-
(CANCER RESEARCH 50, 5504-5514. September I. 1990]
Cytotoxicity Phase of Diethylnitrosamine-induced Hepatic
Neoplasia in Medaka1
Barrel J. Lauren, Swee J. Teh, and David E. Hinton2
Department of Medicine, School of Veterinary Medicine,
University of California, Davis, California 95616
ABSTRACT
Adult Oryzias ¡atipeswere exposed to 50 mg of diet h\
Initrosamine perliter of water for 5 »k and then transferred to
clean water for an additional15 wk. Responses of the liver during
the first 6 wk were analyzed byenzyme histochemistry and by
high-resolution light and transmissionelectron microscopy. After 1
wk, cytotoxicity was apparent at the lightmicroscopic level by
piecemeal necrosis and phagocytosis apoptosis by-
adjacent hepatocytes and resident macrophages. Spongiosis
hepatis andinflammation, found as early as wk 3, were not
widespread until wk 6.Glycogen depletion and multifocal increases
in f-glutamyl transpeptidasewere found as early as 3 wk. At 5 wk,
macrophage infiltration andaggregation and hepatocyte lysosome
proliferation were revealed by anincrease in cells staining for
acid phosphatase. In addition, a subpopula-tion of macrophages
stained positively for glucose-6-phosphate dehy-drogenase during wk
6. Other histochemical biomarkers (Mg2*-ATPase,
DT-diaphorase, uridine diphosphoglucuronyl dehydrogenase) were
notaltered. Mitotic figures were rare for the entire 6-wk period.
At theultrastructural level, necrotic alterations of some
hepatocytes were seenwithin 24 h. Within 48 h, an apparent
reduction of hepatocyte glycogenand cell volume characterized the
majority of hepatocytes; this wasaccompanied by an increase in
interhepatocytic space and the length andcomplexity of the
hepatocyte microvillous projections found in the spaceof Disse.
Lipid vacuolar inclusions inhabited space previously occupiedby
glycogen. Margins of hepatocyte nuclei were irregular, and
mitochondria were condensed and their shape altered so that
crescentric andelongated profiles were abundant. Lysosomes and
residual bodies wereincreased after 1 wk. The cytoplasmic processes
delineating spongioticlesions were identified as originating from
Ito cells. After 4 wk, apparentproliferation of smooth endoplasmic
reticulum and retention of transportlipid within its cisternae were
seen. The toxic depletion of hepatocytesand the attendant altered
cellular environment are discussed in relationto cell-to-cell
interactions and the possible contribution of stromal
andextracellular matrix changes to liver regeneration and
neoplasia.
INTRODUCTION
The medaka Oryzias ¡atipes,a teleost fish, has proven sensitive
to a number of hepatocarcinogens including dimethylni-trosamine,
DEN,3 (methylazoxy)methanol acetate, aflatoxin BI,aflatoxin d,
sterigmatocystin, o-aminoazotoluene, and
N-methyl-/V'-nitro-nitrosoguanidine (1, 2). Small size, ease of
rearing, short time to tumor formation, economy of
multiorganhistológica! examination, and lower cost per assay make
medaka a very promising adjunct to conventional assays for chemical
carcinogenesis (3). However, the medaka model has becomepopular
only within the past decade, and the endpoint of moststudies has
been restricted to tumor formation. Thus, it is notsurprising that
little attention has been paid to the cytotoxiceffects of
carcinogens on this fish species.
Understanding the cytotoxic responses of fish liver cells is
Received 11/14/89; revised 5/7/90.The costs of publication of
this article were defrayed in part by the payment
of page charges. This article must therefore be hereby marked
advertisement inaccordance with 18 U.S.C. Section 1734 solely to
indicate this fact.
1This study was funded by NIH Grant CA45131.2To whom requests
for reprints should be addressed.'The abbreviations used are: DEN,
diethylnitrosamine; GOT, 7-glutamyl
transpeptidase; ATPase, Mg!*-adenosine triphosphatase; DT.DT
diaphor-ase:quinone reducÃ-ase;AP, acid phosphatase: GóPdh,
glucose-6-phosphate dehydrogenase; UDPGdh. uridine
diphosphoglucuronyl dehydrogenase; HRLM,high-resolution light
microscopy; TEM, transmission electron microscopy; ER,endoplasmic
reticulum; GMA, glycolmethacrylate; PAS, periodic acid-Schiff;GER,
granular endoplasmic reticulum; SER, smooth endoplasmic
reticulum.
important because the cell types that are affected provide
important clues about the ability of these cells to
metabolizexenobiotics, and those cell types which repopulate the
depletedliver provide information about the origin of neoplastic
lesions.However, cytotoxicity has not received attention in
medakastudies with only minimal coverage in studies with other
fishspecies. Even when treated with such classical
mammalianhepatotoxins as CC14 and with acetaminophen, fish have
shownrelatively little hepatic necrosis (4, 5). Furthermore, when
cytotoxic responses have been described, conventional light
microscopy has been the primary technique (5, 6).
Unfortunately,this approach lacks resolution to distinguish between
differentcell types of teleost liver (7-9). Thus, information on
themetabolic role, toxic response, or proliferative capacity of
specific parenchymal and nonparenchymal liver cells is lacking
infish.
The purpose of this paper is to provide a detailed descriptionof
acute and chronic changes in medaka liver during DENexposure using
enzyme histochemistry and high-resolution lightand electron
microscopy. Subsequent papers will describe re-population of the
toxically depleted hepatic parenchymal compartment, tumor
morphology, and the spread of neoplastic cellsto other tissues.
MATERIALS AND METHODS
Fish. Medaka, reared from embryonated eggs at the Aquatic
Center(Institute of Ecology), University of California, Davis, are
consideredadults at 11 wk posthatch, when egg production is first
observed. Adultfish (n = 100) were exposed in a static system to 50
mg/liter of DENfor 5 wk, with weekly replacement of 25% of the
volume. DENconcentrations were confirmed by capillary gas
chromatography. Pairedcontrols were held under identical conditions
except for the absence ofDEN. Fish (n = 2, controls; n = 5, DEN
exposed) were collected atdaily intervals for 3 wk after the onset
of exposure and at weeklyintervals thereafter. Exposures were
conducted within a negative pressure glove box, and exhaust was
passed through a HEPA filter prior toatmospheric release. Waste
solutions were concentrated by evaporationwithin the glove box and
disposed of by Environmental Health andSafety personnel.
