Origin and Fate of Oval Cells in Dipin-Induced

American Journal ofPathology, Vol. 145, No. 2, August 1994Copynight © American Societyfor Investigative Pathology

Origin and Fate of Oval Cells in Dipin-InducedHepatocarcinogenesis in the Mouse

Valentina M. Factor*t, Svetlana A. Radaeva,and Snorri S. ThorgeirssontFrom the Laboratory of Cytology,* Institute ofDevelopmentBiology, Moscow, Russia; and Laboratory ofExperimentalCarcinogenesis,t National Cancer Institute, NationalInstitutes ofHealth, Bethesda, Maryland

We have studied the development and differentia-tion of oval ceUs in the Dipin model of hepato-carcinogenesis in the mouse and compared thisprocess togeneration ofbiliary epithelial ceUs bybile duct ligation using light and electron micros-copy. The Dipin model of hepatocarcinogenesisconsists ofa single injection ofan alkylating drug,Dipin (1,4-bis[N, N'-di(ethylene)-phosphamideJ-piperazine), folowed by partial hepatectomy.The Dipin treatment resulted in irreversible dam-age andgradual death ofhepatocytes by necrosisand apoptosis. Earlier work provided evidencethat regeneration ofparenchyma occurred viaoval ceU proliferation and subsequent differen-tiation into hepatocytes that replaced the degen-erating hepatocytes. Both autoradiographic andmorphological data indicated that oval ceUs werederivedfrom ductular ceUs ofHering canals. Thefirst oval ceUs labeled with [3H]thymidine weresimilar in size and ultrastructure to ductular ceUsofHering canals with whom intracellular connec-tions existed. The proliferation of ductular ceUsofHering canalsgave rise to a new system ofovalceU ducts that spread into the liver acinus. In theperiportal areas, the transition ofoval ceUs intohepatocytes was observed inside the ducts. Bothgrowth patterns and ultrastructure of oval ceUswere differentfrom the biliary epithelial ceUls inbile duct-ligated liver. Also, oval ceUs retained theproperty to interact with adjacent hepatocytesthrough desmosomes and intermediatejunctions.Oval ceUpopulation was heterogeneous in termsofproliferating potentiaL A proportion ofprolif-erating ceUs (38 to 45%) in the Hering canals andsmaU oval ceU ducts located in the periportal ar-eas was similarthroughout theperiod ofovalceU

development. The extent ofproliferation of ovalceUs decreasedfrom 62% at the stage of activemigration into the acinus to 22% at maximumfor-mation ofoval ceUl ducts. These data suggest thatin the mouse liver ceUs of the terminal biliaryductules harbor the hepatic stem ceU compart-mentfrom which oval ceUs, capable of differen-tiating into hepatocytes, may be derived (AmJPathol 1994, 145:409-422)

The existence of epithelial stem cells in the adultmammalian liver and their role as possible target cellsin carcinogenesis is currently receiving increasingsupport.1- This is largely due to development of newmodel systems and the use of reliable cell lineage-dependent markers9'10 in combination with new invitro11'12 and in vivo13 assays for liver epithelial celldifferentiation. However, the existence of liver stemcells still poses many unanswered questions and keyissues that still remain to be resolved include the pre-cise localization of the stem cell compartment anddefinition of their differentiation options.

The concept that stem cells are present in the adultliver is based mainly on extensive research in experi-mental rodent carcinogenesis.1-9 One of the earliestcellular responses to chemical hepatocarcinogensinvolves the proliferation of a new cell population ofnonparenchymal epithelial cells that are distinct bothmorphologically and biochemically from maturehepatocytes. These cells, commonly referred to asoval cells because of the shape of their nuclei,14 havebeen shown to be multipotent cells that can differen-tiate into hepatocytes and other cell lineages. Evi-dence indicates that oval cells are derived from cellsof terminal biliary ductules (Hering canals)14-19where the presence of the hepatic stem cells is hy-pothesized. 16 The first ductular cells lining the Heringcanals, which connect the biliary canaliculi of hepa-tocytes with the system of bile ducts, possess unique

Accepted for publication April 4, 1994.

Address reprint requests to Dr. Snorri S. Thorgeirsson, NationalCancer Institute, Building 37, Room 3C28, 9000 Rockville Pike, Be-thesda, MD 20892-0037.

409

410 Factor et alAJP August 1994, Vol. 145, No. 2

characteristics. Unlike all other cells of the biliary epi-thelial cell lineage, they interact with hepatocytesthrough a system of complex specialized contactsand belong to the smallest and structurally the mostprimitive cells in the liver.17'20-22 In the embryonicliver, structures analogous to the Hering canals areformed in the region of the future portal tracts by primi-tive hepatoblasts,23 25 which manifest the propertiesof bipotential progenitor cells.12'26'27

The possibility that oval cells arise from cells of un-clear morphology localized outside the system of bileducts has also been suggested.5'28'29 This sugges-tion is based on autoradiographic studies conductedat light and electron microscopic levels in which it was

shown that at early stages of oval cell proliferationlabel was found in small periductular cells lying freelybetween the basement membrane of the bile ductsand the hepatocytes. Our laboratory has recently re-ported that the first cells entering DNA synthesis dur-ing activation of the hepatic stem cell compartment inthe AAF/partial hepatectomy model are OV-6 anddesmin-positive oval and Ito cells, respectively, in theperiportal area.30

Recently, a novel model of experimental hepato-carcinogenesis in mice was developed that proved tobe valuable in exploring both the activation of hepaticstem cells and their differentiation.31 In this model allhepatocytes are irreversibly damaged by combina-

