelve hours [25]. The resultingclot was removed and the samples assayed.

Statistical Methods

Significantly elevated titers were defined as antibody titers greater than the meanlevel plus two standard deviations observed in healthy control subjects [16]. Inde-pendent groups were compared using T statistics for independent data, heparin vs.serum comparisons were made using the paired T test.

RESULTS

THP Antibody

The titer of Tamm-Horsfall antibody in the serum of patients with end-stage renaldisease and in healthy volunteers is shown in Fig. 1 and is expressed in cpm as well asng/ ml of antibody. The mean anti-THP antibody titer in ten healthy subjects was1050 ng/ ml and the mean plus two standard deviations was 1920 ng/ ml. Any titer

141

K10

i. x

J_, 0

x j

I

DN IN(x) MM(*)CPN(o) UTO(x)

9 6 3

CONTROLS

10

FIG. 1. Measurement of anti-body to Tamm-Horsfall Protein

3260 (THP) in serum from patientswith various types of kidney dis-

2470>ease and control subjects. Left2470 ordinate: cpm x 103 are actual

>- counts in a 1:8 dilution of serum0

1720 ° and bar graphs represent onep standard deviation. Right ordi-z< nate: ng/ml of antibody calcu-

1040 a. lated from a standard curveI- which is not linear. Dotted barF (control) represents two standard

430 , deviations. PCKD = polycystickidney disease; DN = diabeticnephropathy; GN = glomerulo-nephritis; CPN = chronic pyelo-nephritis; IN = interstitial nephri-tis; MM = multiple myeloma;UTO = urinary tract obstruction.

l._0

0

cl 6a.)

4

2

.

.0l00%

PCKD GN

n= 13 14

136

I12-

ANTIBODY AND END-STAGE RENAL DISEASE

above this value was considered significantly elevated. The mean anti-THP antibodytiter in sera from the thirteen patients with PCKD was 721 * 341 ng/ ml and none wassignificantly elevated. The mean value of anti-THP antibody in sera from fourteenpatients with GN was 1348 1503 ng/ ml, which included two patients with signifi-cantly elevated titers of 5132 and 4437 ng/ ml. The mean titer in sera from ninepatients with DN was 1328 ± 1161 ng/ ml and included one patient with a titer of4329ng/ ml. The mean antibody titer in sera from the three patients with CPN was1485 ± 338 ng/ ml and none was significantly elevated. The mean titer in sera fromn thethree patients with IN was 1591 ± 506 ng/ ml and included a significantly elevatedtiter of 2045 ng/ ml in one patient. The mean titer in sera from the two patients withmultiple myeloma (MM) was 340 ±37 ng/ ml and was the lowest for any singlediagnostic group. The serum from the one patient with UTO gave a titer of 1090ng/ml.The mean values for anti-THP antibody for each of the groups were not

significantly different from the mean value for healthy subjects. The highest mean of1591 ng/ ml was seen in patients with IN, followed by 1485 ng/ ml in patients withCPN, 1348 ng/ml in GN, 1328 ng/ml in DN, 1090 ng/ml in UTO, 721 ng/ml inPCKD, and 340 ng/ ml in MM. Urine cultures were obtained from twenty-threepatients, twelve of whom had significant bacteriuria. Urine cultures could not beobtained from twenty-two patients who were anuric. Sera from the four patients withE. coli bacteriuria had a mean titer of 1640 ± 1280 ng/ ml of anti-THP antibody. Serafrom the eight other patients with infected urine had a mean anti-THP antibody titerof 876 i 257 ng/ ml. Sera from the eleven patients with sterile urine had a mean titerof 850 i 376 ng/ ml and the twenty-two patients who were anuric had a mean titer of1180 775 ng/ ml. Individual results appear in Fig. 2. None of the groups of patientswas significantly different from the healthy subjects. Of the four patients withsignificantly elevated titers, one had an E. coli UTI and the other three were anuric.

