A simplified technique to produce toxemia in the pregnant dog

A simplified technique to produce toxemia in the pregnant dog M.

MAURICE

ABITBOL,

M.D.

Jamaica, Neul York In 39 pregnant and six nonpregnant dogs the abdominal aorta was exposed extraperitoneally through a 1eft paralumbar incision, and a vascular occluder and a flow transducer were placed around it. The tubing of the occluder and the wiring of the flow transducer were slipped under the skin and advanced to the back of the neck of the animal. Thus, the aortic blood flow could be progressively reduced and continuously monitored. After 4 to 20 days of blood flow reduction of 35% to 75%, the picture of a moderately severe toxemia appeared in 10 dogs and that of a mild toxemia was seen in the 11 others. Nine animals were eliminated because of premature labor and three were eliminated for surgical complications. The control group (six pregnant animals with a sham operation and six nonpregnant animals with a severe aortic constriction) presented no abnormalities. (Au. J. OBSTET. GYNECOL. 139:526, 1961.)

PREECLAMPSIA appears specific to humans and does not occur spontaneously in any animal species during pregnancy.’ Because decreased uterine perfusion is a characteristic of toxemia, researchers have attempted to reproduce this syndrome in the pregnant animal’-’ by chronically reducing uterine blood flow, in some instances prior to conception4. 5 and in others during pregnancy.3 Success has been rather limited as in most studies abortion or premature delivery immediately follows surgical manipulation. We previously reported on the successful production of experimental toxemia in several animal species. In these studies, constriction of the abdominal aorta was used to compromise uterine flo~.~-‘~ The syndrome we described consisted of hypertension, proteinuria, and weight gain. Light and electron microscopic studies demonstrated endothelial swelling, mesangial proliferation, subendothelial deposits, and “fusion” of the foot processes. lmmunofluorescence was indicative of deposits of fibrinogen or fibrin-related antigens in the From the Department of Obstetrics and Gynecology, The Jamaica Hospital. The work w(w done at the IRM Institute of the New York University School of Medicine, New Ywk, Dr. F. E. Birknrr, Director; Mario Uagnaz and Robert Kolwicz, Assistant Research Scientists. Received RezGed Accepted

for publication November November

June 21, 26,

24,

1980.

1980. 1980.

Reprint requests: M. Maurice rlbttbal, M.D., Chairman, Department of Obstetrics and Gynecology, The Jamaica Hospital, 89th Ave. and Van Wyck Expressway, Jamatca, Neur York 11418.

526

loo/

50-

0,

\ 0.5

WATER

I .o

IN ml.

Fig. 1. This diagram shows the relation between the amount of water pushed with a 1 ml tuberculin syringe into a VO-4 8 mm occluder and the aortic blood flow reduction (Dog R-34). For instance, a 40% blood flow reduction requires 0.4 ml of water to be pushed into the occluder. glomeruli and the liver showed foci of necrosis. Numerous infarcts were observed in the placenta, which is important because infarcts are not seen in the placentas of normal animals, whereas they have been observed in varying degrees in the placentas of nontoxemic patients and to a much higher degree in the placentas of toxemic patients. 9- ‘I In addition, localized intravascular coagulation, I2 elevated plasma renin activity-, and elevated aldosterone levels” were demonstrated in 000%9378/81/050526+09$00.90/0

0 1981 The C. V. Moshv

C<,

Volume 139 Number 5

Technique

to produce

toxemia

in pregnant

dog

527

Fig. 2. The wiring of the Row transducer and the tubing of the vascular occluder (both placed around the abdominal aorta of a pregnant dog) are slipped under the skin to the back of the neck of the dog. As the occluder is compressed, the aortic blood flow is reduced, as shown on the flowmeter. Table

I. Control

group* Blood pressure

Dog

No. 1 4 14 17 20 21 33 36 38 39 40 41

Duration constriction

of (days)