Tissue Preparation. Fish were anesthetized in 3-amino benzoic
acidethyl ester (50 mg/liter; Sigma, St. Louis, MO), and the liver
wasdissected free and divided into approximately equal portions.
One halfwas reserved for HRLM enzyme histochemistry, and the other
wasreserved for routine HRLM and TEM. For the former, the tissue
wasquickly quenched in 2-methyl butane in liquid nitrogen and
transferredto a freeze drier (FTS Systems, Stone Ridge, NY), where
it wasdehydrated under vacuum at —40°C.The freeze-dried tissue
was then
infiltrated and embedded in GMA for HRLM as described by
Hintonet al. (10). For routine HRLM and TEM, the other liver half
wasimmersed in one-half strength Karnovsky's fluid (11) (pH 7.4,
200
mOsmol), containing a small amount of picric acid to preserve
glycogen.The liver was then minced into 1-mm3 pieces and
transferred to vials
containing 2.5% glutaraldehyde in 0.1 M cacodylate buffer (pH
7.6).Tissues were postfixed in osmium tetroxide for TEM. Following
dehydration, tissues were infiltrated with GMA for HRLM or
withMedcast resin (Peleo, Redding, CA) for TEM. Semithin (Medcast)
andHRLM (GMA) sections were stained with toluidine blue or
hematox-ylin and eosin. In addition, some HRLM sections were
stained withPAS reagent. Thin sections were stained with lead
citrate and uranylacetate and examined with either a Phillips 400
or Zeiss Model 10transmission electron microscope.
5504
on June 19, 2021. © 1990 American Association for Cancer
Research. cancerres.aacrjournals.org Downloaded from
http://cancerres.aacrjournals.org/
-
CYTOTOXICITY OF DEN IN MEDAKA LIVER
Fig. 1. HRLM of control medaka liver showing tubules of
hepatocytes. Separated by sinusoids (5). sagittali) sectioned
tubules show bile canaliculi (BC) atcenter. Sinusoidal endothelial
cells (E) are apparent by their elongated nuclei;large pale areas
of hepatocyte cytoplasm correspond to glycogen lacunae (see Fig.2).
Nuclei of biliary epithelial cells (BEC) are also shown (see Fig.
5). H & E.
Enzyme Histochemistry. Serial HRLM sections of freeze-dried
liverwere subjected to histochemical reactions for 60 min at room
temperature and correlated with routine staining by hematoxylin and
eosin.The enzymes analyzed were ATPase (12), DT (13), GGT (14), AP
(15),G6Pdh (16), and UDPGdh (17).
RESULTS
Normal Morphology
HRLM
Basic Architecture. The hepatic parenchyma consists of rowsof
cells separated by sinusoids (Fig. 1). In longitudinal
section,these rows are 2 cells wide, but present a circular profile
of 5to 7 cells when transversely sectioned (Fig. 1). Thus,
medakahepatocytes are arranged in tubules as is also the case in
rainbowtrout Oncorhynchus mykiss (8), amphibians (18), reptiles
(19),and birds (20). Understanding this tubular pattern
facilitatedthe identification of cell types. For example, bile
preductular4
and ductular epithelial cells are centrotubular (i.e.,
surroundedby hepatocytes). Intertubular cell types include
sinusoidal endothelial cells whose abluminal plasma membranes
partially surround tubules, perisinusoidal macrophages whose
extensionsoften penetrate tubules passing between adjacent
hepatocytesto reach outer margins of bile ductular epithelial
cells, and fat-storing cells of Ito whose stellate processes form
the frameworkin which hepatocytes reside. Finally, large bile
ducts, the columnar epithelial cells of which are surrounded by a
plexus ofarterioles and capillaries, are contained within
connective tissuesheaths. These exist as a few biliary-arterial
tracts (8) and areonly partly analogous to the mammalian portal
tract (21), sincethey contain no venules.
Controls. Cytoplasm of hepatocytes contained glycogen la
cunae ranging in size from about Vt,to 3/4the cell area and
a
very few, small lipid droplets. Biliary epithelial cells were
distinguished from hepatocytes by their centrotubular location(Fig.
1), their dark staining nuclei, and their sparse cytoplasm.These
cells are very common in the tubular liver of fishes. Thenuclei of
the endothelial cells of the sinusoidal wall appearedas elongate
dark staining bodies, while the cell bodies appearedas a light
staining cuff around the sinusoid (Fig. 1).
There was no apparent zonal distribution of
histochemicalreaction products in the medaka liver. ATPase was
found insinusoidal and venular endothelium; the luminal surfaces of
thebile duct, ductule, and canaliculus; and in the connective
tissuesurrounding bile ducts. GGT was found only on the
luminalborder of bile ducts. G6Pdh, DT, and UDPGdh were faintlybut
uniformly distributed over hepatocyte cytoplasm, leavingthe nucleus
unstained. The AP reaction product was found onlyover macrophages
and hepatocyte lysosomes, both of whichwere rare in the normal
liver. Glycogen lacunae were apparentwith the PAS reaction (Fig.
2).
TEM
Differences in organelle distribution were observed betweenthe
apical (canalicular or bile preductular) and basal
(sinusoidal)hepatocyte cytoplasm (Fig. 3). Lysosomes were found
primarilyin a pericanalicular location, and vesicles budding from
the ERand Golgi apparatus appeared to fuse with the plasma membrane
at this pole. Secondary lysosomes were very common andranged from
small, roughly spherical, electron-dense bodies tolarge irregular
organelles occupied by fibrillar material of medium electron
density, alternating with homogeneous electron-dense areas. The
latter were identical to liver lysosomes ofchannel catfish
Ictalurus punctatus which reacted positively foraryl sulfatase and
acid phosphatase (22), of trout (7, 9), and ofthose found in the
supranuclear region of English sole Paro-phrys vetulus proximal
tubule cells (23). Peroxisomes weregenerally smaller, more
uniformly spherical in shape, and rel-
' These elongated biliary epithelial cells, along with
hepatocytes. line a transi
tion (bile preductule) between canaliculus and ductule. Bile
ductular epithelialcells form the entire wall of the ductule.
Fig. 2. HRLM of control medaka liver stained for glycogen (C)
using PASreaction. Note extensive dark staining areas of hepatocyte
cytoplasm whichcontrast with pale nuclei (arrows).