14 -

12 _

1, 0 _ 32600

Eo _ -- . 24700

* ~~~~~~TI7O6 -- 1720 a

I 8 . z FIG. 2. Measurement of anti-4 _ . _% _"7

- 1040o X body to Tamm-Horsfall Protein(THP) in cpm in a 1:8 dilution of

p i.i } F ~~~~~~~serumand calculated ng/ml of2 _ F .; & ~ 430 c antibody in serum from patients

.L * * .with end-stage kidney diseasewho had significant bacteriuria

i O N I (>105 organism/ ml), sterileEcoli OTHER NO ANURIC CONTROL urine, or anuria, as compared to

GROWTHn: 4 8 11 22 10 abacteriuric controls. Left and

right ordinates as in Fig. 1.

137

WORK AND ANDRIOLE

The chemical data and antibody levels to Tamm-Horsfall protein in these forty-fivepatients with end-stage renal disease are shown in Table 1.The relationship of each patient's serum anti-THP antibody titer to the number of

months on dialysis at the time of this study is shown in Fig. 3. Figures 4-7 show thesame function for each diagnostic group, PCKD (Fig. 4); DN (Fig. 5); GN (Fig. 6);CPN (Fig. 7) and IN (Fig. 7). Significantly elevated titers were observed in 1/ 11patients dialyzed for twenty months or less and in 3/14 patients dialyzed betweentwenty and forty months. None of the other patients dialyzed for forty months orlonger had elevated titers.

Effect of Heparin

Pre-dialysis samples of blood obtained from patients at Yale-New Haven Hospitalwere considered free of heparin effect since blood was obtained 48 hours afterprevious dialysis and heparinization. Pre-dialysis samples (non-heparinized at Yale-New Haven Hospital, heparinized at the West Haven Veterans AdministrationHospital) and post-dialysis samples (heparinized during dialysis at both hospitals)were compared by hospital. There is a suggestion of significant differences betweenmeans of pre- (1237 ng/ml) and post- (1081 ng/ml) samples at Yale-New HavenHospital (p = .03) while there is no significant difference between pre- (637 ng/ ml)and post- (666 ng/ ml) samples at the West Haven Veterans Administration Hospital(p = 0.64).

Figure 8 presents results of serum and plasma anti-THP titers in ten volunteersexpressed in cpm. Each volunteer gave two samples of blood, one heparinized andone not heparinized. The mean antibody titer for non-heparinized samples was 7530cpm ±4040 cpm, and the mean antibody titer in heparinized samples was 5590cpm i 2780 cpm (Table 2). These means are significantly different (p = 0.003).

4

12

10_0

x

0 8

0~~~~~o a2 6 _ 0

2 %""^ *S FIG. 3. Measurement of anti-@ *@ body to Tamm-Horsfall Protein

(THP) in cpm in a 1:8 dilution ofI I ffi I I serum from patients with end-

20 40 60 80 100 120 stage kidney disease as a functionMONTHS ON DIALYSIS of number of months on dialysis.

138

ANTIBODY AND END-STAGE RENAL DISEASE

t n. N N'IT W0 N 01 Iw

C) o) U)C1 C)O

C:COn m O en 00 oo 00 x.0 w IO-Itw m O O

C4CC-~C-~ IF C C) . C)~

Ct l- l-C) en l- O.2 m

en00 W 00 r 00 ON 0e

O~N N Cl C 0 Cl 00 0%S

- Cl eY Ce Yr %0 N0 0% 0

INit00 -

C:)r

C) Co

o* £*

LL CZ CO

-000%7-+

ooN00

C)u

W) 00

ON (a)

ON 'n 'n^

0o 0

00N 0

rn WII II

C )u

m os N 2

0

_0 N N O NCl - N- l-Cl oo

000WI0 0 (~ 0

dp~ f. f"- 0%_ .-

_ C> a O

C) C) *_ C) C)C)C)C -0n In

WI as '-i 0 ON en

CO LC.- cO CO C CO

ClN_l %O_

ir_

en I- WI -)

+

NCl m oo0 \0 'CJ

0n o- 000a

s0e~

I_ _ _ _ _ _ _0 r

zDzzz00e_ _ f_ f)

139

C)

CO

)

C)

C)

CO

0

mCO

C)

0

D

CO

.0

o CKC)._

.0

U= 2

r- C

o ._

.O

E0.

0

0.

C)4

ZC)

00_- to_d o

CL

.4

WORK AND ANDRIOLE

n Q0mo Xl

o-0 0 C7

0O

= ='{-it~c1*000C)00 tn1~

000NO0-a- - -enIT- en 0) O

o0

o

slo a%Nr-0%. N- soQ(NI - t0 o(NIC1