Maximum aortic jaw reduction

12 8 16 19 7 12 4 8 4 9 12 7

*The first six were nonpregnant

75 50 60 70 75 60 0 0 0 0 0 0

blood (‘%)

Before surgery

I lOif 105180 135190 120170 134160 110155 110160 120170 1OOl.50 135165 110/50 120160

(mm Hg) Highest value after surgery

i 00150 120170 120180 120170 140170 120150 100150 11 O/65 120170 130160 120160 14OJ85

Proteinuria

Immunojluorescence

0 0 0

0

0 0

0

0 0 0 0 0

0 0 Traces to I+

and had an aortic constriction; the last six were pregnant and had a sham operation.

blood samples from the uterine veins of some toxemic animals but not in the control group. There are many shortcomings in this experimental model. The degree of blood flow reducrion is critical and there is a narrow range between success and failure which is hard to maintain. Severe forms of toxemia were not observed, and with time the reappearance of collateral circulation lessened the disease. Another reason for these shortcomings is that while the blood reduction necessary to produce toxemia seems to be in the range of 40%, it appears that for each animal there

is a specific blood flow reduction which will lead to toxemia, above which fetal death will occur and below which nothing happens.13 This report describes a new technique for producing, controlling, and monitoring aortic blood flow reduction in gravid dogs.

Material and methods A total of 39 pregnant and six nonpregnant dogs were selected for this study. The pregnant ones were approximately within the last 3 weeks of gestation as determined by known date of conception or by abdom-

528

Abitbol

Fig. 3. There is marked enlargement of the mesangial area. Protein droplets in the visceral epithelial layers are obvious at 10 o’clock, and can also be seen at 12 o’clock. Some endothelial thickening can be seen between 10 and 12 o’clock. (Phosphotungstic acid-hematoxylin. X 300.) Table

II. Experimental

toxemia

in 21 dogs Blood pressure

Dog NO.

5 6 9 10 11 12 13 16 18 19 26 29 30 32 34 35 37 42 43 44 45

Duratzon constriction

4 7 8 4 4 iL 12 12 15 16 9 20 4 4 9 -: 7

of (days)

&faximum

pow

oortic blood redurtlon (70)

55 40 50 60 50 60 50 40 75 50 60 65 50 35 60 40 55 75

(mmiHg) Highe.ct

Before

surgery

ajier

value

surgery

120170 llOl80 105180 134160 140160 100150 120170 155180 150/90 110/40 135190

145195 1451110 180/l 15 1701100 1801100 2701170 145195 2201110 220/l 20 140/70 1801120

150190

2401180

150180 100/60 135180 145185 17ot70 140/75 11o/55

1801110 1801100 2201110 1351110 2001100 2201140 16Oi95

120/50

180185

120/70

156/100

inal palpation during anesthesia (pregnancy in the dog lasts 9 weeks). No animals were hypertensive since the initial diastolic pressure was below 90 mm Hg. Among the pregnant group, 12 animals were eliminated from the study because of surgical complications or premature labor. Six had a sham operation and th~rs

Proteinuria

0 0 0 0 0 4+ 1+ Ii3+ 0 Trace to 1t 1+

Traces Traces 3+ 2+ 2+ 0 2t 3+ 0

Immunofluorescence

3+ 3+ 3+ 2-t 2+ 3+ 1+-2+ If 2t-3+

3+ 3+ 2t

3t 3t

2+ 1+ 2t 2t

1 t-2+ 3+

served as controls. The remaining 21 had an aortic constriction lasting 4 to 22 days. Six nonpregnant dogs served as controls. They had an aortic constriction that reduced the blood Row by 50%’ to 7%. which was maintained for 3 to 18 days. The Gould-Statham p&table blood Howmeter (SP

Technique

Fig. 4. In addition to the lesions cytoplasm of proximal tubules hematoxylin. x 300.)

described as seen

in

in Fig. 3, there is deposition the center of the picture.