Figure 1. Morphology of the mouse liver at different stages of Dipin-induced hepatocarcinogenesis. A: Week 10: numerous oval cells occupy approximately one-half of the liver lobule. The oval cells are arranged mainly in a duct-like pattern. B: Week 12: a focus of small new hepatocyteswvith round nuclei and basophilic cytoplasm isformed around the portal vein. Old parenchyma and oval cells are displaced toward the center ofthe liver lobule. C: Week 15: oval cells penetrate into the central part of the hepatic lobule coming in close contact and surrounding the preexistinggiant hepatocytes with enormous malformed nuclei. New hepatocytes are considerably smaller (upper left) and have the appearance of mature ba-sophilic cells withouit signs ofdamnage or degeneration. D: Week 26: Preexisting damaged hepatocytes are replaced by newlyformed ones. Oval cellsdo not disappear completely from the tissue but remain in noticeable numbers around portal tracts and inside the parenchyma. PV, portal vein;CV, central vein; H, foci composed ofsmall new hepatocytes; thin arrows, ducts of oval cells; thick arrow, apoptotic cell. Histoplast-embedded liversections. Indirect immunoperoxidase staining with monoclonal antibody A6 counterstained with hematoxylin and eosin. Magnification: A and CX200, B andDXlOO.

IIq

Oval Cells in Mouse Liver 411AJP August 1994, Vol. 145, No. 2

Table 1. Characterization of the Oval Cell Response

Zone of theIntensity Number of Liver Lobule Depth of Penetration

of Oval Cells (5-p Histoplast into Liver CordReaction (%)* Section) (1-p Epon Sections) Duct Type

Weak (2 weeks) 14 Periportal 1-2 Hepatocytes Terminal biliary ductules, small(n = 2) 17 oval cell ducts

Moderate (2 to 4 weeks) 46 Periportal and 1-5 Hepatocytes Terminal biliary ductules, small(n = 4) 54 intermediate and large oval cell ducts

7272

Strong (4 to 6 weeks) 82 Periportal and 1-7 Hepatocytes Small and large oval cell ducts(n = 2) 82 intermediate

Percentage of total number of hepatic epithelial cells (bile epithelial cells, hepatocytes, and oval cells).

Figure 2. Oval cells in Epon-embedded semithin liver sections stained with toluidine blue. A: Week 2: weak oval cell response. Hering canals(shown by open arrows) and oval cell ducts (shown by asterisks) form joint ducts. B Week 4: moderate oval cell response. Note an increase in thenumber ofjoint oval cell ducts. PV, portal vein; A, artery. Magnification X 1000.

tion of a single dose of Dipin (1,4-bis[N,N'-di(ethyl-ene)-phosphamide]-piperazine) and partial hepat-ectomy. Parenchyma is recovered largely due to ex-tensive oval cell proliferation and subsequent forma-tion of numerous foci of newly formed hepatocytesthat replace the damaged and dying preexistinghepatocytes. Dipin produces many types of damagein the target cells as a result of the alkylation of DNA,RNA, and protein. The most important are the DNAlesions that result in multiple chromosome breakage,rearrangement, and cell death after mitotic stimula-tion.32'33 After 10 to 12 months, multiple liver tumorsdevelop in all experimental mice.

Previously, studies in the Dipin model provided evi-dence that oval cells can differentiate into hepato-cytes.3435 The transition of oval cells into hepato-cytes was found to occur in ducts located in theperiportal area of the liver lobule.34 We have used theDipin model to study the origin of oval cells in mouseliver and to characterize the proliferative capacity ofthese cells, depending on their localization in the liverlobule. To identify the first proliferating cells and theirprecise localization in the liver structure, we used acombination of autoradiography on a series of semi-

thin sections with subsequent electron microscopicexamination of labeled cells on the ultrathin sections.We also attempted to further verify the stem cell na-ture of oval cells by comparing the growth pattern andultrastructural characteristics of oval cells and biliaryepithelial cells proliferating after a ligation of the com-mon bile duct. The results provide evidence that ovalcells may be a progeny of a hepatic stem cell com-partment localized in the terminal biliary ductules.

Materials and Methods

Animals

Adult (2 months old) male F1 (CBA x C57BL/6) miceweighing 21 to 23 g were used in all experiments.

Experimental Models

Dipin Model

Dipin (1 ,4-bis[N,N'-di(ethylene)-phosphamide] pi-perazine), an alkylating agent, was obtained fromChemical Pharmaceutical Research Institute (Rus-

I

'-'- 1, . f- ---M. .-W,I,O '.. .l

. .L Ii'Alii.%- e- i"Ok... k il-Lb A." li. I, %a

412 Factor et alAJP Atgust 1994, Vol. 145, No. 2

Figure 3. Incorporation ofl3Hlthymidine into cells ofHering canal 2weeks after treatment (A) anid oval cells 4 weeks after treatment (B).Lutmen of Hering canal is designzated by an arrou and that of ovalcell ducts by asterisks. Epon-embedded semithinz section stained withtoluidine blue. Magtiification: A anid B x 1000.

sia). Dipin at a dose of 60 pg/g body weight was ad-ministered to mice intraperitoneally. Two hours afterDipin administration, partial hepatectomy was per-

formed. A total of 200 mice were examined. Liverspecimens were taken at weekly intervals up to 6months and then at 12 and 18 months and fixed forlight and electron microscopy.

Ligation of the Common Bile Duct

The ligation of the common bile duct was done on

24 male mice under pentobarbital anesthesia (50 mg/kg). The animals were sacrificed (6 mice per week) bycervical dislocation during the following 4 weeks.

Light Microscopy

Samples of liver tissue were fixed with 4% paraform-aldehyde and embedded in histoplast. The 5-p sec-

tions were stained with hematoxylin and eosin.

