JOURNAL OF SURGICAL RESEARCH 38,272-280 (1985)

Islet Cell Isolation in Experimental d,l-Ethionine Pancreatitis in Dogs’

LUTHER F. COBB,M.D.,* FRANCESCOMARINCOLA, M.D.,*

ATSUSHI HORAGUCHI, M.D.,’ NORIO AZUMI, M.D.,?
AND RONALD C. MERRELL, M.D.*
Departments of *Surgery and TPathology, Stanford University School of Medicine, Stanford, California 94305

Submitted for publication January 23, 1984

Bulk isolation of islets of Langerhans for biochemical studies or tranplantation has generally been
associated with significant or even prohibitive contamination by acinar tissue. High-purity islet
preparations are associated with an extremely low yield. The acinar tissue can be ablated with d,I-
ethionine, but previously, this tactic has been restricted to rodents because of toxicity problems in
larger animals. A dosage regimen in dogs which reduces amylase content of the pancreas to 0.3% of
normal with complete preservation of insulin content is reported. Ductal perfusion with collagenase
permits the recovery of 45% of the islet cell mass as determined by extractable insulin. These islets
prove functional in in vitro perifusion studies and in ahografts. 0 1985 Academic FIW, inc.

INTRODUCTION a dosage schedule which permits acinar de-

struction and islet processing in mongrel
The ethyl analog of methionine, d,l-
dogs.
ethionine, is an extremely toxic compound
which can destroy the acinar pancreas while
preserving the endocrine function. Payne and MATERIALS AND METHODS
colleagues [2] used this property to facilitate
islet recovery from rats. After a period of A 2.5% (w/v) solution of d,l-ethionine
dietary ethionine, a single rat pancreas pro- (Sigma Biochemical, St. Louis) was prepared
vided sufficient processed islets to correct in normal saline at 50°C. A unit dose of 50
streptozotocin-induced diabetes mellitus in ml was administered via the cephalic vein.
three to four isogeneic recipients. This en- A variety of dosagescheduleswere studied
hanced yield has proved useful in routine but the compromise of frequency and dura-
donor preparation for other islet studies [ 17, tion was three times per week for 3 weeks.
26, 271. This stands in contradistinction to The regimen was reasonably well tolerated
the normal untreated pancreas in that islets and consistently gave near-total ablation of
from multiple donors must be pooled to the acinar tissue based upon histologic study.
restore glucose homeostasis to a single recip- If the drug was given more frequently in a
ient. The systemic toxicity of ethionine is greater total dose or for a longer period,
extremely significant in large animals and hepatic toxicity was prohibitive. The final
successful acinar ablation in dogs has not total dose varied from 10.0 to 12.5 g per dog.
previously been reported. We have developed During the final week of drug administra-
tion, listlessness and loss of appetite be-
came marked and pancreatectomy followed
’ This work was supported by a Basil O’Connor grant shortly.
from the National Foundation. Ronald C. Merrell is a After completion of the ethionine course,
recipient of a ResearchCareer Development Award from the dogs were anesthetized with intravenous
the NIH. Luther F. Cobb is a recipient of a National
Research Service Award, Fellowship 5F32AM06493-02.
thiamylal sodium (SuritaI) and maintained
r Current address: Department of Surgery, Tohoku on positive-pressure ventilation with 100%
University, Sendai, Japan. oxygen and 1.O to 1.5% halothane. The pan-
0022-4804/85 $1.50 272
Copyright Q 1985 by Academic Press, Inc.
All rights of reproduction in any form reserved.
 COBB ET AL.: ISLETS IN ETHIONINE PANCREATITIS 273