2202) and flow transducers were used to measure the aortic flow. Zero flow calibration was easily obtained by using the nonocclusive zero feature of this flowmeter. The zero flow was always determined first by producing a complete occlusion of the measured artery. However, reactive hyperemia, which follows occlusion and requires several minutes to subside, frequently precluded performance of this maneuver. With the nonocclusive zero feature. the apparatus can be realigned to zero flow at any time without occluding the artery. The values given by the precalibrated transducer and the values obtained bi allowing heparinized blood to flow through a resected arterial segment should not show more than 10% difference. If it did, the How transducer was returned to the factory for recalibration. Vascular occluders, Models VO 3 and VO 4 (Rhodes Medical Instruments), were used to produce a progressive and calibrated constriction of the abdominal aorta. This constriction was produced by water injected with a tuberculin syringe and pushed into polyethylene tubing connected to the vascular occluder (Fig. 1). Arteriosonde 1010 or 1020 was used to measure the blood pressure on the forelimb of the dog. The abdominal aorta in each dog was exposed extraperitoneally through a left paralumbar incision, as previously placed origin

described.+‘” around of the

the abdominal renal arteries.

The

vascular

occluder

aorta, 1 to 2 cm below The polyethylene tube

to produce

was the was

toxemia

of protein droplets (Phosphotungstic

in pregnant

dog

529

in the acid-

delivered through a small incision in the back of the neck of the dog by sliding it under the skin along the vertebral spine (Fig. 2). Similarly, the flow transducer could be placed around the abdominal aorta and its wiring also was delivered to the back of the neck of the dog. The flow transducer is connected to a Rowmeter recording system, which can accurately register the pulsatile and mean blood flow of the abdominal aorta. By forcing water into the polyethylene tubing with a syringe, one can modify the blood flow within the abdominal aorta. Any blood flow variation within the abdominal aorta is integrally transmitted to the uterine arteries, as previously demonstrated.lJ With this simple technique, blood flow to the uterine horns can be modified at will and its variations can easily be recorded. On the first postoperative day the blood flow was reduced by only 35%. Then the blood flow reduction was increased by 5% every day, or every second day, until hypertension, then proteinuria appeared. For dogs near term, the blood flow was reduced over shorter intervals. The litters were not at risk as long as blood

pressure

was

normal

and

one

could

decrease

blood flow reduction by another 5% or 10%. When aortic constriction was excessive, partial paralysis of the hind

legs

Blood

was

noted

pressure

tervals. A catheterized the day of surgery

and was

constriction measured

was daily

urine specimen and every second

decreased.

at frequent was obtained day thereafter.

inon

530

MaKh .4m. J. Obstet.

Abitbol

Fig. 5. Toxemic dog. Electron normal glomerular basement foot processes at the bottom.

Urine

was

qualitatively

of the

sulfosalicylic

urine

specimens

surgery The

were animals

ery,

which

checked acid

occurred

ation

in

were

killed

13 of the

killed before and placenta for histologic taken from

technique.

were included were

for

negative

in this observed within

Only for

eight

examination. kidneys

immunofluorescence cording to techniques

for

shows subendothelial electron-dense deposits between and the enlarged mesangial area. There is focal “fusion”

The

whose prior

to deliv-

the oper-

These other

the liver, buffered

animals dogs

were

kidneys, formalin

In addition, tissues electron microscopic

studies and were processed previously described.Hm’0

were and ac-

pregnant

sented and

after

the ot

Results

means

spontaneous

bitches.

while

by

animals protein

4 to 21 days

delivery. Tissues from were fixed in neutral the

protein

study. until

21 pregnant

immediately,

micrograph membrane (X 15,800.)

1, 1981 Gynecol.

no no

and

proteinuria.

and

revealed studies

The

results were

indicative

toxemia

(Table

II).