5505
on June 19, 2021. © 1990 American Association for Cancer
Research. cancerres.aacrjournals.org Downloaded from
http://cancerres.aacrjournals.org/
-
CYTOTOXICITY OF DEN IN MEDAKA LIVER
Fig. 3. TEM of control medaka liver showing glycogen lacunae
ifi) separating bands oforganelles. the perinuclear cuff of
organellesincluding Golgi apparatus «iti. and the apicallocation
of lysosomes (L). A biliary epithelialcell (BEC) contributes to the
bile preductule(HI'), while hepatocytes surround bile canali-
culus (AC). Endothelial cells (£).erythrocytes(Kill }, and the
space of Disse (D) are alsoshown. Inset at upper left is a higher
magnification of boxed area, upper right. P, peroxi-some; M,
mitochondrion. Lead citrate and ur-anyl acetate.
M' ••'¿HIv •'•*- ^ '-''
RB.Q-"E-r-
atively rarely seen (Fig. 3, inset). They were bounded by a
singleelectron-dense, closely applied membrane, and they were
filledwith a homogeneously dispersed, granular material of
mediumelectron density, which lacked a nucleoid core (Fig. 3,
inset).Mitochondria were numerous and, with GER, were restrictedto
the perinuclear region, or they formed thin bands extendingbetween
large glycogen lacunae. SER was rarely seen. Hepato-cyte nuclei
were round with dispersed heterochromatin and asingle nucleolus.
Hepatocytes possessed microvillous structureswhich projected into
the lumen of the bile canaliculus. Hepatocytes also formed frequent
junctional complexes with biliaryepithelial cells. These very small
cells had a high nu-clearrcytoplasmic ratio and medium electron
density over cytoplasm but they did not generally contribute
microvilli to thecanalicular system nor did they form a smooth
portion of thatstructure (Fig. 3).
The basal pole of hepatocytes and the abluminal plasmamembrane
of sinusoidal endothelial cells formed the space ofDisse (Fig. 3).
Hepatocytes contributed microvillous structuresto this space.
Although a very fine flocculent material was oftenpresent, there
was no evidence of collagen or fibrillar material.
Stellate cells of Ito were found in the perisinusoidal space
ofDisse, where they extended long processes between
endothelialcells and hepatocytes. Except at the porta hepatis and
nearlarge blood vessels, collagen fibers were not seen in the
spaceof Disse. The perisinusoidal location of medaka Ito cells
andthe presence of intermediate filaments were two of their
mostdistinguishing features. Lipid droplets were found only
infrequently. Macrophages were also occasionally found in
thisperisinusoidal location and extented between hepatocytes
toreach bile ductular epithelial cell basal poles. As in trout
(24),Küpffercells were not found.
The livers of some adult medaka showed evidence of an
activereproductive state. Two different morphologies were found.
Inmale hepatocytes, large areas of the cell were filled with
glycogen depots. In vitellogenic females, large portions of the
cytoplasm were filled with parallel arrays of GER, and much
lessglycogen was seen. A small perinuclear cuff of GER and
mitochondria surrounded the nucleus and sent extensions to
theperipheral cytoplasm as previously described in other fish
species (22). These extensions enclosed large lacunae of
glycogen,which frequently contained single or multiple lipid
vacuoles.
5506
on June 19, 2021. © 1990 American Association for Cancer
Research. cancerres.aacrjournals.org Downloaded from
http://cancerres.aacrjournals.org/
-
CYTOTOXICITY OF DEN IN MEDAKA LIVER
DEN-dependent Cytotoxic Changes
HRLM. During the first wk of exposure to 50 mg/liter ofDEN, no
changes were found at the light microscopic level.However, by wk 2,
cytotoxicity was apparent by multiple, butscattered, markedly
swollen hepatocytes. a dramatic increase inthe number of
eosinophilic, apoptotic vesicles (see Figs. 4 and6) within
hepatocytes and resident macrophages, and singlehepatocyte necrosis
(Fig. 4). No evidence for zonal necrosis wasfound. The space
between adjacent hepatocytes was enlarged,and numerous
interhepatocytic bridges could be seen (Fig. 5).Many hepatocytes
were smaller than normal and deeply stained
Fig. 4. HRLM of medaka liver after 1-wk exposure to DEN showing
apoptoticvesicles (AV) within hepatocytes. These vesicles were
strongly eosinophilic. M.macrophage. H & E.
Fig. 5. HRLM showing increased intercellular spaces (IS) between
and lipidaccumulation (LA) in hepatocytes after 2-wk exposure to
DEN. Note enlargedspace of Disse (D). Biliary epithelial cells
(BEC) are abundant. Toluidine blue.
Fig. 6. HRLM showing glycogen (G) which is depleted in
hepatocytes of thisfield after 2-wk DEN exposure (compare with Fig.
2). PAS positivity overmacrophages (A/) was not altered by diastase
pretreatment. Arrowheads point tolipid vacuoles within
hepatocytes.
(Fig. 5); glycogen lacunae were reduced (Fig. 6), and
lipiddroplets were prevalent (Fig. 5). As the exposure
continued,large lipid droplets were found (Fig. 5). Mitotic figures
wererare for the entire 6-wk period, but binucleated
hepatocyteswere relatively common after wk 2. No extravasation of
bloodand consistently normal endothelial cell structure indicated
alack of damge in this cell type. Likewise, other
nonparenchymalcell types were also not apparently affected. During
wk 3 ofexposure, multifocal aggregates of inflammatory cells were
seenin the parenchyma! compartment (see below). This correlatedwith
the appearance of small to large cystic spaces filled withan
eosinophilic flocculent material (Fig. 7). This material didnot
react with PAS or Alcian blue (pH 2.5), indicating anabsence of
neutral and acid mucopolysaccharides. Thin, verydarkly staining
strands of tissue surrounded and transversedthese spaces (Fig.
7).
Histochemical changes were not found during the first 2 wk,but
single cell and multifocal increases in GGT were commonduring wk 3
of exposure (Fig. 8). The GGT reaction productwas located primarily
in pericanalicular, apical, region of hepatocytes, but occasionally
a more diffuse reaction was seen.Macrophages in aggregates,
apoptotic vesicles, and granularbodies of hepatocytes and resident
macrophages all showedenhanced reaction products for AP (Fig. 9)
during wk 5. Inaddition, during wk 6, a subpopulation of
macrophages stainedpositively for G6Pdh (Fig. 10).