1) *Q .~ . *~ o

COS CO CO3 CO CO

so0 ^ - 4 rONN - + _+ ++

i- o -%

(710%I0 1.(NI t 00

00Qo o o o o o_ en en N WI

e WI N 00

C4 . . e

4)Co c

~ ~ C)C)~ t 0

O C)-Cl) 0 CO cs7 CO

d- - _+ +

ONN- 00

- 0%'

E

= 0Fo en WI_

00 _% It 00 _CO.

enI WI

-n en-

°° ° ° F1

140

00 -

en -

,t C74 11 11

_^ II

E

O *C.0 > _*- Cd Ei< 0

C

F_ -

CO0

OD u

0.C)

0.C?C)4

4.)z

..j .z

*04)C.0

C)0 C

CO 0

+

N N

X~N

0. 0

e00£ 'l eII II

2 C

ANTIBODY AND END-STAGE RENAL DISEASE

lq.00 q

0O( II IICIT-~~ -Z

4))')

.C)=I.

.

r- U)

U)

0

4)w0

4.

c.

erIerI" I4 C1

8 e ,oD 8 8 t un

0N II IIo frCRg ,l,tn lt W) Ce d

4)U)

0 o

0000iC-4

4) U)

s0 elt It 1- (0en en II 11 0-r-

0%v

E

4.) U).0

4)-4)~

- ... -

8 C 4)

O ~~~~~~~~~I I

et

0

4C)to toI(a _

0.-,~. _ _

c, :;0'It

141

~000

U)

(NI (ON 00'I q R

0

4) 4) .

U)U C4

4)00.

0% s

08

C>' 0%§ 4 -

00 'n 00 11

om̂ on CX

N- oo- o~ o = 08o l

(L(ntA

0 _

t 00 alen en en

0

r-

,

0)

v4).43.E

0

U*

0~kZ

I--, Z. I

:. >F .z>1

lq:. C> (7,tn "grI.-, iC14 en"It :z "Izr

WORK AND ANDRIOLE

PCKD.

*S * FIG. 4. Measurement of* ** antibody to Tamm-Horsfall

* Protein (THP) in cpm in a1:8 dilution of serum frompatients with polycystic kid-

I I I ney disease (PCKD) as a20 40 60 80 100 120 function of number of

MONTHS ON DIALYSIS months on dialysis.

The effect of adding heparin to hyperimmune rabbit serum is shown in Fig. 9. Thefinal concentration of added heparin was 5, 10, 50, and 100 times the calculatedconcentration of heparin in patients' serum at the end of dialysis. THP antibodylevels remained the same in hyperimmune rabbit serum regardless of the concentra-tion of added heparin, which suggests that the heparin molecule itself does not affectthe method for measuring antibody.

Effect of Other Anticoagulants

The effect of other anticoagulants on the level of Tamm-Horsfall antibody inserum was studied in ten healthy subjects, each of whom contributed four freshlydrawn samples of blood. Either heparin, sodium citrate, EDTA, or no anticoagulant(control) was added to one of the samples from each subject, after which the level ofantibody was determined. Calcium chloride was then added to each sample and theantibody level was measured again. The results are presented in Table 3 and show a

12 -D N

10 k

8

6

4 .0

0

2 _

20 40 60 80 100MONTHS ON DIALYSIS

FIG. 5. Measurement of anti-body to Tamm-Horsfall Protein(THP) in cpm in a 1:8 dilution ofserum from patients with diabetic

120 nephropathy (DN) as a functionof number of months on dialysis.

to0xCMz

0co

a- 2

0x

zM0

a.C.

142

6%F

4 e

ANTIBODY AND END-STAGE RENAL DISEASE

0

GN

000

0

me 0 FIG. 6. Measurement of anti-0 body to Tamm-Horsfall Protein

(THP)in cpm in a 1:8 dilution of20 _

0 120 serum from patients with glomer-20 40 60 80 100 120 ulonephritis (GN) as a function

MONTHS ON DIALYSIS of number of months on dialysis.

significant difference in measured antibody titers between control serum (non-anticoagulated) and sodium citrated blood (p = 0.003), between control serum andblood anticoagulated with EDTA (p = 0.004), as well as between control serum andheparinized blood (p = 0.01). All three anticoagulants significantly reduced the titerof measured antibody. However, the addition of calcium chloride to blood anticoag-ulated with sodium citrate resulted in a return of measured antibody titer towardnormal. Although the antibody titer in citrated blood to which calcium chloride hadbeen added was lower than that observed in control serum with (p = 0.01) or without(p = 0.02) calcium chloride, it was significantly higher than that observed in citrated

