A chronic conscious dog model for direct transhepatic studies in normal and pancreatic islet cell transplanted dogs

A Chronic Conscious Dog Model for Direct Transhepatic Studies in Normal and Pancreatic Islet Cell Transplanted Dogs

D. W. O’BRIEN, H. A. SEMPLE,G. D. MOLNAR, Y. TAM, R. T. Couns, R. V. RAJOTTE,AND j. BAYENS-SIMMONDS

A chronic

conscious,

a period hepatic

of 6-8 effects

of the hepatic food

hormone associated

and studies

blood

pre-

cations were

collected

IRI fluxes curves.

samples

Total

pling

interval.

period.

catheter

portal,

hepatic

for plasma

Hepatic

studies studies

Key Words:

Conscious

transplantation;

organs

subject,

other

(IRI) and glucose. of flux X time the curves

for each sam-

to determine

The techniques which

allows

Freartery

hepatic

described

for repeated

may also have application

in tran-

in other

than the liver.

physiological

Food and drug

model,

compli-

and carotid

was calculated

it was possible

conditions.

animal

in the same conscious involving

veins,

of

prepara-

are presented.

by interpolation

extraction

of analysis,

non-steady-state

of this complex

surgical

as the sum of the areas under

insulin

and

undergone

and possible

concentrations

period

have

The preparation

meal (OMT)

and jugular

studies

metabolism

probes,

maintenance,

insulin

for each time

By this method during

the development shepatic

care,

flow

over of the

metabolism, drug

in dogs that

and blood

IRI flux was determined

IRI extraction

chronic

from

determined

for each sampling

balance

studies

to the study

and glucose

high first-pass

Data from an oral glucose

and analyzed

were

with

of insulin

transhepatic

for application

secretion

catheters

and postoperative

blood

for repeated

and islet cell auto-transplantation.

sampling

are described.

quent

model developed

of pancreatic

pancreatectomy

specialized tions,

has been

mechanisms

interactions,

previous

large mammal

weeks

model;

interaction

Hepatic

in first-pass

mechanisms; drug

Islet

cell

metabolism

INTRODUCTION

A large mammal model of pancreatic islet cell transplantation into the spleen has been developed by Warnock et al., 1983 and 1987. An acute model for transhepatic pharmacokinetic studies has been employed in numerous investigations in our laboratory (Wieczorek et al., 1985; Berzins et al., 1986; O’Brien et al., 1988). We report on the development of a chronic preparation studies for a useful period of 6-8 weeks. Dogs

in dogs for repeated transhepatic are implanted with four blood sam-

From the Department of Medicine (D.W.O., G.D.M., R.V.R.), Faculty of Pharmacy and Pharmaceutical Sciences (H.A.S., Y.T., R.T.C.), and Health Sciences Laboratory Animal Services fJ.B-S), University of Alberta, Edmonton, Alberta, Canada Address reprint requests to: Dr. D. W. O’Brien, Surgical Medical Research Institute, 1070 Dentistry/ Pharmacy Centre, University of Alberta, Edmonton, Alberta T6G 2N8, Canada. Received June 19, 1990; revised and accepted October 9, 1990. 157 Journal

of Pharmacological

0 1991 Elsevier

Science

Methods Publishing

25, 157-170 Co.,

Inc.,

(1991)

655 Avenue

0160.5402/91/$3.50 of the Americas,

New

York,

NY 10010

158

D. W. O’Brien et al. pling

catheters

carotid

placed

artery,

in the portal,

and two

simultaneous

intermittent

measurement

of hepatic

liver can be measured the model

(among

flow

hepatic

probes

blood blood

sampling

flow, other

from

animal.

applications)

veins

artery.

and With

and continuous

of a physiological

in a conscious potential

jugular

vein and hepatic

all four catheters

all parameters

simultaneously

numerous

and right external

on the portal

model

of the

The applications

include

hepatic effects of pancreatic hormone secretion and carbohydrate patic mechanisms associated with high first-pass drug metabolism,

investigation

of of

metabolism, heand food inter-

actions (Semple et al., 1990), and the study of transhepatic hormone pancreatic islet cell transplanted dogs (O’Brien et al., 1990a and b).

balance

in

METHODS

Model

Animal Implant

Preparation

Preparation

Four silastic catheters, 0.062 in id x 0.125 in od (lengths: carotid, 70 cm; jugular, 70 cm; hepatic, 80 cm; and portal, 70 cm) (Dow Corning, Midland, MI) and two ultrasonic

transit

time flow

probes

(Transonic

Systems

Inc.,

Ithaca,

NY), 4 mm for

the hepatic artery and 8 mm for the portal vein, were cleaned with chlorhexidine scrub (Hibitaine, Ayerst Laboratories, Montreal, Canada) and rinsed with distilled water.

Double

the external

velour

discs and a 2-mm onto a catheter the catheter, ing, Midland, then

dacron

(Meadox,

ends of all the devices.