cresswas excisedcompletelyvia an upper from the periductal area; however,when

midline incision, carefully preservingthe other areaswereassayed, little differencein
blood supplyto the gland until the end of insulin values was found.
the dissectionto limit the warm ischemia The homogenizedpancreasand the di-
time. The two limbs of the major pancreatic gestedcell suspensionwere assayedfor insulin
duct were cannulated with No. 20- or 22- and amylase. Insulin was measured in acid
gaugeplastic iv catheters (Jelco Laboratories, alcohol extracts by the double-antibody ra-
Raritan, N. J.) which were sutured tightly dioimmunoassay technique of Morgan and
into place. Lazarow using porcine insulin standards; re-
The pancreas was processedby the ductal sults were expressed in milliunits of insulin
perfusion method [9] with minor modifica- per gland or per cell preparation. Samples
tions. Briefly, the gland was placed into a for amylase were sonicated and centrifuged.
recirculating pump chamber and perfused The supernatant was tested with the Amy-
via the ductal cannulae with Hank’s balanced lochrome dye assay(Roche Diagnostics, Nut-
salt solution (HBSS, GIBCO, Grand Island, ley, N. J.), and the results were expressed in
N. Y.) plus 25 mM Hepes (n-2-hydroxyethyl dye units per gland or per cell preparation.
piperazine-N-2-ethane sulfonic acid, Sigma). Three of the islet preparations were ran-
After 10 min, the HBSS was removed and domly selected for in vitro glucose challenge
replaced by 100 ml of 0.5% collagenase to assessinsulin secretion using a perifusion
(Sigma Type I) in HBSS plus 25 mM Hepes system as described by Lacy [ 111.Cells were
and perfusion was resumed. Forty to fifty cultured overnight in Medium 199 ( 100 mg/
min were required for digestion. Cannulae dl glucose) plus 10% calf serum at 24” in a
were then removed and the gland was stirring flask. The next day, cells were washed
chopped with scissors in cold calcium- and three times in phosphate-buffered saline by
magnesium-free HBSS (CMF, GIBCO). The centrifugation and resuspended in fresh Me-
tissue was filtered through 200~pm steel mesh dium 199 with 0.1% bovine serum albumin.
strainer and washed three times in CMF by Approximately lo6 cells in suspension were
centrifugation. placed in a filter chamber behind a 5-pm
The cells were resuspendedin tissue culture membrane filter (Metric& Gelman, Ann Ar-
medium (Medium 199, GIBCO) with 25 mM bor, Mich.) and two prefilters (Gelman Me-
Hepes. Cell counts were done in a hemacy- trigard Superfine). They were perifused at a
tometer using Turk’s solution (crystal violet rate of 1.0 ml/min with a peristaltic pump
in 2% acetic acid); cell viability was assessed (Polystaltic, Buchler, Fort Lee, N. J.). After
by exclusion of trypan blue after 5 to 15 perifusion with 5.5 mM glucose for 1 hr, the
min. A fragment of unprocessed pancreas perifusate was changed to 16.6 or 22.0 mA4
was reserved and homogenized in a Dounce glucose in Medium 199. Serial aliquots of
tissue homogenizer (Bellco Glass, Vineland, the perifusate were taken for baseline and
N. J.). This sample was usually removed stimulated insulin output, which was mea-

TABLE 1

source Amylase (dye units) Insulin (milliunits)

Normal pancreas (40 -+ 8 g) 676,000 k 184,000 53,800 + 24,300

Processedcells from normal pancreas (28 k 12 X lo* cells) 119,000 + 71,000 30,500 f 19,600
Ethionine pancreas (5.1 + 0.7 g) Mean 1749 54,100 2 23,300
(range 196 to 5535)
Processedcells from ethionine pancreas (5.1 + 1.0 X 10’ cells) Mean 159 24,500 f 19,900
(range 0 to 5 15)
274 JOURNAL OF SURGICAL RESEARCH: VOL. 38, NO. 3, MARCH 1985

sued by the double-antibody radioimmu- method autoanalyzer (Yellow Springs Instru-

noassay. ment Corp.).
Biopsies of the pancreas were fixed in RESULTS
neutral buffered 10% formaldehyde and pro-
cessedfor histology. Sections were prepared Initial trials with a daily course of ethionine
with hematoxylin and eosin as well as with injections proved unsatisfactory. Toxic
insulin-specific immunoperoxidase staining symptoms appeared early, and insufficient
for &cell identification (Imulok, Carpenteria, time had passed to allow complete acinar
Calif). atrophy. After a three dose per week regimen
One dog was given a sixdose course of was adopted, the drug was much better tol-
ethionine over a 2-week period, and then erated and histological acinar atrophy was
was allowed a 3-week period for recovery. virtually complete after 3 weeks. Table 1
This gland was processedfor cell preparation summarizes data from five normal dogs
and histology in the same manner. weighing 14 to 19 kg which received this
One apancreatic control dog with estab- ethionine dosage, and compares these data
lished diabetes received a transplant of d,l- to normal pancreata processed in the same
ethionine-treated islet tissue processedin the manner for islet isolation. At laparotomy,
above manner. The cells were placed into its the pancreata from these animals were
spleen via the splenic vein. Serum glucose strikingly different from normal. A very
was measured daily using a glucose-oxidase small contracted pale remnant of pancreas