8 p .

0

I NCPN

0 .

FIG. 7. Measurement ofantibody to Tamm-HorsfallProtein (THP) in cpm in a1:8 dilution of serum frompatients with chronic pyelo-

I I nephritis (CPN) and inter-

20 40 60 80 100 120 stitial nephritis (IN) as afunction of number of

MONTHS ON DIALYSIS months on dialysis.

14 F

0

12

10 _

C)0

0zM0

a-C.)

8 _

6 k

4

I _0x6az

0 4)

2 -

143

2 [4

WORK AND ANDRIOLE

NON-HEPARI Nn= 10

FIG. 8. Measurement of antibody to Tamm-HorsfallProtein (THP) in cpm in a 1:8 dilution of serum in non-

HEPARIN heparinized and heparinized samples from volunteers. Bar10 graphs represent one standard deviation.

TABLE 2Levels of Antibody to Tamm-Horsfall Protein (THP) in Heparinized and Non-Heparinized Serum

Samples Collected from Healthy Subjects

THP Antibody in cpm*

Subject Non-Heparinized Heparinized

KB 5,140 3,270 -36GC 4,340 4,290 -IBF 2,800 2,490 -11ML 11,090 6,430 -42BS 5,320 4,730 -11JW 5,180 4,930 -5DB 9,970 7,400 -25JB 16,310 12,280 -25JK 6,510 3,980 -39BS 8,660 6,030 -30

mean = 7,530 mean = 5,590s.d. = 4,040 s.d. = 2,780

*cpm = counts per minute

144

16 -

14 -

120

0I')

0x

z

00m

10O

8 _

6 _

0

0

0o

0

0

00

00

0

0

4

2

ANTIBODY AND END-STAGE RENAL DISEASE

TABLE 3The Effect of Three Anticoagulants With and Without Calcium Chloride on Tamm-Horsfall Protein

Antibody Levels in Healthy Subjects

Control Citrate EDTA HeparinSubject Control Citrate EDTA Heparin +CaCl2 +CaC12 +CaCl2 +CaCl2

AM 3570* 2500 2280 2230 2380 2100 2340 1860SD 2630 2230 2610 2400 2950 2135 2890 2120TM 2710 2334 2400 2250 3150 2820 3070 2590BC 3620 2590 2820 3090 3200 3190 3170 2820RD 1830 1430 1510 1780 2130 1970 1990 1560JS 2950 2300 2330 2660 3020 2560 2680 2410SD 5120 3670 4300 3770 3770 3370 4130 4070JW 3180 2060 2380 3090 3290 3360 4050 3070SS 6290 4070 5310 4180 5200 4530 6480 5310DB 7020 4040 4560 4230 5390 5410 4270 4120

MEAN = 3890 2720 3050 2970 3450 3140 3510 2990S.D. = 1690 894 1230 860 1080 1110 1300 1180

*Counts per minute of I125

blood without calcium chloride (p = 0.05). Similar results were observed whencalcium chloride was added to blood anticoagulated with EDTA. In fact, there wasno significant difference in antibody titer between control serum with (p = 0.78) orwithout (p = 0.27) calcium chloride and EDTA anticoagulated blood to whichcalcium chloride had been added, whereas there was a significantly higher antibodytiter in EDTA plus calcium chloride treated blood as compared with blood anticoag-ulated with EDTA alone (p = 0.04). In contrast, the addition of calcium chloride toblood anticoagulated with heparin did not change the measured antibody titer(p = 0.87). Specifically, antibody titers remained significantly lower in heparinizedblood to which calcium chloride was added as compared with control blood with(p = 0.01) or without (p = 0.01) calcium chloride (Table 3).

70.-

60

50 F

10

FIG. 9. Measurement of anti-body to Tamm-Horsfall Protein(THP) in cpm in a 1:8 dilution ofimmune rabbit serum with hep-arin added to make final con-

centrations 1, 5, 10, 50, and 100times the concentration of hep-

I lo arin in the serum of patients atIX 5x lOx 50x IOOx completion of dialysis.

0x

0zD0

0-aU

145

WORK AND ANDRIOLE

DISCUSSION