Oakland,

The dacron

NJ) cuffs were velour

hole was made in each disc. Two such discs were

or flow probe

lead. Between

then threaded

the discs and 15 cm from one end of

a bolus of medical grade silicone elastomer glue (Silastic, Dow CornMl) was applied around the catheter and to the two discs which were

squeezed

together

to form

a laminate

held together

and to the catheter

the glue (Figure 1). At the edge of the disc, the glue thickness the middle,

placed 15 cm from

was cut into 2-cm diameter

about

3-4

mm.

To provide

an anchor

point

by

was 1 mm, and towards at the entrance

into the

vessel, each catheter had a 4-mm diameter silicone glue “button.” The button positions, in cm from the vessel end were carotid and jugular, 15-20, hepatic and portal, 10. The preparations were allowed to dry for at least 24 hr before being packaged and sent for ethylene oxide sterilization. Some of the flow probes had pin connectors, which were unsoldered from their leads before

preparation, and some were fitted with Konigsberg skin button conbuttons were not nectors (Transonic Systems Inc., Ithaca, NY). The Konigsberg prepared as above, but dacron velour discs, the same diameters as the buttons, were cut and glued to the flat surfaces with silicone adhesive. To fit the connector portion of the skin buttons, we cut a 4-mm hole in the dacron disc for the top surface

and then drew

it over the connector

to be attached

with silicone

adhesive

to the top of the base. They were also gas-sterilized before use. Probe reflectors, screws, and a screw driver were sterilized in a separate pack for use during Stage II surgery, which included probe placement on the appropriate vessels.

A Chronic Conscious Dog Model

bie velour

Medical

dacron

grade

(dia. 2 cm)

silastic

adhesive

Top and bottom are pressed together leaving a thick conical centre and thin peripheral edges

FIGURE 1. Construction of a two-layer double velour dacron flange adhered together and to the catheter or probe lead with medical grade silastic adhesive. The skin grows into the dacron from above and below sealing off the exit site wound and preventing sinus tract infections. Surgery Anesthesia and Preparation Random source, intact male,

mixed

breed

dogs weighing

20-25

kg were

selected

for study. After an overnight fast, each dog was administered a preanesthetic mixture (0.1 mL/kg) consisting of acepromazine, meperidine, and atropine. The mixture was made up of 1 mL acepromazine atropine (0.5 mg/mL) to equal 1000 mg i.m.)

and another

and antibiotic

was continued

When

the animal

bated and prepared

1000 mg of Cloxacillin

plane

a 2% halothane/L

of anesthesia

(50 mg/mL), and 5 mL antibiotic (Cloxacillin

postoperatively

for 3 days (500 mg Cloxacillin

was sedated,

mask until a surgical

(IO mg/mL), 4 mL meperidine a volume of 10 mL. A pre-op

O2 mixture

was reached.

BID)

was given

postoperatively.

was administered

The subject

for sterile surgery on the back, neck,

(i.m.)

was then

right side,and

by intu-

abdomen.

Stage I: Subcutaneous Placement of Catheters Preparation of Skin Interface Sites and Subcutaneous Pockets. The subject was first placed in sternal recumbency with the head and hind limbs turned to the right to expose the right aspect of the neck and the right flank. After preparation for surgery and draping, a dorsal midline incision was made from the cranial point of

159

160

D. W. O’Brien et al. the shoulder

caudally

for about

made by blunt dissection,

18 cm. A row of six subcutaneous

extending

pockets

about 4 cm to the right of the incision.

3-cm incision was made in the neck cranial to the point of the shoulder large enough

to hold two catheters

vein by blunt dissection.

formed

in a ventral

A third 3-cm incision

direction

and a pocket

towards

was made vertically

were

Another

the jugular

in the right flank

about 8-10 cm caudal to the last rib. A pocket extending cranially to cover the costovertebral angle was formed by blunt dissection. It was made large enough to hold two catheters Placement neck pocket

and the two flow probes.

of Catheters. through

Intestinal

forceps

were

the edge of the cranial-most

then

passed

pocket

bluntly

from

the

on the back. The carotid

catheter was grasped and pulled through into the neck pocket until the dacron flange lay in the dorsal pocket. A small (I-2 mm) stab incision was made in the middle of the pocket about 3-4 cm from the midline incision using a No.11 scalpel blade

and the

hemostatic

external

forceps,

end

of the catheter

until the flange

was pulled

lay against

through

the subcutis.

using

mosquito

This process

was re-

peated with the jugular catheter in the second pocket, and the vessel end of both catheters were coiled and placed in the neck pocket. The neck pocket was closed with two subcutaneous was closed

with

layers of simple continuous

simple

was repeated with pocket. Tunnelling

interrupted

3-O stainless

were

Skin Closure.

sutures,

and the skin

The same

process

the remaining two catheters and two flow probes in the flank was accomplished using 30-cm De Lee dressing forceps. If the

flow probes were fitted with Konigsberg incisions

3-O Dexon

steel sutures.

made

skin buttons,

3-4-mm

holes instead of stab

at the skin exit points to accommodate

the pin connectors.