Immunohistochemistry

Indirect immunoperoxidase staining with monoclonalantibody A6 that recognizes common surface-

Figure 4. Ultrastructure of a terminal bile ductuile (Herintg canal)formed by hepatocyte (H) and three ductular cells 2 weeks after treat-ment. Inset: light microscopic autoradiography of the same area onEpon-embedded semithin section. Magnification: X 1000. A: Ductuilarcells have a round to oval nuclei and scanty cytoplasm presentinigrare profiles of rouigh endoplasmic reticuilum, well-developed Golgiapparatuis, electron denise) mitochondria, and abundant free polyr-bosomes. Ductular cells enclose lumen with microvilli and rest on abasal lamina that is intemrpted on the border with the hepatocyte.Cell labeled uwith PHithymidine designated with asterisk. B: Fragmentof (A) under larger magnification shous a desmosome (arrow) be-tween a ductular cell and hepatocyte. Itn the region of their contact abasal lamina (arrowheads) terminates. Magnification: A X 7,200, Bx 35, 000.

exposed antigens of mouse oval and biliary epithelialcells was performed as described previously on 5-phistoplast sections.35

Electron Microscopy

Liver specimens approximately 1 mm3 in size werefixed in 2.5% glutaraldehyde in 0.1 M phosphatebuffer, pH 7.4, postfixed with 1% osmium tetroxide,dehydrated, and embedded in Epon. Serial semithinsections (45 to 50) were cut from three to four piecesof each liver. A series of ultrathin sections (18 to 22)were obtained from a selected area of semithin sec-tions. The sections were stained with uranyl acetateand lead citrate and studied by a JEM electron mi-croscope.


Figure 5. Ultrastructure ofjointlyformed Hering canal and duct of oval cells 2 weeks after treatment. The Hering canal isformed by two ductularcells (1 and 2) and a hepatocyte (l) adjacent to them. Cell 2 is sharply hypertrophic and simultaneously lines the Hering canal and the duct oftheoval cells. All cells of thisjoint structure are labeled with f3H]thymidine (inset). The intercellular contacts with hepatocyte are indicated by arrows.Magnification X5800. Insert: light microscopic autoradiography of the same area on Epon-embedded semithin section. Magnification X 1000.

Autoradiography

Groups of mice (3 animals in each) received injec-tions of [3H]thymidine (21 Ci/mmol) at a dose of 1pCi/g body weight at 10, 17, 24, 31, and 38 days afterDipin treatment. Each animal was injected 16 times at6-hour intervals and then sacrificed by cervical dis-location 1-hour after the last injection. The 5-p thickhistoplast and 1-p semithin Epon sections were cov-ered with emulsion, exposed for 3 months at 4 C, de-veloped, and stained with hematoxylin and eosin andtoluidine blue, respectively. The labeling index of ovalcells was determined by counting 500 to 1500 cellson semithin sections from each animal.

Results

Morphology of Dipin-Damaged Liver

The most common histological abnormalities after theDipin administration were massive oval cell prolifera-tion, degenerative change of preexisting paren-chyma, and development of newly formed hepato-cytes. Oval cells were recognized as small epithelialcells with ovoid nuclei and scanty cytoplasm radiatingfrom portal spaces. Oval cells were also specificallystained with monoclonal antibody A6 (Figure 1).35

The first oval cells appeared in the liver 1 to 3 weeksafter the Dipin treatment and were arranged as smallducts located in close vicinity to the portal veins. By3 to 5 weeks small oval cell ducts increased in numberand spread further into the liver acinus. The maximumexpansion of the oval cell population occurred at 8 to10 weeks, at which time oval cells occupied more

than half of the hepatic lobule (Figure 1, A). During thisperiod the surrounding parenchyma showed pro-gressive histopathological lesions (Figure 1, B andC). Hepatocytes became enlarged and vacuolized.They contained large irregular nuclei with alteredchromatin pattern and high ploidy level.33'36 Both ne-crosis and apoptosis of hepatocytes occurred andfrequent mitotic figures of parenchymal cells were ob-served. At the peak of oval cell proliferation (week 8to 10), the first foci of small basophilic hepatocyteswith round nuclei appeared in the vicinity to portaltracts close to oval cells (Figure 1, B). These smallhepatocytes were distinctly different from the preex-isting enlarged hepatocytes because of the differ-ence in sizes and morphology (Figure 1, C). The smallhepatocytes proliferated actively. The size of the fociof small hepatocytes increased and preexisting dam-aged hepatocytes were gradually displaced towardthe central veins. At this time numerous ducts com-posed of oval cells penetrated into the central part ofthe hepatic lobule, coming in close contact with andsurrounding the original damaged giant hepatocytes.The preexisting damaged hepatocytes had enor-mous atypical nuclei with numerous pseudoinclu-sions and dystrophic cytoplasm (Figure 1, C). Al-though nearly complete replacement of damagedparenchyma occurred in 22 to 26 weeks, the ovalcells did not disappear completely from the tissue butremained in noticeable numbers throughout the pa-renchyma (Figure 1, D). The normal structure of theliver was not restored before development of nodularlesions that developed into multiple hepatocellular tu-mors usually 10 to 12 months after Dipin administra-tion.


Table 2. Localization of the First Oval Cells Labeled withPHiThymidine in the Dipin Model*

LabeledOval Oval

Duct Type Ductst Cellst Cellst

Terminal biliary ductules 6 15 7Ducts connected with terminal 4 15 8

biliary ductulesSmall ducts, located in 3 12 10

periportal area

* Data obtained from ultrastructural analysis.t Number of ducts, oval cells, and labeled ovals studied.

Figure 6. Connection of small oval cell duct with Hering cantal 2weeks after treatment. Insert: light microscopic autoradiograpbj' onEpon-embedded semithin section. Small duct lined bJ1 three oval cells,tuo of which (1 and 2) are labeled with /3Hlthymidine. Adjacent tothe duct are PHithymidine-labeled fibroblast (f) and unlabeledendothelial cell (e). A, B: Electron micrographs of two sectionsfrom aseries of ultrathin sections showing a connection of the duct u'ithHering canal. H, hepatocyte; arrouheads, basal lamina; thin arrous,tight junctions between oval cell and hepatocyte. Magnification: insetx1000, A andB X6800.