Hypertension. hypertension. pressure

in the

21 pregnant of mild

and

1 lo/40

microscopic

the

dogs

mm

with

to moderate

Hg

pre-

hypertension

electron

Almost all of these In some cases it was mild, of

animals

no

no abnormal revealed no

deposits in the kidneys, any infarcts (Table I).

striction

control

control

specifically Light

studies of the kidneys Immunofluorescence fibrin-like not show

nonpregnant

abnormalities,

findings. fibrin or

placenta

did

aortic

con-

to severe

animals had for instance, a

before

the

opera-

Volume Number

Technique to produce toxemia in pregnant dog

139 5

Fig. 6. Toxemic

dog. Electron micrograph shows that the mesangial area contains massive electron-dense deposits. The endothelial cell (EN) is swollen lumen. US: Urinary space. (Original magnification X 18,000.)

rion

increased

surgery.

to 140/70

In other

cases,

blood pressure blood pressure

recorded

control

mm

Hg

on the

it was severe,

fifth and

day the

after

highest

was 280/150 mm Hg with a of 150/90 mm Hg before

surgery.

Proteinvria. Proteinuria ( 1+ to 4+) was observed in 10 dogs. Variations in weight. Weight variations were not recorded in ail the dogs because of technical difficulties, but a weight gain of 7, 5, and 4 pounds within an interval of 10 days was observed in three different dogs. Renal lesions on light microscopy. The main difference between toxemia in the experimental dogs and that

in

the

human

is that

in

the

former

there

was

531

(M) is enlarged and and fills the capillary

marked swelling in the mesangiat area, while in the latter the swelling is mostly seen in the endothelial cells (Figs. 3 and 4). Some additional lesions, also typical of human toxemia, were observed in the experimental dogs, while others, equally typical of human toxemia, were not observed (Table III). More important, negative observations, such as normal basement membrane, absence of ischemic tubular lesions, absence of fibrinoid changes in the media of arteries, and absence of elastica beads, rule out any other renal pathology to account for the clinical picture. Compared to human toxemia, as described by Sheehan and Lynch,14z I5 this couid be called an early atypical picture. Renal lesions on electron microscopy. These lesions

532

Abitbol

March 1. 1981 Am. J. Obstet. Gynecol.

Fig. 7. Toxemic dog. Glomerulus fibrinogen serum. (X 160.)

with diffuse fluorescent

Table III. Renal findings on light microscopy experimental toxemia in the dog as compared human toxemia Abnornzal$ndings

Human toxemia

in to

+++

+

++

0

++

+++

++

+++

+++ +++ + Normal

+ + 0 Normal

++ ++

+ 0

Rare

Absent

Absent

Absent

Tubules:

Protein droplets and fat in epithelium of proximal convoluted tubles Protein casts in lower nephron Artm’es:

Ischemic lesion in tubules Fibrinoid changes in the media of arteries

Table IV. Electron seven experimental

microscopic dogs

findings

Charactetitic

ExQerimental toxemia

Glomeruli:

Endothelial swelling Collagen fibrils in endothelial cells Enlargement of the mesangial areas Collagen fibrils in mesangial areas Visceral epithelial cells: Protein droplets Cytoplasmic swelling . Large vacuoles Basement membrane

deposits after exposure to anti-dog

were more informative and were consistent with those of human toxemia, that is, endothelial cell swelling, subendothelial deposits. enlargement of the mesangial areas with swelling of cells and electron-dense deposits, and occasional “fusion” of foot processes. The glo-

Result

Endothelial cell swelling Subendothelial electronlucent “fluff” Subendothelial electron-dense deposits Mesangial enlargement Circumferential mesangial extensions Mesangial electron-dense deposits Effacement of foot processes Findings in two control

in

No.

+ to +++ +to++

S/7 417

+ t,‘+++ +to++ ++ to +++ +to++

% 317 617 417

animals were essentially normal.

merular basement membrane was always intact (Figs. 5 and 6, Table IV). Immunofluorescence. Immunofluorescence revealed deposits of fibrinogen or fibrin-related antigens in the glomeruli (Fig. 7). Stains for immunoglobulin were negative in all the animals. Placental infarcts. Placental infarcts of different ages were observed in the placentas of these experimental animals (Fig. 8). These 21 toxemic animals were arbitrarily divided in two groups. In the first group were 11 animals which had mild toxemia, that is, some hypertension and occasional traces of proteinuria. Placental infarcts generally were present, and the glomeruli showed fibrinoid deposits on immunofluorescence. However, on electron microscopy there was only occasional endothelial swelling without obvious subendothelial deposits.