The first suture layer of the dorsal midline

incision

was 3-O Dexon,

placed close to the velour flanges to eliminate dead space. A second simple continuous subcutaneous layer of 3-O Dexon was followed by simple interrupted 3-O stainless

steel skin sutures.

Sealing of External Ends of Catheters. (14 ga x 2 in) were ends of the catheters.

Teflon

intravenous

then cut to about 2.5 cm and inserted

placement

They were glued in place using cyanoacrylate

lock caps with rubber septa (PRN adaptor, Deseret used to seal the ends of the catheters. The sealed Anticoagulant-citrate-dextrose

(ACD)

Solution

Medical, catheters

B, U.S.P.

catheters

to the hilt into the open glue. Male luer

Inc., Sandy, UT) were were then filled with

with

1.5%

formaldehyde

added to sterilize the catheter lumens. The drapes were then removed and the subject turned over onto its back, prepared, and draped for stage II of the surgery. At this time, about 1.5 hr had elapsed since the subject was anesthetized. Stage II: lntraabdominal

Placement of Catheters and Flow Probes An extensive ventral midline incision was made from the xiphoid to the prepuce. Bleeders were cauterized. The duodenum was retracted to expose the portal vein area and right flank. The small intestine was packed off using a moist laparotomy sponge.

Approach to the Abdomen.

A Chronic Conscious Dog Model

CatheterRetrieval. The location of the right flank pocket containing the catheters and flow probes was palpated at the costovertebral angle next to the right kidney. After confirmation of the location, a blunt incision was made with Metzenbaum scissors through the abdominal wall into the pocket. A Love nerve retractor was used to retrieve the catheters and flow probes. Angled peripheral vascular Debakey forceps were then used to puncture the membrane that attaches the caudal vena cava to the dorsal wall of the abdomen, just cranial to the cranial pole of the right kidney. The two flow probes and the portal vein catheter were passed through this hole beneath the vena cava to lie near the portal vein. In this position, the catheters could not interfere with the small intestine. The hepatic arterial flow probe was Hepatic Arterial Flow Probe Placement. placed first. The hepatic artery is a major branch of the ciliac artery, and is visible as it courses cranially next to the caudal vena cava for a distance of about 6 cm. It is covered by a nerve net. To place the probe, we freed the artery by blunt dissection, usually along a reasonably straight portion where there was a gap in the nerve net. The artery was lifted with Mixter forceps and the reflector of the probe was placed under the vessel (The probe lead exited caudally). The retainer plate was then bolted into place. A stay suture (3-O silk) was placed to fasten the flow probe to the wall of the vena cava in such a way that the artery passed unobstructed through the probe lumen. A second stay suture attached the probe lead to the vena cava. Portal Venous Flow Probe Placement. The portal venous probe was placed next. The entrance of the gastroduodenal vein into the portal vein was located and a 2cm section of portal vein caudal to this was freed by gentle blunt dissection. The freed section of portal vein was elevated using Mixter forceps and the g-mm probe positioned with the probe lead exiting caudally. The retainer plate was bolted in place and the probe and lead were anchored to the surrounding tissue with 2-O silk sutures or to the hepatic artery probe. Protruding pieces of fatty tissue were trimmed from around the probe. This was necessary because fat has poor acoustical conduction and could lead to instrumental measurement errors if it were allowed to enter the lumen of the probe. Gastroduodenai Artery ligation. The hepatic artery sends branches to each liver lobe, then continues distally as the gastroduodenal artery. To eliminate extrahepatic blood flow, this branch was ligated. The regions of the stomach, pancreas, and duodenum supplied by the gastroduodenal artery have a collateral blood supply that compensates adequately following its ligation. Portal Vein Catheter Placement. The gastroduodenal vein empties into the portal vein about 1.5 cm from the hilus of the liver, downstream from the flow probe. The vessel was utilized as an entry point for the portal vein catheter. About 1-2 cm from the portal vein the gastroduodenal vein emerges from the body of the pancreas. At this position, it was freed by blunt dissection, ligated upstream, and a loose ligature was placed downstream close to the entry to the portal vein. Backflow from the portal vein was prevented by putting tension on the downstream ligature. A

161

162

D. W. O’Brien et al. small cut was made in the wall of the vessel and the portal vein catheter was inserted and fed downstream into the portal vein. It was very important to palpate the position of the catheter

because

lodged

branch

in the portal

it tended supplying

to pass across the portal vein and become the caudate

process

and right lateral

In this position, the catheter was extremely prone to blockage. Therefore, positioned ideally by palpation, then removed, and trimmed, if necessary, the tip lay in the hilus when The ligature

on the catheter. angle.