Previous attempts to identify antibody to kidney tissue in patients with pyelone-phritis and chronic nephritis using complement fixation [26], latex agglutination [27],and hemagglutination [28] have been unsuccessful. However, those assays were lesssensitive than the present one and could not identify circulating antibody in normalcontrols. Using the present assay, Marier et al. [16] demonstrated small amounts ofantibody to THP in normal subjects and significantly elevated titers in patients withacute urinary tract obstruction and infection. Hanson et al. [15] also observedelevated titers of antibody to THP in 9/10 school-age girls with acute symptomaticupper urinary tract infection using an enzyme linked immuno-absorbent assay.

The objective, in the present study, was to determine, using our new assay, whetherelevated titers of antibody to THP were present in patients with various forms of end-stage kidney disease. The results of the present study indicate that four of the forty-five patients with chronic renal disease had significantly elevated titers of antibody toTHP, but they did not share any other unifying characteristic. Two patients withglomerulonephritis dialyzed for three and twenty-six months, respectively, hadelevated titers, one of whom also had anE cli urinary tract infection, though theother was anuric. One patient with diabetic nephropathy dialyzed for thirty-onemonths, had elevated titers and was anuric, and another anuric patient with elevatedtiters and also dialyzed for thirty-one months, had kidney failure based on long-standing interstitial nephritis. The present work indicates that none of the groupsstudied had mean antibody titers significantly different from controls. Furthermore,in all groups of patients studied, no general trend was apparent between levels ofantibody to THP and number of months on dialysis.The present data do not exclude the possibility of elevated antibody titers at some

point in the progression of various types of kidney disease, but this study indicatesthat titers of antibody to THP are not useful in diagnosing the cause or presence ofend-stage kidney disease. The patients reported in the present study have end-stagerenal disease and have been on dialysis for periods of one month to ten years. Thefailure to detect elevated antibody titers in many of these patients might have beenexpected since these patients were not likely to have active or progressive renaldisease at the time of study, and elevated antibody titers may be dependent on activekidney disease. If this concept is true, then our observations would suggest thatfurther study to identify patients with elevated titers of antibody should involvepatients with active kidney disease. Such studies should include sequential determina-tions to better define the time course of this antibody response.

During the course of these studies we observed a significant difference in antibodytiters between pre- and post-dialysis samples of blood obtained from patients at theYale-New Haven Hospital but not in samples obtained from patients at the WestHaven Veterans Administration Hospital. The pre-dialysis samples (serum) frompatients at Yale-New Haven Hospital were not anti-coagulated, while the samples(plasma) from patients at the West Haven Veterans Administration Hospital were

anti-coagulated with heparin. Post-dialysis samples (plasma) at both hospitals were

collected identically and both were heparinized. These results suggest two importantfindings: first, heparin interferes in some way with the quantitative measurement ofcirculating antibody to THP, which explains the difference noted between pre

(serum) and post (plasma) dialysis samples obtained from patients at Yale-NewHaven Hospital; and second, dialysis has no effect on the quantity of circulatingantibody since pre- and post-dialysis samples (plasma) obtained from patients at theWest Haven Veterans Administration Hospital were the same. Further studies were