The development of oval cell reaction in differentanimals was asynchronous and sometimes delayedup to 1 to 2 months.

Characterization of the OvalCell Response

The detailed analysis of oval cell development was

done on a series of semithin liver sections. The ovalcell response was classified as weak, moderate, andstrong, depending on the number of oval cells and thedegree of their penetration into the hepatic lobule(Table 1). In a weak response oval cells comprised 14to 17% of the total number of cells. They formed smallducts located strictly around the portal veins and notpenetrating beyond the boundary of the first to sec-

ond hepatocyte in the plate (Figure 2, A). In moderateoval cell response the ductular oval cells expandedinto the parenchyma, reaching the fourth to sixth hep-

Figure 7. Dynamics of[-/HlthVmidine incorporation itnto cells of ter-minal and small ducts located in the region of the portal triads(hatched columns) and large ducts infiltrating the parenchyma(dark columns). Horizontal axis: mouse number(see Table 1), inten-sitv of oval cell response increasesfrom 1 (weak) to 8 (strong). Verti-cal axis: index of labeled cells, %.

atocyte in the plate and increasing in number to 46 to72% of the total cell number (Figure 2, B). At the maxi-mum response the oval cells formed an integratedsystem of branched anastomosing ducts, occupiedmore than half of the hepatic lobule, and outnum-bered the hepatocytes by four- to fivefold.

In a weak oval cell response it was evident thatsome oval cell ducts were connected with neighbor-ing Hering canals (Figure 2, A) and a structural con-nection with the Hering canals could be identified ina series of semithin sections. In a strong oval cell re-sponse it was impossible to find the "start points" ofoval cell growth.

Origin of Oval Cells

To further characterize the oval cell response, we ana-lyzed the cell types undergoing DNA synthesis at theearly stages in the Dipin model. Repeated injectionsof [3H]thymidine were needed to obtain an adequatenumber of labeled cells due to slow and asynchro-nous development of oval cells in this model. Auto-radiographs of semithin and ultrathin liver sectionswere then analyzed by light and electron microscopy,respectively.


Figure 8. Ultrastructure of oval cells liningducts located in penportal area (A) and insidetheparenchyma (B) 7 weeks after treatment. A:Oval cells have a scanty cytoplasm with a feuwsmall mitochondria and short singular cister-nae of rough endoplasmic reticulum. B: Ovalcells have more developed cytoplasm. There isan increase in the number of mitochondriaand profiles of rough endoplasmic reticulum.The cell have a well-developed Golgi apparatusand numerous polyribosomes. Basal lamina isshown by arrowheads. Inset shows one of thedesmosomes designated on B with open arrowsunder higher magnification. Magnification: in-set x58,000, A x9,600, B x 7,200.

After 4 days of [3H]thymidine administration duringweak oval cell reaction labeled cells were found in: 1)the canals of Hering, 2) small oval cell ducts bothconnected to the canals of Hering, and 3) isolated inthe region of the portal tracts (Figures 3 to 6, Table 2).Cross-sections of the canals of Hering were made upof hepatocyte and two to three ductular cells (Figure4). The mean diameter of a ductular cell and a nucleuswas 4.3 ± 0.4 and 3.2 + 0.2 p, respectively. On theapical surface all cells had microvilli. On the lateralcell surfaces near the lumen there were tight junc-tions. Proliferation of Hering canals cells resulted inthe increase of cellularity (Figure 3, A) and, in somecases, formation of the first labeled small oval cellducts that were luminally connected with the originalcanals. Figure 5 shows an example of joint structureformed by terminal biliary ductules and the first ovalcell duct. All cells of this joint structure are labeled

(Figure 5, inset). Sometimes the same cell partici-pated in the formation of both Hering canal and theoval cell ducts (Figure 5, indicated on the photographas 2).The first small oval cell ducts located in the peri-

portal areas were surrounded by the basal lamina andconsisted of two to four oval cells on cross-section.Analysis of serial ultrathin sections of three such ductsshowed that they were also linked to the canals ofHering (Figure 6). On three to four sections from theseries of ultrathin sections one or two cells of the smallducts had two apical surfaces, one turned toward thecommon lumen with a hepatocyte, and the other to-ward the lumen of the duct of oval cells. The hepa-tocyte and the ductular oval cell in contact with it hadmicrovilli on the surface turned toward the commonlumen. The lumen was isolated by tight junctions fromthe intercellular space (Figure 6, B).


Figure 9. Ultrastructure of a large duct linedby seven oval cells (1 to 7) located in the inter-mediate zone of the liver lobule 6 weeks aftertreatment. A: The duct is surrounded by abasal lamina only partly (arrowbeads). B,C:Larger magnification of cells 2 and 3. They losttheir basal lamina along the extended surfaceattacbed to a neigbboring hepatocyte (H). Notea close contiguity ofplasma membranes of ovalcells and hepatocyte and development of des-mosomes and intermediate junctions (arrows).Adjacent old hepatocyte (H) shows signs of de-generation, degranulation, and vesiculation ofrough endoplasmic reticulum and mitochon-dria swelling. Magnification: A X5,800, B,x 11,600.

The thymidine labeling was also found in hepato-cytes, endothelial cells, Kupffer's cells, and in fibro-blasts (Figure 6, inset).

Proliferating Pool of Oval Cells

The proportion (38 to 45%) of proliferating oval cellsin Hering canals and small ducts surrounding portaltracts remained similar throughout the period of for-mation of the oval cell population. The maximum value(62%) for the proliferating pool was recorded duringthe period of active migration of oval cells into theacini. At the peak of the oval cell response, when thesystem of ducts was already formed, the proportion(22%) of proliferating oval cells declined (Figure 7).