the catheter

was gently tightened,

In order

fully to the anchor

vein enters

the catheter

more

the portal

vein. The vein was then transected stabilize

the portal

its position,

between

vein. The catheter and the position

vein almost

in the direction

vein and to prevent blockage of the catheter, the following formed. A second ligature was placed above the catheter in line with

point.

tied, and then tied again above the anchor

The gastroduodenal to direct

was inserted

lobe. it was so that

the ligatures

was fastened

was checked

of flow

point

at a right

in the portal

manipulation was peron the gastroduodenal and the catheter

to the portal

by palpation.

flow

moved probe

to

The duodenum

was replaced in position. The technique in pancreatectomized islet auto-transplanted dogs differed slightly with respect to portal vein catheterization depending on whether, when the pancreatectomy

was performed,

a I-2-cm

portion

of the gastroduodenal

served at the site of entrance into the portal vein. accessible, then the portal vein was catheterized scribed.

If this site was not available,

vein was pre-

If this remnant was available via the vessel as previously

then a branch

of the caudal

mesenteric

was utilized approximately 5 cm distal to the entrance of the gastroduodenal The catheter tip was advanced to lie in the main portal vein at the entrance liver as before. This procedure

In this situation,

the catheter

had no discernable

passed through

effect on blood

and devein

vein. to the

the flow probe window.

flow measurement

vein. Because it was not possible to know in advance which the portal vein catheter was constructed with two “anchoring

in the portal

site would be used, buttons.” If the pri-

mary site (via the gastroduodenal vein) was not available, then the button nearest the catheter tip was removed to allow the catheter to slide further into the portal vein and was anchored tectomized

at the second

islet cell auto-transplants

Hepa tic Vein Catheter

button. were

Placement.

To

All other

as previously install

the

procedures

in the pancrea-

described. hepatic

vein

catheter,

we

grasped the left lateral lobe of the liver using a wet sponge and drew it into the field of view. The diaphragmatic surface was palpated to detect a depression where a large branch

of the hepatic

vein

ran through

the parenchyma.

This depression

could usually be palpated along its length to the common hepatic vein, although the intrasubject variation in anatomy was large. At the distal end of the depression, usually about 3-6 cm from the tip of the lobe, an incision was made in the liver capsule and with the end of a scalpel handle, the liver parenchyma was bluntly dissected to expose the vein. The inevitable oozing of blood from the parenchyma was minimized by applying pressure from beneath the lobe. A stay suture was placed through the exposed wall of the vein to anchor the catheter. No other tissue in the liver is strong enough to use as an anchor. Just proximal to the stay suture, the vein

A Chronic Conscious Dog Model

was nicked and the catheter introduced. As it is possible in some animals to mistake a branch of the portal vein for the hepatic vein, two methods for confirming the placement of the catheter were used. First, the catheter ends were made longer so that they could be inserted well into the vena cava and palpated at the intersection of the vena cava and diaphragm. Second, the back pressure in the vein was checked by evaluating leakage when manual pressure on the vessel was removed. A portal branch would quickly flood the site with blood, whereas the low-pressure hepatic vein would leak very little. After confirming that the vein was indeed a branch of the hepatic vein, the catheter was trimmed so that when the anchor point abutted the entry point into the vein, the tip of the catheter lay about 1 cm into the common hepatic vein. The stay suture was then fastened above the anchor point on the catheter. The capsule was gently closed with three loose, horizontal, mattress sutures to prevent bleeding. The catheter was anchored to the body wall to the right of the xiphoid with a stay suture. Extra lengths of catheter and flow probe leads were gathered in coils along the right body wall and fastened there with stay sutures to prevent interference with the intestines. Ensuring Catheter Patency. Before closure of the abdomen, catheter patency was checked. The ACD solution was removed and the catheters were flushed with sterile saline, followed by a 1000 U/mL heparin lock. About 0.1 mL more heparin solution than the dead space of the catheters was used. Use of a large excess was avoided to minimize bleeding. Abdomen Closure. The linea alba was closed with simple interrupted 0 Dexon sutures, the subcutaneous layer was closed with simple continuous 3-O Dexon, and the skin was closed with simple interrupted 3-O stainless steel sutures. Placement of Carotid Arterial Catheter. To place the jugular and carotid catheters, we made a 4-cm ventral midline incision in the caudal portion of the neck. The subcutaneous fascia was bluntly dissected to expose the pocket containing the catheters. A hole was made in the pocket and the catheters were retrieved. The neck muscles were separated down to the trachea and the loose fascia was separated to expose the right external carotid artery. The artery was exposed, stripped, and clamped at least 2 cm upstream from the catheterization site. A 2-O silk ligature was placed downstream from the catheterization point and used to elevate the vessel. Two loose ligatures were placed around the vessel downstream from a bulldog clamp. A small incision was made in the wall of the artery, and the catheter was introduced through the incision and advanced to the level of the clamp. One of the loose ligatures was tightened around the artery to prevent leakage of blood around the catheter as it was fed towards the heart. The clamp was then removed and the catheter was fed into the artery until the anchor point of the catheter was reached. The catheter was doubly ligated in place and the downstream ligature was fastened above the anchor point on the catheter. The artery was then replaced in position and the catheter coiled one revolution to prevent kinking of the catheter and to allow for some lengthening, if necessary. The neck muscles were apposed with simple continuous 3-O Dexon sutures.