146

ANTIBODY AND END-STAGE RENAL DISEASE 147

performed to attempt to clarify the effect of heparin on the assay. Specifically,heparin was added directly to hyperimmune serum obtained from rabbits immunizedwith human THP and no changes in antibody titers were noted. These resultssuggested that the heparin molecule, which exists as a heterogenous assortment ofpolymeric chains [29], did not directly interfere with the quantification of antibody toTHP when that antibody was in serum obtained from clotted blood. That is, heparindid not appear to interfere directly with antigen-antibody binding but instead seemedto act through or because of some substance present in plasma and not present inserum, such as fibrinogen or heparin activated anti-thrombin III. Additionalexperiments with other anti-coagulants, i.e., EDTA and sodium citrate, furthershowed that the inhibition on lowered levels of measured antibody to THP observedin heparinized -samples (plasma) also occurred when other anti-coagulants were usedto collect the blood and the assay was performed with the plasma obtained from thesesamples. Antibody levels in samples of blood collected with these two chelators ofcalcium returned near to control values when fibrinogen was removed by the additionof CaCl2 to the samples, but did not return toward normal when CaCl2 was added toheparinized samples. These results suggested that anti-coagulants interfere indirectlywith the quantification of circulating antibody to THP and that clotting factors areprobably the source of this interference. Since calcium chelators halt the clottingcascade by preventing conversion of prothrombin to thrombin (a calcium-dependentstep), which also prevents conversion of fibrinogen to fibrin, and heparin also stopsthe clotting cascade by inhibiting certain activated factors: XIa [30], IXa [31], Xa[32], and thrombin [33], and since the defect could be corrected by the addition ofCaCl2 to EDTA and citrated but not heparinized samples, the protein most likely tointerfere with this assay appears to be fibrinogen. Fibrinogen is the most plentiful ofthe coagulation proteins and is preserved when all the above anti-coagulants areused. These results indicate that anti-coagulated blood should be avoided in futureuse of this assay to determine THP antibody levels, in order to prevent obtainingspuriously low levels of THP antibody.

ACKNOWLEDGEMENTS

The authors wish to thank Martha Jansen and Susan Marino for technical and secretarial assistance.

REFERENCES

1. Tamm I, Horsfall FL Jr: Characterization and separation of an inhibitor of viral haemagglutinationpresent in urine. Proc Soc Exp Biol Med 74:108-114, 1950

2. Schrenk EA, Schwartz RH, Lewis RA: Tamm-Horsfall protein. 1. Localization in the kidney. LabInvest 25:92-95, 1971

3. Pollack VE, Arbel C: The distribution of Tamm-Horsfall mucoprotein (uromucoid) in the humannephron. Nephron 6:667-672, 1969

4. McQueen EG, Engel GB: Factors determining the aggregation of urinary mucoprotein. J Clin Path19:392, 1966

5. McKenzie JK, Patel T, McQueen EG: The excretion rate of Tamm-Horsfall protein in normals and inpatients with renal disease. Aust Ann Med 13:32, 1964

6. McQueen EG: The nature of urinary casts. J Clin Path 15:367, 19627. Grant AMS, Baker LRI, Neuberger A: Urinary Tamm-Horsfall protein in certain kidney disease and

its content in renal and bladder calculi. Clin Sci 44:377-384, 19738. Patel R, McKenzie JK, McQueen EG: Tamm-Horsfall urinary mucoprotein and tubular obstruction