FIG. I. Hematoxylin-eosin-stained normal canine pancreas. Lightly stained islets are surrounded by a
matrix of exocrine acinar tissue arranged in lobules. X250
 COBB ET AL.: ISLETS IN ETHIONINE PANCREATITIS 275

was noted. However, the gland was not In contrast to amylase, the insulin content
densely fibrotic or scarred as seen in the was well preserved after ethionine, averaging
duct-ligated chronic pancreatitis model. 54,100 + 23,300 milliunits per gland. This
Weight of the glands averaged 5.1 g, which value was comparable to that previously
is only about 8-128 of the weight expected found in normal glands where the average
in normal dogs of this size [ 101. The ducts content of insulin was 53,800 -t 24,300 mil-
were still widely patent and easily cannulated. liunits per gland. After cell preparation, an
The tissue amylase values confirmed acinar averageof 45% of this insulin was still present
atrophy. The average amylase was 1749 dye in the cell suspension, indicating a substantial
units per unprocessedgland, which represents recovery of available islet tissue [ 191.
less than 0.3% of the amylase found in a Figure 7 shows insulin output curves for
normal gland (Table 1). Isolation of cells by perifused islets from three of the cell prepa-
ductal perfusion resulted in a further reduc- rations in this series. These islets show a
tion in amylase, as had previously been noted prompt insulin output in response to glucose
in ductal perfusion of normal pancreas [9]. with a prolonged second-phaseresponse.This
Only 8.5% of the extractable amylase in the pattern is indistinguishable from that of islets
ethionine-treated gland remained in the cell prepared from normal pancreata, as seen in
preparation after processing (Table 1). In one Fig. 6.
animal in this series, no amylase was detect- The apancreatic dog which received the
able in the cell suspension. intrasplenic transplant of ethionine-treated

FIG. 2. Canine pancreas after full-dose d,Z-ethionine treatment. Islet tissue remains but it is difficult to
distinguish from degenerating acinar cells within fibrotic lobules. Hematoxylin and eosin. X250
276 JOURNAL OF SURGICAL RESEARCH: VOL. 38, NO. 3, MARCH 1985

FIG. 3. Low-power view of d,Z-ethionine-treatedpancreas treated with immunoperoxidase stain specific

for insulin. Darkly stained areas are intact islets. These are morphologically similar to islets seen in the
normal untreated pancreas. X 100

islets became normoglycemic postoperatively pancreatitis [3]. The histological pattern is

and insulin was discontinued. The islets were somewhat like nesidioblastosis although hy-
promptly rejected, however, and diabetes re- perinsulinism is not associated. The islet
curred 5 days after the graft. proliferation was most striking in the dog
The histology of these glands after ethio- that received a 2-week course of six doses of
nine-treated glands deserves comment. The ethionine and then was allowed a 3-week
principal changeswere degeneration of acinar recovery period before pancreatectomy. Fig-
cells with loss of zymogen granules. However, ure 5 is a photomicrograph of a hematoxylin
we observed a marked amount of closely and eosin-stained section of this pancreas.
packed intact islet tissue. These cells were While acinar tissue had regenerated to repro-
seenin groups much larger than normal islets duce a normal appearance (amylase levels
as well as in small groups of cells surrounding and pancreas weight in this animal were
small ducts (Figs. 2, 3, 4). The spreading of almost normal), marked islet cell proliferation
thesecells along the ducts as well as occasional was also seen, especially around small ducts.
mitotic figures seen in islet cells most likely In addition, a mitotic figure in an islet cell is
represents proliferation of islet cells. This seen in this photograph.
kind of islet proliferation has been noted DISCUSSION
previously by us in the duct-ligated pancre-
atitis model [lo], and it has also been reported QEthionine was first observed to be se-
as mimicking neoplasia in human chronic lectively toxic to pancreatic acinar cells by
 COBB ET AL.: ISLETS IN ETHIONINE PANCREATITIS 277