163

164

D. W. O’Brien et al.

Placement of Jugular Venous Catheter.

The right external

jugular

vein was ex-

posed by blunt dissection and stripped for a 3-cm length. It was ligated with 2-O silk at the upstream end of the exposed section and a loose ligature was placed downstream.

An incision

was made

in the in-wall

of the vein, and the catheter

was

introduced and fed to the anchor point. The loose ligature was tightened and tied, and the upstream ligature was fastened above the anchor point. The vessel was replaced

and the extra catheter

material

coiled

beneath

not kink. The dead space was closed with 3-O Dexon and skin layers were closed I and II was about 4 hr.

Recovery and Aftercare. lowed

to recover.

respective

During

connectors,

as for the abdomen.

The anesthetic recovery,

probe

and the subcutaneous

The total surgery

was turned

the flow

and the flow

the skin so that it could

sutures,

probe

time for stages

off and the animal

leads were

connections

were

was al-

resoldered checked.

to their Acoustic

errors were often initially encountered, presumably due to air spaces between the vessels and probes, but these disappeared in a few days as the probes healed in place.

The

dogs were

fitted

with

jackets

with

side pockets

Medical Arts, Los Angeles, CA) to protect the catheters probe leads and catheter leads were brought through

(Alice

the pocket. The ends could then be accessed from the pocket. Upon waking, the animals were allowed to recover in heated given 0.1-0.2 mg/kg buprenorphine gesia as required. The day following protein,

serum

lowed

to heal for at least 10 days before

Catheter Care. Inc.,

liver enzymes,

Dilute

gentamicin

Pte Claire,

cages and were

(Reckett and Colman (PL), Erie, IRL) for analsurgery, blood was taken for a complete blood

count,

ing, Canada

King Chatham

and flow probe leads. The the side of the jacket into

and serum

amylase.

experiments

spray (gentamicin

Quebec,

Canada)

The preparation

were

sulphate

was applied

was al-

started. 1 .I mg/mL,

around

Scher-

the skin inter-

faces twice daily when oozing was encountered. The catheters were flushed every 3-4 days by cleaning the PRN adaptors with disinfectant, removing the heparin locks, instilling

0.9% sodium

chloride,

followed

by ACD solution

with formaldehyde

(0.4%

anhydrous citric acid, 1.32% sodium citrate (dihydrate), 1.47% dextrose (mono H,O), and 1.5% of 37% formaldehyde) to fill the catheter dead space. The ACD was left for 5 min to sterilize the catheters lumens and then removed. The catheters were flushed with sterile saline and then filled with heparin solution. After 1 week of healing, the heparin strength was increased from 1000 U/mL to 10,000 U/mL. The PRN adaptors

were

changed

regularly

as required.

A sling was used for restraint

during the procedures. After the surgery, the subjects were fed canned dog food or a combination of canned dog food and dog biscuits depending on the nature of the study. Pancreatectomized dogs received twelve Cotazym tablets (Organon, Montreal, Canada) with each meal. TRANSHEPATIC

PHARMACOKINETIC

STUDIES

On the day of an experiment, the dog was brought to the laboratory early in the morning, having been fasted overnight, and placed in a sling frame that both re-