TEM. Hepatocytes were the major target of
DEN-inducedmorphological change. Individual necrotic hepatocytes
werefound within 24 h of the onset of exposure, and cell
debris,including organelles most frequently found in hepatocytes
(i.e.,mitochondria, GER, glycogen), was found in the
sinusoids.Sublethally intoxicated hepatocytes were much less
electrondense and displayed highly swollen mitochondria, ER
cisternae,and nuclear membranes, with shearing of ribosomes from
theGER. Surviving hepatocytes were either large and glycogen richor
small and glycogen poor (Fig. 11). The latter were more
5507
on June 19, 2021. © 1990 American Association for Cancer
Research. cancerres.aacrjournals.org Downloaded from
http://cancerres.aacrjournals.org/
-
CYTOTOXICITY OF DEN IN MEDAK.A LIVER
•\>• .iv
.
200pm
Fig. 7. HRLM showing early stage of spongiosis hepatis (SH) seen
after 2-wk , «•. , ,,,»exposure to DEN; venule (H. Hepatocyfe
remnants occupy some spaces. Tolui- F'8- *•HRLM histochemica
reaction for AP 5-wk after the ons« of DENdine blue exposure. The
reaction product is found over macrophage aggregates (MA) and
as granules in hepatocyte cytoplasm.
Fig. 8. HRLM histochemical reaction for GGT during wk 3 of
exposure toDEN. The reaction product (arrowheads) is locali/ed in
hcpatocyte apical plasmamembrane at the bile canaliculus (BC) in
the light gray area, and it impartsadditional grayness to the
hepatocyte within the dashed area. This comprises afocus of
GGT-reactive cells. S, sinusoid. Hematoxylin countcrstain.
electron dense than were control hepatocytes because the
gly-cogen reduction allowed the organelles to pack more
densely.Lipid droplets were initially restricted to glycogen
lacunae. Asthe exposure progressed, glycogen was nearly depleted,
andlarge lipid droplets, displacing other organelles, were
foundthroughout the cytoplasm. Except in obviously necrotic
cells,no abnormal swelling of either nuclear envelope or ER
cisternaewas found.
By Day 2, an apparent expansion of the interhepatocytic
Fig. 10. G6Pdh reaction product after 6-vvk DEN exposure. Note a
darkreaction localized over a small area of cytoplasm of a
subpopulation of macrophages (M) and a fainter reaction over larger
portions of hepatocyte (//) cytoplasm: venule (V).
spaces and the space of Disse was seen (Fig. 11).
Flocculentmaterial was seen in both spaces, but no indication of
collagendisposition was apparent. The lateral membranes of
adjacenthepatocytes were frequently interdigitated. The expansion
ofthe space of Disse was associated with elongation and branchingof
the microvillous extensions that project from the basalsurface of
hepatocytes. These processes were often vacuolatedor enclosed
myelin figures or glycogen rosettes. No changes
5508
on June 19, 2021. © 1990 American Association for Cancer
Research. cancerres.aacrjournals.org Downloaded from
http://cancerres.aacrjournals.org/
-
CYTOTOXICITY OF DEN IN MEDAKA LIVER
Fig. 11. TEM after 24-h exposure to DENshowing glycogen
depletion and an apparentconcentration of organelles (arrow) in one
population of hepatocytes and an abundance ofglycogen (G) in other
large hepatocytes. Notealso the very long microvilli (mv) of
thesehepatocytes and the increased intercellularspaces (IS) and
space of Disse (D). Lysosomes(/.) are abundant. Lead citrate and
uranyl acetate.
. :•
were found in the apical microvilli lining bile canalicoli.After
1 wk of exposure, large intrahepatocytic vacuoles (het-
erolysosomes) containing remnants of hepatocyte organelleswere
frequently found (Fig. 12). These appeared prior to thetime at
which eosinophillic bodies were seen by light microscopy, but
correspond to the same apoptotic vesicles. Increasednumbers of
smaller lysosomes (Fig. 11) and residual bodies(Fig. 14) were also
found. Nuclear profiles were irregular,nucleoli were condensed, and
in many hepatocytes, 2 nucleiwere present. A variety of
mitochondria! alterations was seen,including elongated, curled, or
swollen forms and those in whichthe intercristal matrix was either
electron dense (condensed) orabsent (high amplitude swelling).
Mitochondria were frequentlysurrounded by GER. In some instances,
electron-dense inclusions or intramitochondrial myelin figures were
found.
Cystic spaces, equivalent in area to that occupied by
severalhepatocytes, were found after 3 wk (Fig. 13). These spaces
weresurrounded by the stellate processes of Ito cells and
probablyrepresent areas once occupied by hepatocytes; occasionally
theycontained hepatocyte organelles. A fine flocculent,
osmiophilicmaterial filled this space and was associated with an
increase
in the presence of cytoplasmic vesicles in the surrounding
Itocells. Macrophages were sometimes seen at the periphery ofthese
spaces, and remnants of hepatocytes, with their characteristic
organelles, were visible within their cytoplasm. Smallcystic
spaces, equivalent to less than one cell volume, surrounded the
bile canaliculus, and some swelling of biliary epithelial cells was
found. There was no evidence of endothelialcell damage.
During wk 4, the cisternae of hepatocyte ER were found tobe
swollen and filled with an electron dense, paniculate substance
(Fig. 14). Apparent proliferation of SER was also foundin the same
cells, but in some profiles, ribosomes could be seenattached to the
outer surface of the membrane. The SER wasshown to be continuous
with the dense-staining ER. In thesecells, no perinuclear cuff of
GER and few Golgi stacks wereseen. No additional ultrastructural
changes were found duringwk 5 and 6.