by casts in acute renal failure. Lancet 1:457, 19649. Vernier RL, Resnick J: Medullary cystic disease: the possible role of Tamm-Horsfall protein. Kidney

Inter 9:450, 197610. Zager RA, Cotran RS, Hoyer JR: Pathologic localization of Tamm-Horsfall protein in interstitial

deposits in renal disease. Lab Invest 38:52, 1978

148 WORK AND ANDRIOLE

11. Hoyer JR: Autoimmune tubulointerstitial nephritis induced in rats by immunization with rat Tamm-Horsfall protein. In Abstracts of the Ninth Annual Meeting of the American Society of Nephrology.Thorofare, NJ, Charles B Slack, 1976, p 58

12. Hodson CJ, Maling TMJ, McManamon PJ, et al: Reflux nephropathy. Kidney Inter 8:S50-S58, 197513. Rolleston GL, Maling TMJ, Hodson CJ: Intrarenal reflux and the scarred kidney. Arch Dis Child

49:531-539, 197414. Bell RD, Ornitz RD, Trautman R, et al: The significance of the renal pelvis and intrarenal veins in

renal lymph formation. Invest Urol 9:149, 197115. Hanson LA, Fasth A, Jodal U: Auto-antibodies to Tamm-Horsfall protein, a tool for diagnosing the

level of urinary tract infections. Lancet 1:226, 197616. Marier R, Fong E, Jansen M, et al: Tamm-Horsfall protein antibody in urinary tract obstruction and

vesicoureteral reflux. J Infect Dis 138:781-790, 197817. Bakshandeh K, Lynne C, Carrion H: Vesicoureteral reflux and end stage renal disease. J Urol 116:557,

197618. Hodson CJ, Edwards D: Chronic pyelonephritis and vesicoureteral reflux. Clin Radiol 11:219-231,

196019. Salvatiena 0 Jr, Kountz SL, Belzer FD: Primary vesicoureteral reflux and end stage renal disease.

JAMA 226:1454, 197320. Lenagham D, Whitaker JG, Jensen F, et al: The natural history of reflux and long term effects of

reflux on the kidney. J Urol 115:728, 197621. Schaub IG, Foley MK, Scott CG, et al: Diagnostic Bacteriology. 5th Edition. St. Louis, Mosby, 195822. Marier R, Jansen M, Andriole VT: A new method for measuring antibody using radiolabeled Protein

A. Clin Res 26:518A, 197823. Marier R, Jansen M, Andriole VT: A new method for measuring antibody using radiolabeled Protein

A. J Immunol Methods, in press24. Olsson P, Lagergren H, Ed S: The elimination from plasma of intravenous heparin: an experimental

study on dogs and humans. Acta Med Scand 173:619-630, 196325. Pitney WR, Dacie JV: A simple method of studying the generation of thrombin in recalcified plasma.

J Clin Path 6:9, 195326. Mackay IR, Larkin L: The significance of the presence in human serum of complement fixing

antibodies to human tissue antigens. Aust Ann Med 7:251-258, 195827. Kramer NC, Watt MF, Howe JH, et al: Circulating antihuman kidney antibodies in human renal

disease. Am J Med 30:39-45, 196128. Kalmanson GU, Glassock RJ, Harwick HJ, et al: Cellular immunity in experimental pyelonephritis.

Kidney Inter 8:S35-S43, 197529. Jaques LB, McDuffie NM: The chemical and anti-coagulant nature of heparin. Semin Thromb Hemo

4:277, 197830. Amir T, Ratnoff DD, Pensky J: Partial purification and some properties of a plasma inhibitor of

activated plasma thromboplastin antecedent factor. J Lab Clin Med 80:786, 197231. Rosenberg JS, McKenna PW, Rosenberg RD: Inhibition of human factor IX. by human antithrom-

bin. J Biol Chem 250:883-888, 197532. Rosenberg RD: Function of heparin. In Heparin Chemistry and Chemical Usage. Edited by VV

Kakker, DP Thomas. New York, Academic Press, 1976, pp 101-12033. Rosenberg RD, Damus PS: The purification and mechanism of action of human anti-thrombin-

heparin cofactor. J Biol Chem 248:6490, 1973