A Chronic Conscious Dog Model

strained and supported the animal gently. The dog was able to stand freely in the sling or be supported along its entire ventral surface, thus taking the weight off its legs. The vest was usually removed to facilitate access to the flow probe connectors. Blood flow was measured with a transit-time ultrasonic flow meter (Model T201) produced by Transonic Systems, Ithaca, NY. The PRN adapters were removed from the catheter ends and replaced with minimum volume extension sets (Cutter Biological, Elkhart, IN). On the other end of the extension set, a three-way stopcock was attached. Blood was withdrawn from the catheter to check its function and then the catheter was flushed with heparinized saline (IO U/mL). During the experiment, samples were drawn by first removing the catheter deadspace of 3-4 mL, then switching the stopcock to the other port to which was connected a 3-mL sample syringe. After the sample was removed the deadspace volume was reinjected. After each sample, the catheters were flushed with 3 mL of heparinized saline. Intermittent samples were withdrawn from four catheters simultaneously according to a sampling schedule. Various protocols were carried out on a single animal. However, for the purposes of this report only the protocol developed for an oral glucose meal test (OMT) will be described. The OMT consisted of 1 m/kg dextrose mixed with 5 g/kg Science diet canine maintenance (Hills Pet Products, Topeka, KS). The meat and dextrose bolus was placed in a dog dish and covered with parafilm. The meal was presented to the dog in a way which minimized any neurogenic arousal prior to the experimental period, The time of presentation was recorded as was the time to consume the meal. At this time the deadspace volume was withdrawn and the first samples (0 time) were taken. Prior to the “0” sample, baseline samples were withdrawn at -30, -20, and -10 min, and baseline blood flows were established. Additional samples were taken at 0, 2.5, 5, 7.5, 10, 12.5, 15, 30, 60, 90, 120, 150, and 180 min. Blood samples were placed in 3-mL chilled vacutainer tubes containing 45 USP units of sodium heparin. The samples were centrifuged and the supernatant was transferred to 3-mL stoppered tubes. Plasma glucose and insulin were measured as previously described (O’Brien et al., 1988). Blood flow in the hepatic artery and portal vein readings were extrapolated from the flow recordings at each time point in the study. Fluxes of insulin and glucose in the portal (PV) and hepatic vein (HV) and hepatic artery (HA) were determined (concentration x flow). The LAGRAN program (Rocci ML, Jusko WJ, 1983) was used to determine the area under the concentration vs. time curve (AUC) defined by the sampling time points. Interval AUCs were summed to yield a cumulative area value extending from the initial to the last time point. Hepatic insulin extraction (HIE) for each interval was determined according to the following equation: (PVAUC

+ (PvAlJC

HAAuC) +

HAAUC)

HVAUC ’

Extraction calculated in this manner represented the average extraction over the interval and therefore reduced the variability due to small differences in sampling time or transit time through the liver when blood flow and sample concentrations

165

166

D. W. O’Brien et al.

may be changing rapidly. By this method of analysis it was possible to determine hepatic insulin extraction and clearance during non-steady-state conditions.

RESULTS

A schematic representation of the circulation showing the major vessels supplying the liver, placement of flow probes, and vessel sampling points is depicted in Figure 2. Table 1 shows the data from a representative experiment including the plasma flows (Q) in the portal vein (QW) and hepatic artery (QHA); the plasma insulin (IRI) concentratrons (C) in Cpv, CHA, the hepatic vein (C,v) and right external jugular vein (Cv); the IRI plasma flux (F) in these vessels (Fpv, FHv, and F,.+J; and the area under the curve calculated for the time intervals beginning at zero. For each area calculation the end point of the interval is shown. For example, the interval O-2.5 min, AUC = 86.2 mU in PV. In the final column the interval hepatic insulin extractions are shown for each interval from 0 to 180 min. Insulin fluxes are plotted versus time and the HIE calculated for each time interval is shown. HIE is high when hepatic influx of IRI is high and decreases as influx of IRI decreases (Wieczorek et al., 1985).

Schematic

Representation

Showing

Sampling

of Circulation Points

Corporeal Clrculatlon

Pulmonery Clrculetlon

Vene Cere

@ifA Flow probe

Sempllng Cethetere (A): cerotld ertery right externel Juguler rein portel rein hepetlc rein

FIGURE 2. A schematic diagram of the circulation showing the positional relationships of the flow probes and the sampling catheters.

A Chronic Conscious Dog Model TABLE 1

Data from a Representative Experiment INTERVALAUC

PLASMA FLOW TIME (MIN)

QPV ML/

IRI Cont. mU/L

QHA MIN

CPV

CHV

CHA

FLUXES (MU/MIN)

Cv

FPV

OF FLUXES IN

INTERVAL IRI%

MU

FHV

FHA

FpvAUC

FHvAUC

FHAAUC

EXTRACTION

-

0

421

34

165

21

12

9

69.5

9.55

0.40

2.5

405

28

21

10

12

10

8.5

4.33

0.34

86.20

16.93

0.95

5.0

354

28

5

3

6

6

1.77

1.15

0.17

6.53

6.35

0.63

11

7.5

337

47

4

2

3

3

1.35

0.77

0.14

3.20

2.07

0.38

42

10.0

337

47

4

2

2

3

1.35

0.77

0.09

3.38

1.89

0.28

48

12.5

287

74

3

2

2

2

0.86

0.72

0.15

2.45

1.64

0.27

31

15.0

354

54

11

7

8

7

3.89

2.85

0.43

5.38

4.12

0.69

32

30

371

47

8

3

3

4

2.96

1.25

0.14

82.93

54.55

7.52

40

60

472

41

20

8

15

18

9.44

4.10

0.61

168.39

71.87

9.90

60

90

421

41

40

7

8

9

16.85

3.23

0.33

417.73

113.58

14.73

75

120

421

34

11

7

9

10

4.63

3.18

0.30

331.69

95.44

9.06

72

150

371

41

12

7

8

9

4.45

2.88

0.33

124.39

93.51

9.70

30

180

337

34

5

2

3

3

1.68

0.74

0.10

98.42

58.90

7.10

44

81

Abbreviations as in text.

