| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Regular Articles |
From the Department of Pathobiological Sciences,*
School
of Veterinary Medicine, University of Wisconsin-Madison, Madison,
Wisconsin; and the Division of Developmental
Biology,
Children’s Hospital Research
Foundation, University of Cincinnati, Cincinnati, Ohio
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsReferences |
|---|
Several experimental animal model systems recently have been developed
that permit detailed examination of the cellular basis of hepatic
regeneration.7
Each model requires induction of hepatocyte
injury, which produces an environment conducive to proliferation of
healthy transplanted hepatocytes. The first model involved targeting of
a urokinase-type plasminogen activator (uPA) transgene to hepatocytes
using the albumin (AL) enhancer/promoter.8,9
AL-uPA
transgenic mice displayed diffuse vacuolization within hepatocyte rough
endoplasmic reticulum that increased in severity with age and was
accompanied by sporadic hepatocyte death. Remarkably, transgene DNA was
deleted from a small fraction of hepatocytes, which then clonally
proliferated once released from the toxic effects of uPA expression.
Progeny of these cells appeared as reddish foci in the otherwise pale
transgene-expressing liver beginning at 
2 weeks of age, and by 2 to
3 months of age the hepatic parenchyma was replaced by healthy,
transgene-deficient hepatocytes.9
Furthermore, this
process of clonal hepatic repopulation could be reproduced after
splenic transplantation of healthy donor cells isolated from a normal
mouse liver into young transgenic recipients. Transplanted cells
underwent up to 14 cell doublings in recipient liver to produce small
foci of donor-derived hepatocytes that subsequently became integrated
into host parenchyma.10
These results demonstrated the
effectiveness of a crude liver cell preparation at repopulating damaged
parenchyma, a medically important finding.
This approach was extended through the use of fumarylacetoacetate hydrolase (FAH)-deficient mice. The FAH gene, which encodes an enzyme in the tyrosine catabolism pathway, was disrupted in embryonic stem cells, and these cells were used to generate mice.11 FAH-null mice died as neonates because of liver dysfunction. However, they could be rescued with 2-(2-nitro-4-trifluoro-methylbenzyol)-1,3-cyclohexanedione, which prevents the accumulation of toxic metabolites in the tyrosine catabolism pathway.11 Transplanted normal hepatocytes were able to repopulate the liver and restore FAH activity in FAH-null mice that had been removed from 2-(2-nitro-4-trifluoro-methylbenzyol)-1,3-cyclohexanedione.12 This system was used to demonstrate that the replicative capacity of adult mouse liver cells extends to at least 77 cell doublings, using serial host-to-host transplantation of a marked donor cell population.13 Thus, it seems that at least a subset of liver cells must not be subject to typical constraints on cellular longevity.14
The most recent model involves administration to rats of the pyrrolizidine alkaloid retrorsine (12,18-dihydroxysenecionan-11,16-dione), followed by partial hepatectomy. This model, like those described above, may permit near-total replacement of liver parenchyma by donor cells, but in this case in rat.15 Collectively, these three models provide experimental tools that can be used to address medically important issues relating to the replicative and differentiation potential of selected populations of liver cells.
Our objective for this study was to explore in detail both the process of focal hepatic repopulation that follows hepatocyte transplantation and the cells that can participate in this process. To accomplish this, we modified the uPA model to overcome two limitations associated with the AL-uPA transgene: first, the frequent neonatal hemorrhaging accompanying uPA-induced hypofibrinogenemia,8 and second, the need to transplant cells into very young recipients so that these donor cells could compete effectively with endogenous transgene-deficient hepatocytes during repopulation.10 We describe below the characteristics of a new uPA transgenic mouse model, and our use of that model to address 1) the engraftment frequency and kinetics of repopulation of diseased liver by donor hepatocytes, 2) the influence of donor cell age on liver repopulation, 3) the long-term fate of donor-derived parenchyma, and 4) the proliferative capacity of donor hepatocyte populations of different ploidy.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsReferences |
|---|
Mice were housed and maintained in accordance with The Guide for the Care and Use of Laboratory Animals in AAALAC-accredited facilities. All experimental procedures were approved by the Animal Care and Use Committee of the School of Veterinary Medicine, University of Wisconsin-Madison. Certain transgenic lines used in these studies have been assigned the following genetic designations: MUP-uPA line 350-2, TgN(MupPlau)1Eps; MT-nLacZ line 379-4, TgN(Mt1nLacZ)4Eps; MT-hPAP line 456-3, TgN(Mt2ALPP)3Eps; and R26-hPAP line 808-6, TgN(R26ALPP)5Eps.