DISCUSSION
DEN exposure produced cytotoxic changes in medaka liverwith
hepatocytes as, by far, the major cellular target. However,
5509
on June 19, 2021. © 1990 American Association for Cancer
Research. cancerres.aacrjournals.org Downloaded from
http://cancerres.aacrjournals.org/
-
CYTOTOXICITY OF DEN IN MEDAKA LIVER
Fig. 12. TEM showing nuclear (A') and ER(/ /Õ)remnants within
an apoptotic vesicle(A V) of an hepatocyte; a perisinusoidal
macrophage (A/) is also shown. Lead citrate anduranyl acetate. RBC
indicates erythrocytes.
as in previous studies of the cytochrome P-450-dependent he-
patotoxins, CC14, and acetaminophen (4), no repeating patternof
necrosis was found associated with afferent or efferent elements of
the vascular system. Since CC14, acetaminophen, andDEN must be
metabolized to exert their toxicity, we wouldexpect toxicity to be
limited by cytochrome P-450 activity,which is relatively low in
fish compared with mammals. Wyllie(25) has suggested that, at
relatively low toxicant concentrations, piecemeal necrosis (as
found in the present study) resultsfrom the stimulation of
endogenous lethal mechanisms in cellsalready primed to undergo
apoptosis; focal necrosis occurs onlyat higher toxicant
concentrations. Indeed, when the alcoholdehydrogenase-dependent
hepatotoxin, allyl formate, is administered to trout, a repeating
pattern of necrosis is seen (4). Thus,if P-450 activity was
uniformly high in hepatocytes, we mightexpect to see loci of
contiguous, necrotic hepatocytes. Thepresence of randomly scattered
necrotic cells implied cellularheterogeneity, but overall low P-450
activity.
The first signs of cytotoxicity were apparent only at
theultrastructural level. Sublethal effects were most apparent
bythe elongation of the microvillous projections which form
cell-
to-cell connections between adjacent hepatocytes and
endothe-lial cells. This was particularly obvious in the space of
Disse,suggesting that a reduction in hepatocyte volume had
occurred.Wester and Canton (26) and Ishikawa et al. (1) have
illustratedsimilar cytotoxic effects with ß-hexachlorocyclohexane
andDEN, respectively. Ogawa et al. (27) and Toyomori et al.
(28)have shown that preneoplastic and neoplastic hepatocytes aswell
as hepatocytes regenerating after partial hepatectomy
showsignificant dilatation of the intercellular spaces and
increasedabundance of microvilli in rats. They suggested that this
was aresponse to the increased bile load imposed on each
hepatocyte.However, since glycogen usually occupies a major portion
ofthe fish hepatocyte, we believe that the reduction in
glycogenwith DEN toxicity may be a major factor in reduction
ofhepatocyte volume. This also accounts for the apparent increasein
mitochondria! abundance in intoxicated cells. Bannasch
(29)suggested that glycogen depletion is a nonspecific response
tocell injury in mammals. Apparently, despite the slower metabolic
conversion rates and greater glycogen stores in fish hepatocytes,
this may also be a general toxic response in medaka.
Changes in glycogen and lipid content were detected at
both5510
on June 19, 2021. © 1990 American Association for Cancer
Research. cancerres.aacrjournals.org Downloaded from
http://cancerres.aacrjournals.org/
-
CYTOTOXICITY OF DEN IN MEDAKA LIVER
SH
Fig. 13. TEM showing spongiotic spaces(SH) surrounded by
stellate projections of Itocells (1C), and flocculent nature of
spongioticmaterial (right). Lead citrate and uranyl acetate.
the light and electron microscopic levels. Such changes
werefound as early as 1 day after the onset of DEN exposure,
andthey corroborate the findings of Hinton et al. (10).
Bannasch(29) suggested glycogen storage as a characteristic of
preneo-plastic hepatocytes in mammals. Given the preexposure
morphology of the medaka hepatocyte, clear cell or
eosinophillicfoci might be expected after DEN exposure. However, as
introut (30), such lesions are not commonly seen in medaka.Glycogen
depletion was followed by accumulation of lipid droplets and, in
some cells, by retention of lipid within the ERcisternae. The
near-total absence of ribosomes on the outerleaflet of the nuclear
and ER membranes suggested prior shearing of these organdÃ-es. Such
shearing could have inhibitedapoprotein synthesis, leading to
impaired lipid transport. Theabsence of Golgi-derived secretory
vesicles in the same cellsimplies that more distal steps in lipid
transport within hepatocytes were also compromised. Little is known
of the mechanisms leading to either triglycérideretention or fatty
degeneration in fish liver, but such effects likely reflect
cytotoxicityrather than genotoxicity per se. We found occasional
enlargement of hepatocytes (10), but no evidence of megalocytosis,
ashas been reported in flatfish from polluted harbors (31) and
in
trout exposed to pyrrolizidine alkaloids (32).Alterations in
enzyme histochemical reaction products have
been used to distinguish presumptively transformed cells earlyin
carcinogenesis (33, 34). Focal histochemical alteration precedes
detection by conventional hematoxylin and eosin tinctorial changes.
However, toxic change and cellular hyperplasia ofregeneration may
also alter tinctorial and enzyme changes. Ofthe 6 histochemical
reactions performed on sections during thecytotoxic phase, only GGT
was enhanced above control levelsin parenchyma! cells. During
exposure to many toxicants, cellular stores of glutathione in both
fish and mammals are depleted. Since GGT serves the dual roles of
catalyzing the uptakeof glutathione from the bloodstream into
hepatocytes, as wellas cleaving the glutamic acid moiety from
glutathione conjugates so that the more water-soluble mercapturic
acid conjugate
can be excreted, induction of this enzyme could represent
anadaptive response to DEN toxification. Although the productof DEN
metabolism, the alkyldiazonium ion, does react withglutathione, it
is so reactive that it binds indiscriminantly toany nucleophile
present (35). Thus, the aqueous bath exposureand the slow rate of
DEN metabolism (36) probably explainwhy Fong et al. (37) found that
liver glutathione in rainbow
5511
on June 19, 2021. © 1990 American Association for Cancer
Research. cancerres.aacrjournals.org Downloaded from
http://cancerres.aacrjournals.org/
-
CYTOTOXICITY OF DEN IN MEDAKA LIVER
"' •
Fig. 14. TEM showing lipid accumulation(LA) in large vacuoles
and ER cisternae (IT).Note lipid-filled ER is continuous (arrow)
withapparent SER (MV) and that some ER hasattached ribosomes (r);
residual body (RB).Lead citrate and uranyl acetate.
trout is not depleted during DEN exposure; a similar
responsewould be expected in the present experiments. On the
otherhand, normal cell division also requires ample
glutathionestores, and GGT is induced following partial hepatectomy
inmammals (38). Thus, induction of GGT could be the firstindication
of an increase in the rate of cell division by survivingor
resistant hepatocytes. Since these hepatocytes are only asmall
subset of the total liver cell population, such focal changesare
not likely to be detected by biochemical methods
usinghomogenates.