DISCUSSION We have reported

on the development

of a chronic

dog preparation

in which

all

parameters of a physiological model of the liver can be directly measured simultaneously in a conscious animal. This model has proven to be a useful method to study transhepatic insulin balance in normal and pancreatic islet cell-transplanted dogs (O’Brien et al., 1990a and b) and the interactions between food and drugs that undergo

extensive

first pass metabolism

(Semple

et al., 1990). Other

workers

have

reported on methods in conscious swine (Mustard et al., 1986; Olesen et al., 1989) that utilize similar vessels as described by our methods. Olesen et al. (1989) described

a direct cannulation

procedure

is different

from

of an hepatic the direct

vein from the margin of a liver lobe. Their

cannulation

technique

that we describe

in

our methods and although the resulting catheter placement may be similar, an incision of the margin of the liver lobe may result in excessive bleeding if a portal vessel is transected. The catheterization site that we employ on the diaphramatic surface of the left lateral lobe and the technique for hepatic vein catheterization result in minimal perturbation of the liver. As seen at necropsy the site heals in around the catheter. Direct palpation of the catheter tip as we have done eliminates any uncertainty as to the position of the hepatic vein catheter. This method represents zation.

an advantage

over previously

reported

methods

of hepatic

vein catheteri-

Various factors contribute to the success of our model. One of the most important factors is the application of transit time ultrasonic flow meter (Model T201 Transonic Systems Inc., Ithaca, NY), which has significant advantages over other blood flow

167

168

D. W. O’Brien et al. meters (e.g., electromagnetic, doppler) include direct measurement of volume or flow velocity

profile,

insensitivity

for chronic application. These advantages rate of flow regardless of vessel dimension

to blood flow profile,

alignment

and flow sensor, nonconstrictive fit on the vessel, and accuracy long-term implant (Burton and Gorewit, 1984). In our studies, implanted for up to 8 weeks with trouble-free operation. A second factor for success has been the development for application This device terface

at the skin exit sites of catheters has minimized

with subsequent

sistent serious

the incidence

problem

in chronic

animal

of an implantable

and flow probe

of infections

sinus tract infection

between

studies (Dennis

flange

leads or connectors.

developing

and possibly

vessel

over the period of probes have been

systemic

at the skin ininfection,

a per-

et al., 1986) and in clinical

situations (Gruneberg, 1983; Hampton and Sherertz, 1988). Although other types of catheter flanges utilizing other materials have been reported (Schmidt et al., 1988; Gray et al., 1985), acquisition

factors

of prefabricated

such as availability, devices

function,

and prompted

and expense

us to develop

limited

our

and construct

our

own implantable catheters that have proven to be successful. Problems encountered included blocked catheters, skin interface ticemia,

and chewing

of catheters

and probe

the surgery were invariably due to mechanical by manipulating the position of the subject, another

day, they would

the preparation

spontaneously

(usually after 4 weeks

leads.

Blocked

infections,

catheters

sep-

soon after

blockage. They could often be freed but occasionally if simply left until

become

free.

postsurgery,

Blockages

later in the life of

unless an episode

of septicemia

had occurred) were more serious, because they were due to growth of a fibrous sheath around the catheter. Fluid could be infused but not withdrawn, because the sheath would collapse around the catheter tip, blocking it. This was confirmed on postmortem

examination.

Such a blockage

meant

termination

of experiments

on

that subject. The hepatic and portal vein catheters were most prone to malfunction due to mechanical obstruction. This problem was considerably alleviated by the use of catheters

that had side-ports

in addition

to the tip opening.

The side ports

were constructed so that the area of any two side-ports equalled the area of the tip opening. Six side-ports were made at various orientations within 0.5 cm of the catheter

tip.

Figure

3 shows

a drawing

of the catheter

tip with

side-ports.

Skin

interface infections became rare as experience was gained. If they did occur, they were usually localized. Occasionally, an animal would chew through the jacket and damage the catheters and flow probe leads. The flow probe leads were spliced and resoldered, and the catheters ends were trimmed and refitted with teflon catheters and PRN adaptors. A serious consequence of chewing off a catheter was that bacteria could be introduced directly into the bloodstream via the damaged catheter to cause septicemia. If this occurred, blood cultures were performed and appropriate antibiotics administered on the basis of culture results (penicillin, cloxacillin, or gentamicin, because they are excreted mostly unchanged in the urine). After recovery, the subject was not used for an experiment for at least a l-week washout period after the antibiotic treatment had finished. Another serious consequence of chewing was hemorrhage from the carotid artery catheter if the end was bitten off. This only

A Chronic Conscious Dog Model

Tip of silastic catheter used to alleviate mechanical blockage in chronic catheterized vessels. Outside diameter: 0.125 inches -

1 cm

Inside diameter: 0.062 incherr

Side ports Area of any 2 port8 equal6 area of catheter tip.