Differentiation of Oval Cells

When oval cells first appeared close to the portalveins, they formed small ducts with basal lamina and

central lumen, numerous microvilli, well-developedtight junctions, and desmosomes (Figure 8, A). Theducts consisted of two to five cells with ovoid or ir-regular nuclei containing dense scattered chromatin.Free polysomes, individual cisternae of rough endo-plasmic reticulum, a well-developed Golgi apparatus,and single mitochondria could be detected in the cy-toplasm. Half of the cell volume (14.56 ± 7.19 p3) wasoccupied by the nucleus (7.43 + 3.24 p3).As the ducts expanded into parenchyma, there

were more cells lining the ducts, with 7 to 12 cells(range 6 to 22) per cross-section. The mean cell di-ameter increased from 4 to 5 p up to 8 to 10 p. Thenucleocytoplasmic ratio, compared with the stagewhen oval cells were first detected, decreased from1.0 to 0.8 (Table 3). Inside the parenchyma some ovalcells accumulated lysosome-like structures and la-mellar bodies but their ultrastructure remained gen-erally the same (Figure 8, B). Primary cilia was foundin 13% of studied cells. The basal lamina surrounding

Oval Cells in Mouse Liver 417AJPAugust 1994, Vol. 145, No. 2

Figure 10. Ultrastructure of a small duct composed of oval cells (0)and an intermediate hepatocytic cell (1) 10 weeks after Dipin treat-ment. The oval cell has only a small amount of cytoplasm. Nucleustakes up most of the area in the cell. The oval cell is in contact withthe neighboring intermediate cell via desmosomes and tight junctions(arrows). The intermediate cell shows features characteristic of hepa-tocyte differentiation: an enlargement of a cell volume, a roundnucleus with well-defined nucleoli, numerous mitochondria, develop-ment of rough endoplasmic reticulum, large number ofglycogen ro-

settes, and organization of bile canicular structure. Magnification:X 10, 000.

the oval cell ducts was well defined. However, in manycases oval cells could loose their basal lamina alongthe extended surface attached to a hepatocyte. Des-mosomes and intermediate junctions developed be-tween oval cells and hepatocytes (Figure 9).

Gradual transition to hepatocytes was observedwithin some oval cell ducts located in the periportalareas of liver lobules (Figures 10 and 11). Cells oftransitional morphology showed rounding of cell andnucleus shape, decrease of nucleocytoplasmic ratio,and development of abundant cytoplasmic or-

ganelles. The signs of hepatocytic differentiation in-cluded centrally placed euchromatic nuclei with one

to two well-defined nucleoli, increased number andsize of mitochondria, well-developed rough endo-plasmic reticulum often arranged in parallel rows,

large number of glycogen rosettes frequently withclose relationship to the smooth endoplasmic reticu-lum, accumulation of lipid droplets, and organizationof bile canicular structures.

Morphological Changes after BileDuct Ligation

One week after ligation of the common bile duct an

increased number of bile ducts were observed in theportal tracts accompanied by the development ofportal inflammation and moderate peribiliary fibrosis.After 4 weeks, the lobular architecture was distorted

by portal-portal bridging septa-containing hyperplas-tic bile ducts. The ducts were located in the zones ofdeveloped fibrosis and did not penetrate into the pa-renchyma.

In comparison with oval cells, biliary epithelial cellsin bile duct-ligated liver were larger, had more de-veloped cytoplasm, and low nucleocytoplasmic ratio(Figure 12, Table 3). They were surrounded by thick-ened basal lamina, supported by stroma-containingfibroblasts, collagen, and inflammatory cells. Manycells showed bleb formation replacing microvilli alongpart of the lumen surface (Figure 12, B and C). Thecells varied considerably in size, shape, and densityoften in the same duct. Despite the heterogeneity ofsize and shape, the proliferating biliary epithelial cellshad similar cytomorphological characteristics andwere represented by the one cell type. Among 250examined cells none showed signs of hepatocyte dif-ferentiation.

DiscussionDuring Dipin-induced hepatocarcinogenesis in micea massive and prolonged proliferation of oval cellsoccurs. Oval cells are arranged as a system of ductsthat arise in the periportal area and spread into theparenchyma. [3H]thymidine autoradiography re-vealed that cells of Hering canals proliferate at theearly stages of oval cell development. Analysis ofsemithin and ultrathin liver sections showed thatductular cells of Hering canals and the first labeledoval cells are similar in size and ultrastructure. Sub-sequently, the oval cells proliferate and expand intothe parenchyma and form a network of connectedductal systems that retain its original connection tothe Hering canal. These data strongly suggest thatoval cells are derived from ductular cells of Heringcanals, although the possibility of periductal progeni-tor cells cannot be excluded.

The similarity of ultrastructural and antigenic char-acteristics of oval cells and biliary epithelial cells inrats has repeatedly been described.9 10,17'20'21'37-40Ductular organization of oval cells proliferating duringboth experimental and human hepatocarcinogenesishas been demonstrated by immunohistochemicalstaining for cytokeratins and monoclonal antibodiesraised against oval cells.30'35'4143 Our data are con-sistent with those from earlier studies indicating thatoval cell proliferation starts as an outgrowth of theterminal biliary tree. 17,19-21,44,45 However, proliferat-ing oval cells in the Dipin model are distinctly differentfrom proliferating bile ductular cells after bile duct li-gation. In comparison with the biliary epithelial cells,


Figure 11. Ultrastructure of a "mixed" ductcomposed of oval cells and young hepatocyteslocated at the penportal end of liver lobule 10weeks after treatment. A: Light microscopy ofEpon-embedded semithin section stained withtoluidine blue. B: Electron micrograph of thesame duct. The cells share the same lumensealed by tight junctions. C: A fragmentmarked by arrows in (B) at a higher magnifi-cation. A bile canalicular-like structure withnumerous microvilli limited by tight junctionscan be seen between oval cell (0) and adjacentyoung hepatocyte (H). PV, portal vein; an as-terisk shows mixed duct. Bars: A 10 p, B 1 p, C0.2 p.