Within hepatocytes there were a number of changes to
organ-elles. The most interesting of these changes was the
obviousincrease in the number of binucleated hepatocytes.
Similarresults were reported by Ishikawa et al. (1) in
DEN-exposedmedaka and by Braunbeck et al. (39) with zebrafish
Brachydaniorerio exposed to 4-nitrophenol. Although approximately
5% of
control rat hepatocytes are binucleated, they are rare in
controlmedaka. Since we saw very few mitotic figures in these
livers,binucleated cells may represent the failure of hepatocytes
tocomplete normal cell division, rather than a stimulation of
celldivision above normal rates.
At least two types of phagocytic processes were found innecrotic
areas after DEN exposure. In the early stages ofcytotoxicity,
hepatocyte necrosis was associated with a highincidence of
apoptosis by resident cells including hepatocytesand perisinusoidal
macrophages. Apoptosis, a normal processfor regulating cell
turnover without inflammation (40), mayalso occur in tissue
subjected to injury by toxic substances (25)and appears to us as an
adaptive response early in DEN-inducedhepatocellular toxicity.
Later in the cytotoxic phase, pleo-morphic macrophages were found
enveloping hepatocytes. Theincreased presence of inflammatory cells
seen within the hepaticparenchyma at 5 and 6 wk may also indicate
an enhancedrelease of chemotactic factors by hepatocytes when
damageexceeded the ability of hepatocytes or resident macrophages
tophagocytose necrotic cells through apoptosis. The histochemi-cal
labeling of these cells for lysosomal hydrolase (AP) appearsto be a
very useful technique for evaluating the intensity of thecellular
immune response in fish. We also found a subpopula-tion of
macrophages staining positively for G6Pdh. These weresmaller,
contained no phagocytosed material, and reacted differently from
macrophages of aggregates. Since this enzyme is
5512
on June 19, 2021. © 1990 American Association for Cancer
Research. cancerres.aacrjournals.org Downloaded from
http://cancerres.aacrjournals.org/
-
CYTOTOXICITY OF DEN IN MEDAKA LIVER
critical to the generation of NADPH used by macrophages
toproduce Superoxide anión and H2O2, this may indicate a
sub-population of active or newly recruited cells as opposed
toolder, exhausted cells containing abundant residual debris.
Inaddition, clusters of thrombocytes were found within the
sinusoids. As there was no concurrent stasis of erythrocytes in
thesefish, this suggests a chemotactic response to damaged
endothe-lial cells. Endothelial cell damage has been reported in
ratstreated with DEN (41), but we found no apparent damage ateither
the HRLM or TEM level. This is surprising since wehave recently
shown inducible cytochrome P 450E in the en-dothelium of the trout
gill (42) and similar enzyme activity inliver sinusoidal
endothelium.5
Cell death is of great importance in that it provides
thestimulus for sublethally injured and/or genetically
transformedstem cells to divide and grow. Thus, the extensive
DEN-inducedhepatocyte necrosis might be thought of as a chemical
equivalent to partial hepatectomy. However, the selective lethal
tox-icity to parenchymal cells apparently permits the survival
ofonly some hepatocytes, but all or most biliary epithelial,
endo-thelial, and Ito cells. Therefore, the stromal component
isrelatively enriched in the DEN-depleted liver.
The selective death of hepatocytes also resulted in the
formation of cystic spaces, surrounded and partitioned by processes
of Ito cells. This lesion in medaka liver resembles spon-giosis
hepatis described by Bannasch et al. (43) in mammalstreated with
yV-nitrosomorpholine. This cystic space containsacid
mucopolysaccharides and is rich in proteins such as collagen and
fibronectin, both of which are normally produced bymesenchymally
derived cells, such as endothelial and Ito cells(44, 45). However,
there are some differences between thelesions formed in fish and
rats. In rats, spongiotic lesions werefound after a long induction
time (without necrosis) or lagperiod (long after necrosis). Thus,
Bannasch et al. (43) did notconsider spongiosis hepatis to result
from hepatocellular necrosis. In medaka, on the other hand, these
lesions appear duringexposure and are associated with, first,
hepatocellular necrosisand apoptosis, and later, with inflammation.
It has recentlybeen shown that CC14 induces macrophages and
Kiipffer cellsto release a factor which stimulates fibroblasts and
Ito cells toproduce collagen in rats (46). In the present study,
ultrastructural examination of the Ito cells surrounding these
spacesrevealed the presence of many apparently exocytotic
vesicles,such as those formed during the secretion of laminin, or
collagen (47). We believe that these spaces may provide the
spacefor the growth of transformed cell clones. Furthermore,
thismatrix may also provide an optimal environment for survivaland
growth of the remaining hepatocytes in vivo, as it does invitro
(48, 49).
ACKNOWLEDGMENTS
We wish to thank Dr. C. S. Giam, Texas A&M University,
Galveston,TX, for confirmation of the DEN exposure
concentrations.
We gratefully acknowledge the timely assistance of L. Heinson
inIllustration Services, University of California, Davis.
REFERENCES
1. Ishikawa, T., Shimamine, T., and Takayama, S. Histologie and
electronmicroscopy observations on diethylnitrosamine-induced
hepatomas in smallaquarium fish (Oryzias ¡atipes).J. Nati. Cancer
Inst., 55: 909-916. 1975.
2. Hatanaka, J., Doke, N., Harada, T., Aikawa, T.. and Enomoto.
M. Usefulnessof screening for the toxicity and carcinogenicity of
chemicals in medaka,
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.5Unpublished observation, this laboratory.
Oryzias latipes. Jpn. J. Exp. Med., 52: 243-253, 1982.Hoover, K.
L. (éd.).Use of small fish species in carcinogenicity testing.
Nati.Cancer Inst. Monogr., 65: 1-409, 1984.Droy, B. F. Effect of
Reference Hepatotoxins on Rainbow Trout Liver. Ph.D.Thesis, p. 169.
Morgantown, WV: West Virginia University Medical
Center,1988.Stanton, M. F. Diethylnitrosamine-induced hepatic
degeneration and neoplasia in the aquarium fish. Brachydanio rerio.