SO” Cross section of tip showing location of side ports.

FIGURE 3. Use of a catheter with six side-ports alleviated the problem of mechanical obstruction of the catheter tip that was a common problem in the hepatic and portal vein catheters.

occurred on one occasion. In most cases, the very low incidence of exit site infections and the wearing of the jacket kept catheter or probe damage to a minimum. Factors, such as the use of preoperative antibiotics, good surgical technique, and careful catheter maintenance, have contributed to the successful development of this complex animal model that allows for repeated metabolic studies in the same subject.

’ Supported by the MSI Foundation and the Muttart Diabetes Research and Training Centre.

REFERENCES Berzins R, Tam YK, Molnar DG, Rajotte RV, Wieczorek KR, McGregor JR, Fawcett DM (1986) Pharmacokinetic approach to the estimation of hepatic removal of insulin. Pancreas 1:544-549. Burton RG, Corewit RC (1984) Ultrasonicflowmeter

uses wide beam transit time technique. Electronics

Medical

15 :68-73.

Dennis MB, Cole JJ, Scribner BH (1986) Vascular Access in Large Laboratory Animals in Methods of Animal Experimentation. Volume VII, Part A.

169

170

D. W. O’Brien et al. Eds., WI

Gray

and

JE Heavner,

NY:

Academic

Olesen

Press. Gray

CG,

White

McKinnan

FC, Crismon

strumentation

techniques catheter.

utilizing

the

June 1985, Abst.

Bristol:

Hampton

AA, Sherertz

fections

GORE-

N America Mustard

C, Finley RJ, Duff JH, Roduik

C (1986) A simplified

studies

in conscious

method

swine.

(1990a)

O’Brien

CD,

Increased

Lab Anim Sci

O’Brien

decrease

DC, Rajotte

RV.

Molnar

CD,

Dia-

insulinemia.

portal

Semple normal,

insulin

islet auto transplanted

HA, O’Brien

DC

but quantitatively

secretion

conscious

in pancreatic dogs. Diabetes

39:311A. O’Brien tersen

Schmidt

nous catheter

LAGRAN

program

in pharmacokinetic

Prog Biomed

W, Dehn

DW, TDP.

patic insulin

Rajotte (1988)

RV, Molnar Somatostatin

extraction.

for

analysis.

16:203-216.

A, Hutter

JF (1988) A central

for long term

fects and pharmacokinetics

Semple

ve-

studies

on drug ef-

in Munich

Minipigs.

13:143-147.

HA,

Tam

YK, O’Brien

DW

logical

modelling

of the hepatic

tween

food

hydralazine

dog.

and

Pharmaceut

Warnock

CL,

(1990)

Physio-

interaction

in the

be-

conscious

Res 7:S223.

Rajotte

RV, Kneteman

NM,

Ellis D, to mixed

long term transplanted

Transplant

lets of Langerhans. Warnock

is-

Proc 19:969-973.

CL, Rajotte RV, Procyshyn

moglycemia

Quantitatively

decreased

Jusko WJ (1983)

in islet transplanted

to peripheral

33:A180.

DW,

(1990b).

ML,

area and moments

meals in dogs bearing Molnar

dogs contributes

betologia

Rocci

Toth K (1987) Oral glucose and response

DW,

in-

Sci 39:429-

for meta-

36:393-395. O’Brien

systemic

access in-

68:57-71.

RA, Lipohar

J, Hohr

pigs. Lab Anim

E f Drug Metab Pharmacokinet

Surgical Clinics

patients.

and tracer

432.

Comput

pp. 162-171.

RJ (1988) Vascular

in hospitalized

loading

in conscious

No. 066.

Wright-PSG,

and

intragastric

fusion

RN (1983)

Peters,

hepatic

J,

of ConferResearch. College

B (1989) Simultane-

portal,

in-

Sepsis and Central Venous Catheter Systems in A Manual of Central Venous Catheterization and Parenteral Nutrition. Ed., JL

Cruneberg

E, Tronier

of

during

Proceedings

ence on Swine in Biomedical Park, MD.

swine

sampling

blood

RP, Wisniewski

D, Bloor CM (1985) Chronic

TEX peritoneal

bolic

HP, Sjontoft

ous

CD,

Toth

K, Pe-

increases

he-

Diabetes Res 7:171-178.

creatic in dogs. Wieczorek

after

fragments

Diabetes

reflux

into the splenic

RV, Molnar

BT, Fawcett

DM

bed

CD,

McGregor

(1985) The effect

hepatic

insulin

influx on insulin

parison

of net flux and per cent extraction

dices of insulin retention.

263.

pan-

vascular

32:452-459.

KR, Rajotte

JR, Lussier

AW (1983) Nor-

of islet-containing

retention:

of

comas in-

C/in invest Med8:257-