Volume Number

Technique to produce toxemia in pregnant dog

139 5

533

Fig. 8. Toxemic dog. Recent placental infarct in the center surrounded by normal placental tissue. The placenta of the dog is epitheliochorial and under normal conditions, has an “endometrioid” appearance. (Hematoxyin aid eosin. x 160.) In the second group were 10 animals which had a severe form of toxemia characterized by proteinuria (1+ to 4+) in catheterized urine specimens, hypertension, numerous placental infarcts, marked glomerular endothelial swelling, and subendothelial deposits on electron microscopy as well as fibrin deposits on immunofluorescence. Positive immunofluorescence and placental infarcts were equally present in both groups. In general, while the highest blood pressures were seen in the second group, there were many exceptions. R42 in the first group had a severe hypertension, and R13 and R43 in the second group had only a mild to moderate hypertension. Evidence of proteinuria, severe endothelial swelling, and subendothelial deposits by electron microscopy were present only with “severe” toxemia.

Comment With this newly developed experimental model, features similar to those of human toxemia were observed more frequently and were more severe in nature than

with the models previously described.8-‘0 Not only was the clinical triad (hypertension, proteinuria, and weight gain) observed in some animals, but the anatomic lesions consistent with toxemia, i.e., endothelial swelling, subendothelial deposits, mesangial enlargement, etc., were present in many animals. The evidence of deposits of fibrinogen and/or its derivatives in the kidneys demonstrated by immunofluorescence was striking. Perhaps the most interesting feature is the presence of placental infarcts. Because placental infarcts do not exist in the normal placenta of the bitch and because it has been previously demonstratedg-I* that the number of placental infarcts and the severity of the experimental toxemia are related to the degree of aortic constriction, one can assume that placental ischemia is the initiating factor in experimental toxemia. The sequence of events could be: aortic constricinfarcts + tion + placental ischemia + placental toxemic manifestations. Investigation of different parameters in the uterine vein blood as compared to those of the uterine artery

534

March 1, 1981

Abitbol

Am. J. Obstet.

or of the peripheral venous blood could be one of the possible future studies in this experimental model. Isolated experimental cases of toxemia have already demonstrated elevated plasma renin activitylO and evidence of localized intravascular coagulationI in the uterine vein blood as compared to the peripheral blood. In this experimental model and with the available equipment, chronic reduction of aortic blood flow in the abdominal aorta was induced in a more precise and consistent manner than with experimental models previously described.“” By continuously monitoring the aortic blood flow and progressively increasing the degree of blood flow reduction, a better relationship was established with the severity of toxemia. However, toxemia was not produced in all the cases. Even if one excludes the 12 cases of premature delivery and surgical complications, severe experimental toxemia, i.e., hypertension, proteinuria, and characteristic lesions of the kidneys, was observed in only 10 of the remaining 2 1 dogs with aortic blood flow reduction. The 11 other dogs presented some hypertension, placental infarcts, minimal glomerular endothelial swelling, and occasional traces of protein in the urine. It is difficult to explain why the same blood flow reduction leads to toxemic manifestations in some animals and not in others. The hypothetical takeover by collateral circulation does not seem to be the answer. For instance, in three cases, aortic blood flow reduction slowly and progressively pushed up to 70% to 80% resulted in the partial or total death of the litter, yet the mother in these cases presented minimal or no toxemic manifestations either clinically or anatomically. It appears that

REFERENCES

GYNECOL.