J. Nati. Cancer Inst., 34: 117-130, 1965.Gingerich, W. H.. Weber.
L. J., and Larson, R. E. The effect of carbontetrachloride on
hepatic accumulation, metabolism, and biliary excretion
ofsulfobromophthalein in rainbow trout. Toxicol. Appi. Pharmacol.,
43: 159-167, 1978.Hampton, J. A., McCuskey, P. A., McCuskey, R. S.,
and Hinton, D. E.Functional units in rainbow trout (Salmo
gairdneri) liver. I. Arrangementand histochemical properties of
hepatocytes. Anat. Rec.,2/.?: 166-175, 1985.Hampton, J. A., Lantz,
R. C., Goldblatt, P. J., Lauren, D. J., and Hinton,D. E. Functional
units in rainbow trout (Salmogairdneri) liver. II. The
biliarysystem. Anat. Ree., 221:619-634. 1988.Hampton, J. A., Lantz,
R. C., and Hinton, D. E. Functional units in rainbowtrout (Salmo
gairdneri) liver. III. Morphometric analysis of
parenchyma.sminuì,and component cell types. Am. J. Anat., 185:
58-73, 1989.Hinton, D. E., Couch, J. A., Teh, S. J., and Courtney,
L. A. Cytologicalchanges during progression of neoplasia in
selected fish species. Aquat.Toxicol., //.-77-112, 1988.Ito, S.,
and Karnovsky, M. J. Formaldehyde-glutaraldehyde fixatives
containing trinitro compounds. J. Cell Biol., 39: 168A-169A,
1968.Wachstein, M., and Meisel, E. Enzymatic histochemistry of
ethionine induced liver cirrhosis and hepatoma. J. Histochem.
Cytochem.. 7: 189-201,1959.Pretlow, T. P., Grane, R. W., Goehring,
P. L., Lapinsky, A. S., and Pretlow,T. G., II. Examination of
enzyme-altered foci with gamma-glutamyl trans-peptidase, aldehyde
dehydrogenase, glucose-6-phosphate dehydrogenase, andother markers
in methacrylate-embedded liver. Lab. Invest., 56:
96-100,1987.Rutenberg, A. M., Kim, H., Fishbein, L. S., Hanker, H.
I., Wasserkrug, H.L., and Seligman, A. M. Histochemical and
ultrastructural demonstration ofgamma-glutamyl transpeptidase
activity. J. Histochem. Cytochem., 17: 517-526, 1969.Burstone, M.
S., and Weisburger, E. K. New diazonium compounds ascoupling agents
in the detection of phosphatases. J. Histochem.
Cytochem.,9:301-303, 1961.Reider, H., Teutsch, H. F., and Sasse, D.
NADP-dependent dehydrogenasesin rat liver parenchyma. I.
Methodological studies on the qualitative histochemistry of G6PDH,
PGDH, malic enzyme, and ICDH. Histochemistry,50:283-298,1978.Lojda,
Z., Ploeg, M., and Van der Duijn, P. Phosphates of the naphthol
ASseries in the quantitative determination of alkaline and acid
phosphataseactivities "in situ"studies in polyacrylamide membrane
model system and bycytospectrophotometry. Histochemistry, //:
13-32, 1967.Goldblatt, P. J., Hampton, J. A., DiDio, L. N., Skeel,
K. A., and Klaunig, J.E. Morphologic and histochemical analysis of
the newt (Notophthalmusviridescens) liver. Anat. Ree., 217:
328-338, 1987.Storch, V., Braunbeck. T., and Waitkuwait, W-E. The
liver of the WestAfrican crocodile Osteolaemus tetrapis. An
ultrastructural study. J. Submi-crosc. Cytol. Pathol., 21: 317-327,
1989.Overton, J., and Meyer, R. Aspects of liver and gut
development in the chick.In: G. C. Schoenwolf (ed.). Scanning
Electron Microscopy Studies of Em-bryogenesis, pp. 293-302. O'Hare,
IL: Scanning Electron Microscopy, Inc.,
1984.Grisham, J. W. Cell types in long-term propagable cultures
of rat liver. Ann.NY Acad. Sci., 349: 128-137, 1980.Hinton, D. E.,
and Pool, C. R. Ultrastructure of the liver in channel
catfishIctalurus punctatus (Rafinesque). J. Fish Biol., 8: 209-220,
1976.Hickman. C. P., and Trump, B. J. The kidney. In: W. S. Hoar
and D. C.Randall (eds.). Fish Physiology. Vol. 1, pp. 91-239. New
York: AcademicPress, 1969.McCuskey, P. A., McCuskey, R. S., and
Hinton, D. E. Electron microscopyof cells of the hepatic sinusoids
in rainbow trout (Salmo gairdneri). In: A.Kirn, D. L. Hook, and E.
Wisse (eds.). Cells of the Hepatic Sinusoid, Vol.1, pp. 489-494.
Leiden: KüpfferCell Foundation. 1986.Wyllie, A. H. Apoptosis: cell
death under homeostatic control. Arch. Toxicol.Suppl., /A-3-10,
1987.
Wester, P. W., and Canton, J. H. Histopathological study of
Oryzias latipes(medaka) after long-term (J-hexachlorocyclohexane
exposure. Aquat. Toxicol., 9: 21-45, 1986.Ogawa, K., Mediine, A.,
and Farber, E. Sequential analysis of hepaticcarcinogenesis: a
comparative study of the ultrastructure of preneoplastic,malignant,
prenatal, postnatal, and regenerating liver. Lab. Invest.,
•//.22-35, 1979.Toyomori. T., Katsuhiro, K., Mori, M.. and
Onoé,T. Ultrastructural changesin the bile canaliculi and lateral
surfaces of rat hepatocytes during restorativeproliferation.
Virchows Arch. [Pathol.], 42: 201-211, 1983.Bannasch, P. The
cytoplasm of hepatocytes during carcinogenesis. Electronand light
microscopic investigations of nitrosomorpholine-intoxicated
ratliver. Recent Results Cancer Res., 19: 1-100, 1968.Hendricks. J.
D., Meyers, T. R., and Shelton, D. W. Histological progressionof
hepatic neoplasia in rainbow trout (Salmo gairdneri). Use of small
fish
5513
on June 19, 2021. © 1990 American Association for Cancer
Research. cancerres.aacrjournals.org Downloaded from
http://cancerres.aacrjournals.org/
-
CYTOTOXICITY OF DEN IN MEDAKA LIVER
species in carcinogenicity testing. Nail. Cancer Inst. Monogr.,
65: 321-336,1984.