FIG. 4. Higher-power view of single islet stained for insulin in ethionine-treated pancreas. The
surrounding matrix is composed of atrophic acinar tissue. X450

Goldberg et al. [8] and by Farber and Popper intracellular activation of zymogens and con-
[4] in 1950. This toxic effect was further sequent cellular destruction [22].
found to be potentiated by a choline-deficient Ethionine has largely been studied in rats.
diet [ 14, 15, 18, 201 and reduced by supple- However, Almeida and Grossman [l] used
mentary dietary methionine [4] and by d,l- dogs, cats, and monkeys in primary patho-
p-3-thienylalanine, a suppressor of zymogen logical and toxicological studies. In dogs,
granule formation [ 181. Early attention was pancreatic damage followed oral, intraperi-
focused on the probable role of substitution toneal, and subcutaneous administration at
of ethionine for methionine in proteins pro- a dose of 0.1 g/kg daily. Severe toxicity was
duced by the acinar cell [8, 251, or in the seen, including decreased activity, lethargy,
induction of ATP deficiency [24]; however, and weakness. After 5 to 6 days of these
these effects, while demonstrable, were not regimens, no food was taken voluntarily and
sufficient to explain the selective toxic ef- the animals were force fed by tube. Many
fect on the pancreatic acinar cells [16]. Re- demonstrated prolongation of the Lee-White
cent evidence implicates the interference of clotting time with bloody diarrhea and oral
ethionine in the role of methionine in trans- hemorrhage. Toxicity generally proved fatal
methylation reactions in the biosynthetic in 10 to 15 days. However, in our protocol,
pathways of phospholipids. Membrane defects while approximately the same dose was used,
in the enzyme-containing secretory granules it was given lessfrequently and toxic problems
or lysosomes, or both, could then lead to the were greatly diminished. Jaundice and loss
278 JOURNAL OF SURGICAL RESEARCH: VOL. 38, NO. 3, MARCH 1985

FIG. 5. Apparent islet proliferation in a dog which received six doses of d,l-ethionine over 2 weeks
followed by a 3-week recovery period. Acinar tissue is present, although not prominent, in this view. A
mitotic figure is seen in the lower midportion of the photograph within an islet of Iangerhans. In this
specimen, mitotic figures were common. X250

of appetite and some weight loss, however, 250

I
were usually seen during the last week of the
drug course.

0, I I I 1 I
01
03 10 15 30 60
TIME IminI
TIME (minutes)
FIG. 6. Insulin output by normal canine islets in an
in vitro perifusion system. These data represent the FIG. 7. Insulin output from three groups of ethionine-
average output of 21 different petiftions; error bars treated islets tested by in vitro perifusion. Insulin is
represent standard deviations. A prompt and sustained released by all three in response to glucose stimulation
insulin releaseis seen in responseto glucose stimulus. in a pattern similar to normal.
 COBB ET AL.: ISLETS IN ETHIONINE PANCREATITIS 279

It is important to give ethionine for at Previously published evidence of islet hy-