the oval duct cells are ultrastructurally more primitiveand show a different growth pattern. Oval cells havea higher nucleocytoplasmic ratio, less developed cy-toplasm, and spread deeply into the liver acinus. Asa progeny of cells of terminal biliary ductules, ovalcells interact directly with adjacent hepatocytes byforming intermediate junctions and desmosomes.These findings are in agreement with previous studiesusing similar conditions to analyze the interaction ofhepatocytes and oval cells.45'46The most remarkable feature of the oval cells is the

capability to differentiate into the hepatocytic lineage.The pattern of hepatocyte differentiation in Dipin-treated liver is similar to that observed in other mouseand rat model systems.1' 47-52 In the Dipin model, ovalductular cells appear to mature through a series oftransitional cells displaying ultrastructural signs of

hepatocytic differentiation. It is worth noting that allthe transitional cells and the first ductular hepatocyteswere detected within the ducts located in the peri-portal area. However, morphological data based onanalysis of fixed tissue sections cannot be taken asdirect evidence. The possibility that small newlyformed hepatocytes derived from replication of pre-existing diploid hepatocytes that might escape or re-cover from Dipin-induced chromosomal damagecannot be discarded. However, given the sequenceof morphological events together with the finding ofintermediate cells and first small hepatocytes withinthe oval cell ducts, we suggest that the oval cell-hepatocyte transition exists in the Dipin model. It is ofinterest to note that differentiation of progenitor ductu-lar cells into the hepatocytes was not a massive pro-cess. In this model de novo formation of hepatocytes,


Figure 12. Ultrastructure of biliary epithelialcells 4 weeks after ligation of the common bileduct. A: Biliary epithelial cells are arrangedaround a narrow lumen. The cytoplasm exhib-its small mitochondria and numerous scatteredshort cisternae of rough endoplasmic reticu-lum. The duct is surrounded by a basal lamina(arrowheads). B,C: Change of an apical sur-face of biliary epithelial cells. Note the reduc-tion in number and size of microvilli. Somecells show bleb-like (B) formation, singlemembrane-boundprojection on the cell surfacecontaining cytoplasmic matrix and ribosomes.Magnification: A X 7200, B X 7500, C X 8500.

Table 3. Comparison ofMorphological Features of Oval Cells and Biliary Epithelial Cells in Bile Duct-Ligated Liver

Feature Oval Cells Biliary Epithelial Cells

Structure Ducts DuctsLocalization Periportal Parenchyma Zone of fibrosisNumber of studied ducts/cells 24/95 63/490 16/256Mean diameter (p ± SEM)* 4.4 ± 0.5/ 9.4 ± 1.2/ 11.5 ± 1.3/

cell/nucleit 3.4 ± 0.4 7.3 ± 1.2 7.6 ± 1.1Nucleocytoplasmic ratiot 1.0 0.8 0.4Basal lamina Discontinuous Discontinuous IntactInteraction with hepatocytes Present Present AbsentIntermediate cells, ductual hepatocytes Present Absent Absent

Cell diameter was measured on electron photographs made at the original magnification x5000.t The mean diameter of cell/nucleus in terminal biliary ductules: 4.3 ± 0.5/3.2 ± 0.2 p, nucleocytoplasmic ratio -1.0.

similar to the compensatory regeneration after partialhepatectomy, occurs mostly as a result of their ownproliferation.The proliferative activity of oval cells depends on

their localization in the hepatic lobule. In the cells lin-ing the canals of Hering and small ducts the propor-tion of proliferating cells does not change during thetime course of the oval cell development. This nu-merically small group of presumptive progenitor cells,no more than 3% of all oval cells at the maximumresponse, supported a stable growth in the popula-

tion of oval cells. The proliferative pool of oval cellsreached a maximum (62%) at the stage of active mi-gration of the cells into the acini and a minimum (22%)when formation of the anastomosing ducts of ovalcells was completed. A similar level of the proliferativepool (50 to 60%) was observed by Sell et al28 in therat liver in which hepatocarcinogenesis was inducedby long-term feeding of N-2-fluorenylacetamide in acholine-deficient diet. However, in the AAF/partialhepatectomy model in which rapid formation of ovalcells is induced, it is possible to achieve labeling of


a considerably larger number of oval cells.50 The het-erogeneity in proliferating activity of oval cells is im-portant, because it indicates the existence of sub-populations of oval cells that differ in growth potentialand differentiation and supports the concept of thepresence in the periportal region of epithelial stemcells.16 However, the influence of different hepatic mi-croenvironment on proliferative potential of oval cellscannot be ruled out.

In summary, these data strongly suggest that ovalcells originate from the cells of Hering canals and rep-resent committed precursor cells capable of differ-entiating into hepatocytes. Such a conclusion isbased on the following observations: 1) the cells ofHering canals are the first epithelial cells undergoingDNA synthesis at the early stages of oval cell acti-vation; 2) the oval cells line the ducts that are struc-turally connected with Hering canals; 3) the oval cellsare similar to ductular cells of Hering canals in theirsize and ultrastructure; 4) gradual transition to hepa-tocytes can be observed within the ducts located onlyin the periportal areas of liver lobules; 5) oval cellsform a system of branched anastomosing ducts thatspread into the liver lobule. Oval cell ducts are com-posed of primitive cell type with a high nucleocyto-plasmic ratio that retains the capability to communi-cate with adjacent hepatocytes; and 6) oval cellpopulation is heterogeneous in terms of proliferativepotential. A proportion of proliferating oval cells (38 to45%) in the Hering canals and small ducts surround-ing portal tracts remained similar throughout the pe-riod of formation of oval cell population. In the ovalcells infiltrating the parenchyma, the proportion ofproliferating cells appeared to depend on the inten-sity of the oval cell response.

