cytotoxicity phase of diethylnitrosamine-induced hepatic neoplasia in … · (cancer research 50,...

(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 Â¡atipeswere exposed to 50 mg of diet h\ Initrosamine perliter of water for 5 Â»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 Â¡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 multiorganhistolÃ³gica! 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: GÃ³Pdh, 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 â€”40Â°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




Fig. 3. TEM of control medaka liver showing glycogen lacunae ifi) separating bands oforganelles. the perinuclear cuff of organellesincluding Golgi apparatus Â«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 (Â£).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' â€¢â€¢'Â¿HIv â€¢'â€¢*- ^ '-''

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),KÃ¼pffercells 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




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



CYTOTOXICITY OF DEN IN MEDAK.A LIVER

â€¢\>â€¢ .iv

.

200pm

Fig. 7. HRLM showing early stage of spongiosis hepatis (SH) seen after 2-wk , Â«â€¢. , ,,,Â»exposure to DEN; venule (H. Hepatocyfe remnants occupy some spaces. Tolui- F'8- *â€¢HRLM histochemica reaction for AP 5-wk after the onsÂ« 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




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.

. :â€¢

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




Fig. 12. TEM showing nuclear (A') and ER(/ /Ã•)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 ÃŸ-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




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 triglycÃ©rideretention 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




"' â€¢

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




critical to the generation of NADPH used by macrophages toproduce Superoxide aniÃ³n 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 Â¡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. (Ã©d.).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.sminuÃ¬,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: KÃ¼pfferCell 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., â€¢//.22-35, 1979.Toyomori. T., Katsuhiro, K., Mori, M.. and OnoÃ©,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




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 fromCCIÂ«-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. ClÃ©ment,B., Rescan, P-Y.. Halli-i. G., LorÃ©al.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.. MÃ¤her,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



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


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/

cytotoxicity phase of diethylnitrosamine-induced hepatic neoplasia in … · (cancer research 50,...

Documents