least 3 weeks. Shorter courseswere less sat- perplasia occurred in settings of clear and
isfactory for inducing complete acinar atro- specific&cell damage.However, both in the
phy. However, these animals must be ob- present model and in the duct-ligatedchronic
served closely during their final doses, as the pancreatitis model, islet damage is at most
toxic symptoms can manifest themselves rap indirect; finding evidence for islet hyperplasia
idly as the schedule is completed. If the in these settings suggests that hyperplasia
animals become listless and totally lose their may be a response of the islet to diverse
appetites during the final week of ethionine injurious conditions, including selective
administration, pancreatectomy should be damage to the acinar pancreas [23].
performed immediately. The ethionine model conhrms our previous
Timing of pancreatectomy in relation to observation in the normal dog pancreas that
the final dose of ethionine also appears to be digestion of the pancreas by ductal perfusion
important in minimizing acinar contamina- with collagenase results in a substantial en-
tion. In dogs which were operated on within richment of insulin content with loss of
2 days of their final ethionine dose, the amylase and, by implication, acinar tissue.
amylase content of the tissue and cell prep- The recovery in the cell preparation of an
arations was very small. After a 5day recov- averageof 45% of insulin present in the gland
ery period, amylase content increased sub- compared favorably with the 7% recovery in
stantially. In one animal which received only the duct-ligated pancreatitis model [lo] which
six doses and was allowed a 2 1-day recovery is another tactic leading to a significant de-
period, amylase content of the tissue ap crease in tissue amylase content. Ethionine
proached normal values. The tendency of treatment is easier and quicker than ductal
the exocrine pancreas to recover rapidly fol- ligation in producing acinar atrophy. The
lowing cessation of ethionine administration fact that tissue insulin is virtually the same
has been noted by other investigators [2, 51. in normal and ethionine-treated pancreata
In the ethionine-treated animals, and par- follows the observation by Payne et al. [21]
ticularly in those allowed a recovery period, in rats. Ethionine treatment preserves the
we observed islet cell proliferation. These islets and their insulin content despite the
proliferating islets were seenespecially around nearly complete elimination of acinar en-
small ducts suggesting a possible origin from zymes. Our perifusion results and allografbng
a multipotentiaI reservecell. This corresponds trial indicate that a good functional capability
with our previous findings in the duct-ligated can be obtained in vitro and in vivo from
pancreatitis model which also demonstrated cells processedin this manner.
islet proliferation [lo]. In summary, this study demonstrates that
Islet cell regeneration has been observed an intravenous course of ethionine can serve
in other models. Logothetopoulos [ 12, 131 to eliminate the acinar enzyme content of
injected mice with a guinea pig anti-insulin the dog pancreas while preserving insulin
antibody to induce islet damage. Histologic content. Islet cell function in preparations
and radioautographic study of these pancreata derived from these pancreata is normal when
clearly showed proliferation in the islets. In studied in vivo and in vitro. This method can
a study of juvenile-onset diabetics, Gepts et serve to provide a more abundant source of
al. [6] also observed islet hyperplasia. Most relatively uncontaminated islets from a large
of the patients in this study died early in animal for in vitro studies as well as for in
their course. Pathologic sections showed rib- vivo transplantation.
bon-like sheetsof cells similar to our findings;
Gepts histochemically identified these as al- ACKNOWLEDGMENTS
most exclusively composed of pancreatic The authors gratefully acknowledgethe expert technical
polypeptide cells. assistance
of Marlene Douglas, assistance in photomi-
280 JOURNALOF SURGICALRESEARCH:
VOL.38,NO.3,MARCH 1985