In Dipin-induced hepatocarcinogenesis in micedamage to original hepatocytes is revealed by ac-cumulation of aberrant cells, development of a highdegree of polyploidy, and aneuploidy resulting ingradual cell death. The development of functional he-patic deficiency stimulate the proliferation of the re-maining hepatocytes.36 The original polyploid hepa-tocytes are capable of maintaining hepatic mass andfunction only for a limited time interval. It seems un-likely that recovery of parenchyma in the irreversiblydamaged Dipin livers is possible without stimulationof the stem cell reserve of the liver and formation ofnew hepatocytes from progenitor cells.

References1. Sell S, Dunsford HA: Evidence for the stem cell origin

of hepatocellular carcinoma and cholangiocarcinoma.Am J Pathol 1989, 134:1347-1364

2. Fausto N: Hepatocyte differentiation and liver progeni-tor cells. Curr Opinion Cell Biol 1990, 2:1036-1042

3. Marceau N: Biology of disease: cell lineages and dif-ferentiation programs in epidermal, urothelial and he-patic tissues and their neoplasms. Lab Invest 1990,63:4-20

4. Reid LM: Stem cell biology, hormone/matrix synergiesand liver differentiation. Curr Opinion Cell Biol 1990,2:121-130

5. Sell S: Is there a liver stem cell? Cancer Res 1990, 50:3811-3815

6. Sell S: The role of determined stem-cells in the cellularlineage of hepatocellular carcinoma. Int J Dev Biol1993, 37:189-201

7. Sigal SH, Brill S, Fiorino AS, Reid LM: The liver as astem cell and lineage system. Am J Physiol 1992, 263:G139-G 148

8. Thorgeirsson SS, Evarts RP: Growth and differentiationof stem cells in adult rat liver. In The Role of Cell Typesin Hepatocarcinogenesis. Edited by Sirica AF. BocaRaton, FL, CRC Press, 1992, pp 110-120

9. Fausto N, Thompson NL, Braun L: Purification and cul-ture of oval cells from rat liver. In Cell Separation:Methods and Selected Applications, vol 4. edited byPretlow TG, Pretlow TP. New York, Academic Press,1987, pp 45-77

10. Marceau N, Blouin MJ, Germain L, Noel M: Role of dif-ferent epithelial cell types in liver ontogenesis, regen-eration and neoplasia. In Vitro Cell Dev Biol 1989, 25:336-341

11. Germain L, Noel M, Gourdeau H, Marceau N: Promo-tion of growth and differentiation of rat ductular ovalcells in primary culture. Cancer Res 1988, 48:368-378

12. Germain L, Blouin MJ, Marceau N: Biliary epithelialand hepatocytic cell lineage relationships in embry-onic rat liver as determined by the differential expres-sion of cytokeratins, alpha-fetoprotein, albumin, andcell surface-exposed components. Cancer Res 1988,48:4909-4918

13. Coleman WB, Wennerberg AE, Smith GT, Grisham JW:Regulation of the differentiation of diploid and someaneuploid rat liver epithelial (stem like) cells by the he-patic microenvironment. Am J Pathol 1993, 142:1373-1382

14. Farber E: Similarities in the sequence of early histo-logical changes induced in the liver of the rat byethionine, 2-acetyl-aminofluorene, and 3'-methyl-4-dimethylaminoazobenzene. Cancer Res 1956, 16:142-149

15. Wilson JW, Leduc EH: Role of cholangioles in restora-tion of the liver of the mouse after dietary injury. JPathol Bacteriol 1958, 76:441-449

16. Grisham JW: Cell types in long-term propagable cul-tures of rat liver. Ann NY Acad Sci 1980, 349:128-137

17. Grisham JW, Hartroft WS: Morphological identificationby electron microscopy of oval cells in experimentalhepatic degeneration. Lab Invest 1961, 10:317-332


18. Grisham JW, Porta EA: Origin and fate of proliferatedhepatic ductal cells in the rat: electron microscopicand autoradiographic studies. Exp Mol Pathol 1964,3:242-261

19. Makino Y, Yamamoto K, Tsuji T: Three-dimensional ar-rangement of ductular structures formed by oval cellsduring hepatocarcinogenesis. Acta Med Okayama1988, 42:143-150

20. Schaffner F, Popper H: Electron microscopic studiesof normal and proliferated bile ductules. Am J Pathol1961, 38:393-410

21. Steiner JW, Carruthers JS: Studies on the fine struc-ture of the terminal branches of the biliary tree. I. Themorphology of normal bile canaliculi, bile pre-ductules(ducts of Hering) and bile ductules. Am J Pathol 1961,38:639-661

22. Phillips MJ, Poucell S: Modern aspects of the mor-phology of viral hepatitis. Hum Pathol 1981, 12:1060-1084

23. Shiojiri N, Lemire JM, Fausto N: Cell lineage and ovalcell progenitors in rat liver development. Cancer Res1991, 51:2611-2620

24. Shah K, Gerber MA: Development of intrahepatic bileducts in humans: possible role of laminin. Arch PatholLab Med 1990, 114:597-600

25. Van Eyken P, Sciot R, Desmet V: Intrahepatic bile ductdevelopment in the rat: a cytokeratin-immunohisto-chemical study. Lab Invest 1988, 59:52-59

26. Desmet VJ, Van Eyken P, Sciot R: Cytokeratins forprobing cell lineage relationships in developing liver.Hepatology 1990, 12:1249-1251

27. Stosiek P, Kasper M, Karsten U: Expression of cyto-keratin 19 during human liver organogenesis. Liver1990, 10:59-63

28. Sell S, Osborne K, Leffert HL: Autoradiography of"soval cells" appearing rapidly in the livers of rats fedN-2-fluorenyl-acetamide in a choline devoid diet. Car-cinogenesis 1981, 2:7-14