31. Murchelano. R. A., and Wolke. R. E. Epizootic carcinoma in
the winterflounder Pseudopleuronectes americanas. Science (Wash.
DC), 228:587-589,1985.
32. Hendricks. J. D., Sinnhuber, R. O., Henderson, M. C., and
Buhler. D. R.Liver and kidney pathology in rainbow trout (Salmo
gairdneri) exposed todietary pyrrolizidine (Senecio) alkaloids.
Exp. Mol. Pathol., 35: 170-183,1981.
33. Dannaseli. P., Moore, M. A., Klimek, F., and Zerban, H.
Biological markersof preneoplastic foci and neoplastic nodules in
rodent liver. Toxicol. Pathol.,10: 19-34. 1982.
34. Peraino. C.. Staffeidt, E. F., Carnes. B. A.. I ink-man V.
A.. Blomquist. J.A., and Vesselinovitch. S. D. Characterization of
histochemically detectablealtered hepatocyte foci and their
relationship to hepatic tumorigenesis in ratstreated once with
diethylnitrosamine or benzo(a)pyrene within one day afterbirth.
Cancer Res.. 44: 3340-3347. 1984.
35. Prasanna. H. R., Raj. H. G.. and Lotlikar. P. D. Lack of
effect of glutathioneon the binding of dimethylnitrosamine to DNA
in vitro. Mol. Toxicol., 1:167-176, 1987.
36. Monti-sano. R., Ingram, A. J.. and Magee. P. N. Metabolism
of dimethylnitrosamine by amphibians and fish in vitro.
Experientia, 15: 599-601, 1973.
37. Fong, A. T., Hendricks, J. D., Dashwood. R. H., Van Winkle.
S., Lee, B. C.,and Bailey, G. S. Modulation of
diethylnitrosamine-induced hepatocarcino-genesis and
O'-ethylguanine formation in rainbow trout by
indole-3-carbinol,d-naphthoflavone, and Aroclor 1254. Toxicol.
Appi. Pharmacol., 96: 93-100, 1988.
38. Bannasch, P. Preneoplastic lesions as end points in
carcinogenicity testing.1. Hepatic preneoplasia. Carcinogenesis
(Lond.), 7: 689-695. 1986.
39. Braunbeck, T., Storch, V.. and Nagel. R. Sex-specific
reaction of liverultrastructure in zebra fish (Brachydanio rerio)
after prolonged sublethal
expousre to 4-nitrophenol. Aquat. Toxicol., 14: 185-202,
1989.40. Wyllie, A. H., Kerr, J. F. R., and Cumi, A. R. Cell death:
the significance of
apoptosis. Int. Rev. Cytol.. 68: 251-306, 1980.41. Stenbeck. F.,
Ala-Kokko, I... and Ryhanen, L. Morphological, immunohis-
tochemical, and ultrastructural changes in diethylnitrosamine
induced liverinjury. Effect of malotilate. Histol. Histopathol., 4:
95-104, 1989.
42. Miller, M. R., Hinton, D. E., and Stegeman. J. J. Cytochrome
P-450Einduction and localization in gill pillar (endothelial) cells
of scup and rainbowtrout. Aquat. Toxicol., 14: 307-322. 1989.
43. Bannasch, P.. Bloch, M., and Zerban, H. Spongiosis hepatis.
specific changesof the perisinusoidal liver cells induced in rats
by A'-nitrosomorpholine. Lab.Invest., 44: 252-264, 1981.
44. Armendariz-Borunda, J.. Greenwel. P.. and Rojkind. M.
Kiipffer cells fromCCI«-treatedrat livers induce skin fibroblast
and liver fat-storing cell proliferation in culture. Matrix, 9:
150-158. 1989.
45. Darnell. J. E.. Lodish. H. F., and Baltimore, D. Molecular
Cell Biology. NewYork: W. H. Freeman Co., 1986.
46. Goulet, F., Normand, C., and Morin. O. Cellular interactions
promote tissue-specific function, biomatrix deposition, and
junctional communication ofprimary cultured hepatocytes.
Hepatology, 8: 1010-1018, 1988.
47. Clément,B., Rescan, P-Y.. Halli-i. G., Loréal.O., Lehry,
D., Campion, J-P.,and Guillouzo. A. Hepatocytes may produce
laminili in fibrotic liver and inprimary culture. Hepatology, 8:
794-803, 1988.
48. Rojkind, M., Gatmaitan, Z., Mackensen, S., Giambrone, M-A.,
Ponce, P.,and Reid, L. M. Connective tissue biomatrix: its
isolation and utilization forlong-term cultures of normal rat
hepatocytes. J. Cell Biol., 87: 255-263,1980.
49. Bisseil. D. M., Arenson. D. M.. Mäher,J. J.. and Roll, F.
J. Support ofcultured hepatocytes by a laminin-rich gel. Evidence
for a functionally significant subendothelial matrix in normal rat
liver. J. Clin. Invest., 79: 801-812. 1987.
5514
on June 19, 2021. © 1990 American Association for Cancer
Research. cancerres.aacrjournals.org Downloaded from
http://cancerres.aacrjournals.org/
-
1990;50:5504-5514. Cancer Res Darrel J. Laurén, Swee J. Teh and
David E. Hinton Neoplasia in MedakaCytotoxicity Phase of
Diethylnitrosamine-induced Hepatic
Updated version
http://cancerres.aacrjournals.org/content/50/17/5504
Access the most recent version of this article at:
E-mail alerts related to this article or journal.Sign up to
receive free email-alerts
Subscriptions
Reprints and
[email protected] at
To order reprints of this article or to subscribe to the
journal, contact the AACR Publications
Permissions
Rightslink site. Click on "Request Permissions" which will take
you to the Copyright Clearance Center's (CCC)
.http://cancerres.aacrjournals.org/content/50/17/5504To request
permission to re-use all or part of this article, use this link
on June 19, 2021. © 1990 American Association for Cancer
Research. cancerres.aacrjournals.org Downloaded from
http://cancerres.aacrjournals.org/content/50/17/5504http://cancerres.aacrjournals.org/cgi/alertsmailto:[email protected]://cancerres.aacrjournals.org/content/50/17/5504http://cancerres.aacrjournals.org/