some animals are predisposed to experimental toxemia and some are not. In brief, placental ischemia may have ill effects on: (1) the mother, with resultant toxemic manifestations; (2) the litter, with death of some or all of the fetuses; (3) both mother and litter. It was difficult to say which were the most severe manifestations of toxemia in these animals because the dogs were kept under observation only a few hours a day. What happened the rest of the time is unknown. However, the severe hypertension and proteinuria, the lesions in the kidneys and placenta, and the frequent fetal deaths and one maternal death can be taken as evidence that some form of severe toxemia might have occurred. Advanced toxemia was produced in only one third of the animals with this experimental model. Modification of the experimental outcome could possibly improve this figure. For instance, isoxsuprine hydrochloride (Vaspdilan), 5 mg given intramuscularly immediately after surgery, may prevent premature labor. Obviously, further studies are needed to determine why toxemia develops in some animals and not in others. In spite of these limitations, this experimental model provides an excellent opportunity for more extensive research. I acknowledge Professor H. L. Sheehan, University of Liverpool Medical School, England, who performed the light microscopic studies, and Dr. Conrad L. Pirani, Columbia University, College of Physicians and Surgeons, New York, New York, who performed the electron microscopic studies for this research project.

tion of experimental

1. Douglas, B. H.: Experimental Approaches to Toxemia of Pregnancy, Springfield, Illinois, 1971, Charles C Thomas, Publisher, p. 52. 2. Oeden. E.. Hildebrand. G. 1.. and Paee. E. W.: Rise of bl:od preisure during’ iscceiia of tie’ gravid uterus, Proc. Sot. Exp. Biol. Med. 43:49, 1940. 3. Douglas, B. H., and Langford, H. G.: Postterm blood pressure elevation produced by uterine wrapping, AM. J. OBSTET. GYNECOL. 97:231, 1967. 4. Hodari, A. A.: Chronic uterine ischemia and reversible experimental toxemia of pregnancy, AM. J. OBSTET. GYNECOL. 97:597, 1969. 5. Cavanagh, D., Rao, P. S., Tung, K. S. K., and Gaston, L.: Eclamptogenic toxemia: The development of an experimental model of the subhuman primate, AM. J. OBSTET. 120: 183,

1974.

6. Chesley, L. C., and McFaul, I. E.: Studies on surviving human placental tissue, AM. J. OBSTET. GYNECOL. 58: 159, 1949. 7. Beker, J. C.: Aetiology of eclampsia, J. Obstet. Gynaecol. Br. Emp. 55:756, 1948. 8. Abitbol, M. M., Gallo, G. R., Pirani, C. L., et al.: Produc-

Gvnecol.

J. OBSTET.

GYNECOL.

toxemia in the pregnant rabbit, AM. 124:460,

1976.

9. Abitbol, M. M., Pirani, C. L., Ober, W. B., et al.: Production of experimental toxemia in the pregnant dog, Obstet. Gvnecol. 48:537. 1976. 10. Abitbol, hi. M., Ober, w. B., Gallo, G. R., et al.: Experimental toxemia of pregnancy in the monkey: With a preliminary report on renin and aldosterone, Am. J. Pathol. 86:573, 1977. 11. Abitbol, M. M., Driscoll, S. G., and Ober, W. B.: Placental lesions in experimental toxemia in the rabbit, AM. J. OBSTET.

GYNECOL.

125:942,

1976.

12. Abitbol, M. M.: Intravascular coagulation in the experimental model of toxemia, N. Y. State J. Med. 78:1397, 1978. 13. Abitbol, M. M.: Hemodynamic studies in experimental toxemia of the dog, Obstet. Gynecol. 50:293, 1977. 14. Sheehan, H. L.: Pathological lesions in the hypertensive toxemias of pregnancy, in Ciba Foundation Symposium, London, 1950, J. & A. Churchill, Ltd., pp. 1622. 15. Sheehan, H. L., and Lynch, J. B.: Pathology of Toxemia of Pregnancy, Baltimore, 1973, The Williams & Wilkins co.