29. Sell S, Salman J: Light- and electron-microscopic au-toradiographic analysis of proliferating cells during theearly stages of chemical hepatocarcinogenesis in therat induced by feeding N-2-fluorenylacetamide in acholine-deficient diet. Am J Pathol 1984, 114:287-300

30, Evarts RP, Hu Z, Fujio K, Marsden ER, ThorgeirssonSS: Activation of hepatic stem cell compartment in therat: role of transforming growth factor alpha, hepato-cyte growth factor, and acidic fibroblast growth factorin early proliferation. Cell Growth Differ 1993, 4:555-561

31. Uryvaeva IV, Factor VM: Total replacement of paren-chymal liver cells induced by Dipin and partial hepa-tectomy. Bull Exp Biol Med 1988, 105:96-98

32. Factor VM, Uryvaeva IV, Sokolova AS, Chernov VA,Brosdky VY: Kinetics of cellular proliferation in regen-erating mouse liver produced with the alkylating drugDipin. Virchow Arch B Cell Pathol 1980, 33:187-191

33. Uryvaeva IV, Factor VM: Development of aberrant pol-ypoid hepatocytes following treatment with Dipin com-

bined with partial hepatectomy. Cytology 1982, 8:911-917

34. Factor VM, Radaeva SA: Oval cells-hepatocytes rela-tionships in Dipin-induced hepatocarcinogenesis inmice. Exp Toxicol Pathol 1993, 45:239-244

35. Engelhardt NV, Factor VM, Yasova AK, Poltoranina VS,Baranov VN, Lazareva MN: Common antigen ofmouse oval and biliary epithelial cells: expression onnewly formed hepatocytes. Differentiation 1990, 45:29-37

36. Factor VM, Eliseeva NA, Tamakhina AYa: The effect ofthe alkylating carcinogen Dipin on the proliferation,level of polyploidy, and micronuclei formation in apopulation of original and newly formed hepatocytes.Izvestiia Akademii Nauk Russia, 1992, 6:821-834

37. Lenzi R, Liu M, Tarsetti F, Slott PA, Alpini G, Zhai WR,Paronetto F, Lenzen R, Tavaloni N: Histogenesis of bileduct-like cells proliferating during ethionine hepato-carcinogenesis: evidence for a biliary epithelial natureof oval cells. Lab Invest 1992, 66:390-402

38. Dunsford HA, Sell S: Production of monoclonal anti-bodies to preneoplastic liver cell populations inducedby chemical carcinogens in rats and to transplantableMorris hepatomas. Cancer Res 1989, 49:4887-4893

39. Hixson DC, Faris RA, Thompson NL: An antigenic por-trait of the liver during carcinogenesis. Pathobiology1990, 58:65-77

40. Faris RA, Monfils BA, Dunsford HA, Hixson DC: Anti-genic relationship between oval cells and a subpopu-lation of hepatic foci, nodules, and carcinomas in-duced by the "resistant hepatocyte" model system.Cancer Res 1991, 51:1308-1317

41. Schmidt WN, Page DL, McKusick K, Hnilica LS: Cellspecificity of rat cytokeratin p39 during azo dye-induced hepatocarcinogenesis. Carcinogenesis 1985,6:1147-1153

42. Carthew P, Edwards RE, Hill RJ, Evans TG: Cytokera-tin expression of the rodent bile duct developing un-der normal and pathological conditions. Br J ExpPathol 1989, 70:717-725

43. Hsia CC, Evarts RP, Nakatsukasa H, Marsden ER,Thorgeirsson SS: Occurrence of oval-type cells inhepatitis B virus-associated human hepatocarcino-genesis. Hepatology 1992, 16:1327-1333

44. Dunsford HA, Maset R, Salman J, Sell S: Connectionof duct-like structures induced by a chemical hepato-carcinogen to portal bile ducts in the rat liver detectedby injection of bile ducts with a pigmented bariumgelatin medium. Am J Pathol 1985, 118:218-224

45. Novikoff PM, Ikeda T, Hixon DC, Yam A: Characteriza-tions of and interactions between bile ductule cellsand hepatocytes in early stages of rat hepatocarcino-genesis induced by ethionine. Am J Pathol 1991, 139:1351-1368

46. Spelman LH, Thompson NL, Fausto N, Miller K: Astructural analysis of gap and tight junctions in the ratduring a dietary treatment that induces oval cell prolif-eration. Am J Pathol 1986, 125:379-392


47. Iwasaki T, Dempo K, Kaneko A, Onoe T: Fluctuationof various cell populations and their character-istics during azodye carcinogenesis. GANN 1972, 63:21-30

48. Dempo K, Chisaka N, Yoshida Y, Kaneko A, Onoe T:Immunofluorescent study on a-fetoprotein-producingcells in the early stage of 3'-methyl-4-dimethylamino-benzene. Cancer Res 1975, 35:1282-1287

49. Ogawa K, Medline A, Farber E: Sequential analysis ofhepatic carcinogenesis: the comparative architectureof preneoplastic malignant, prenatal, postnatal and re-generating liver. Br J Cancer 1979, 40:782-790

50. Evarts RP, Nagy P, Nakatsukasa H, Marsden E, Thor-geirsson SS: In vivo differentiation of rat liver oval cellsinto hepatocytes. Cancer Res 1989, 49:1541-1547

51. Lemire JM, Shiojiri N, Fausto N: Oval cell proliferationand the origin of small hepatocytes in liver injury in-duced by D-galactosamine. Am J Pathol 1991, 139:535-552

52. Bennoun M, Rissel M, Engelhardt N, Guillouzo A,Briand P, Weber-Benarous A: Oval cell proliferation inearly stages of hepatocarcinogenesis in Simian virus40 large T transgenic mice. Am J Pathol 1993, 143:1326-1336

Origin and Fate of Oval Cells in Dipin-Induced

Documents

Transcript of Origin and Fate of Oval Cells in Dipin-Induced