crogmphy by James Taskett, and the secretarialassistance fed with a cholinedeficient diet. Amer. J. Pathol.
of Peggi Polen. 79:465, 1975.
16. Lognecker, D. S., Farber, E., and Shull, K. H. The
REFERENCES biochemical pathology of ethionine-induced pan-
creatic damage: Ethionine incorporation into pro-
1. Almeida, A. L., and Grossman, M. I. Experimental teins, and ATP levels in pancreas of rats. Arch.
production of pancreatitis with ethionine. Gasfre B&hem. Biophys. 127: 601, 1968.
enterology 2: 544, 1952. 17. Mendez-Picon, G., and McGe.orge,M. Effect of total
2. Alvizouri, M., and Warren, S. Effectsof d,l-ethionine lymphoid irradiation and pretransplant blood trans-
on the pancreaseand other organs. Arch. Pathol. 57: fusion on pancreatic islet allogratt survival. J. Surg.
130, 1954. Res. 34: 421, 1983.
3. Bartow, S. A., Mukai, K., and Rosai, J. Pseudoneo- 18. Monto, G. L., and Hruban, Z. Prevention of ethio-
plastic proliferation of endocrine cells in pancreatic nine pancreatitis in the rat by d,f-beta-3-thienylalan-
hbrosis. Cancer 47: 2627, 1981. ine, a suppressor of zymogen granule formation.
4. Farber, E., and Popper, H. Production of acute Digestions 5: 210, 1972.
pancreatitis with ethionine and its prevention by 19. Morgan, C. R., and Lazarow, A. Immunoassay of
methionine. Proc. Sot. Exp. Biol. Med. 74: 838, insulin: Two antibody system. Diabetes 12: 115,
1950. 1962.
5. Fitzgerald, P. J., and Alvizomi, M. Rapid restitution 20. Negro, P., Risetti, A., Ciani, R., DeBemardinis, G.,
of the rat pancreas following acinar cell necrosis Tuscano, D., and Ammicucci, G. Pancreatite acuta
subsequent to ethionine. Nature (London) 170: 929, sperimental nel topo indotta da dieta colina priva e
1952. d,l-etionina. Boll. Sot. Ital. Biol. Sper. 54: 777, 1918.
6. Gepts, W., De Mey, J., and Marichal-Pipeleers, M. 21. Payne, W. D., Sutherland, D. E. R., Mata, A. J.,
Hyperplasia of ‘pancreatic polypeptide”-cells in Gorecki, P., and Najarian, J. S. D,Iethionine treat-
the pancreas of juvenile diabetics. Diubetologiu 13: ment of adult pancreatic donors: Amelioration of
27, 1917. diabetes in multiple recipients with tissue from a
7. Goldberg, R. C., and Chaikoff, I. L. Selective pan- single donor. Ann. Surg. 189: 248, 1979.
creatic acinar destruction by d,f-ethionine. Arch. 22. Rao, K. N., Zuretti, M. F., Baccino, F. M., and
Palhot. 52: 230, 1951. Lombardi, B. Acute hemorrhagic pancreatic necrosis
8. Goldberg, R. C., Chaikoff, I. L., and Dodge, A. H. in mice: The activity of lysosomal enzymes in the
Destruction of pancreatic acinar tissue by d,f-ethio- pancreas and the liver. Amer. J. Pathol. 98: 45,
nine. Proc. Sot. Exp. Biol. Med. 74: 869, 1950. 1980.
9. Horaguchi, A., and Merrell, R. C. Preparation of 23. Rynasiewia, J. J., Sutherland, D. E. R., Kawahra,
viable islets from dogs by a new method. Diabetes K., and Najarian, J. S. Total lymphoid irradiation
30~455, 1981. for prolongation of pancreas and islet allografts
10. Horaguchi, A., Cobb, L. F., and Merrell, R. C. Islet survival in rats. Surg, Forum 31: 359, 1980.
recovery in chronic pancreatitis. J. Surg. Res. 35: 24. Shull, K. H., McConomy, J., Bogt, M., Castillo, A.,
217, 1983. and Farber, E. On thomechanism of hepatic aden-
11. Lacy, P. E., Finke, E. H., Conant, S., and Naher, S. osine triphosphate deficiency by ethionine. J. Biol.
Long-term perifusion of isolated rat islets in vitro. Chem. kl: 5060, 1966.
Diabetes 25: 484, 1976.
25. Simpson, M. V., Farber, E., and Tarver, H. Studies
12. Logothetopoulos, J. Islet cell regeneration and neo-
on ethionine. I. Inhibition of protein synthesis in
genesis. In Handbook of PhysioIogy. Washington,
intact animals. J. Biol. Chem. 182: 81, 1950.
D. C.: American Physiological Society, 1972. Pp.
61-76. 26. Sutherland, D. E. R., Frenzel, E., Payne, W. D.,
13. Logothetopoulos, J., and Bell, E. G. Histological and Matas, A. J., and Najarian, J. S. Transplantation of
autoradiographic studies of the is!& of mice injected adult islet tissue in rats: Spleen vs. portal vein site.
with insulin antibody. Diabetes 15: 205, 1966. Surg. Forum 30: 305, 1979.
14. Lombardi, B. Pa&genesis of ethionine induced 27. Sutherland, D. E. R., Rynasiewicz, J. J., Kawahara,
pancreatic nezrosis.Panminerva Med. lfk 359, 1976. K., Gore&i, P., and Najarian, J. S. Rejection of
15. Lombardi, B., Estes, L. W., and Lognecker, D. S. islets vs. immediately vascularized pancreatic allo-
Acute hemorrhagic pancmatitis (massive necrosis) ~l;8~antitative comparison. J. Surg. Res. 29:
with fat necrosis induced in mice by d,l-ethionine . .