Clinical Chemistry
Blood was obtained from anesthetized MUP-uPA transgenic mice by heart puncture, and plasma samples were analyzed for albumin, total protein, and alanine aminotransferase using a Vitros 250 Chemistry Analyzer (Ortho Clinical Diagnostics, Raritan, NJ).
Generation of MUP-uPA Transgenic Mice
The MUP-uPA transgene was generated by joining the mouse major
urinary protein (MUP) promoter16
to a previously
constructed genomic coding sequence from the mouse uPA gene that
carried the 3' noncoding region and polyadenylation sequence from the
human growth hormone gene8
(Figure 1)
. The parental MUP plasmid was a pUC18
derivative containing the MUP promoter sequence terminating in the
first exon at +29 and flanked by a unique NdeI site upstream
and by a KpnI site downstream within a polylinker. The
2.5-kb MUP promoter element was excised with NdeI and
KpnI, then ligated into the
NdeI/KpnI-cut plasmid vector puPA-human growth
hormone/Nde.8
The resulting plasmid, pMUP-uPA, was
digested with NdeI and NotI, and the excised
transgene DNA was microinjected into fertilized C57BL/6 or FVB strain
mouse eggs using standard methods.17
Transgenic offspring
were identified by polymerase chain reaction, using a forward probe
specific for uPA, 5'-GCGATTCTGGAGGACCGCTTATC-3', and a reverse probe
specific for human growth hormone, 5'-TTAGGACAAGGCTGGTGGGCACTG-3'.
Twenty-five µl of reaction mixture containing genomic DNA from tail
was subjected to the following conditions: 1) 92°C for 2 minutes; 2)
35 cycles of: 45 seconds at 92°C, 1 minute at 60°C, and 1 minute at
72°C; and 3) 72°C for 5 minutes. Transgene DNA displayed an
amplified product band of 300 bp on an agarose gel. To measure
transgene expression, serum from MUP-uPA transgenic mice was subjected
to zymographic analysis as described previously.8
This
method detects increases in serum uPA associated with production and
release of uPA by transgene-expressing hepatocytes.
|
The following transgene constructs used to mark donor hepatocytes were separately injected into fertilized C57BL/6 and FVB mouse eggs. The metallothionein (MT)-nLacZ transgene, encoding nuclear-localized ß-galactosidase, has been described previously.10 MT-nLacZ transgenic mice were identified by polymerase chain reaction analysis using the forward probe 5'-CAGAGCGGGTAAACTGGCTCGGATTAG-3' and the reverse probe 5'-GACACCAGACCAACTGGTAATGGTAGC-3'. Twenty-five µl of reaction mixture containing genomic DNA from tail was subjected to the following conditions: 1) 94°C for 3 minutes, 2) 35 cycles of: 30 seconds at 94°C, 1 minute at 55°C, and 1 minute for 72°C; and 3) 72°C for 7 minutes. Transgene DNA displayed an amplified band of 400 bp on an agarose gel. The human placental alkaline phosphatase (hPAP)-coding region was used to construct two additional transgenes. The first also used the MT promoter fused to hPAP. MT-hPAP transgenic mice were identified by polymerase chain reaction, using the forward probe 5'-CTGATGAATGGGAGCAGTGGTGGAATG-3' and the reverse probe 5'-GCAGACACTCTATGCCTGTGTGGAG-3' (these recognize simian virus 40 DNA, which contributes the 3' noncoding region with polyadenylation signal to the transgene). Reaction conditions were the same as used to identify MUP-uPA transgenic mice. Transgene DNA displayed an amplified band of 360 bp on an agarose gel. The second hPAP transgene used the promoter isolated from the ROSA 26 mouse insertion site (R26).18,19 R26-hPAP transgenic mice were identified by enzyme histochemistry. Tail blood was blotted onto filter paper, allowed to air-dry, then samples were heated to 65°C for 30 minutes in alkaline-phosphatase buffer (pH 9.5) containing 0.1 mol/L Tris-HCl, 0.1 mol/L NaCl, 5 mmol/L MgCl2, and incubated overnight at 37°C in alkaline-phosphatase buffer plus 0.17 mg/ml 5-bromo-4-chloro-3-indolyl phosphate (Sigma Chemical Co., St. Louis, MO). Transgenic mouse blood displayed a blue reaction product. Because hPAP is stable in paraffin-embedded tissue, this marker was used in most transplantation studies. The nuclear localized ß-galactosidase was more readily detectable in single cells, so this was used to mark donor cells when tissues were collected 3 days after transplantation.
Liver Cell Isolation and Transplantation
Hepatocytes were isolated from nLacZ- or hPAP-marked transgenic
mice to facilitate identification of donor hepatocytes in recipient
mice. Donor hepatocytes were isolated using a modified two-step
ethylenediaminetetraacetic acid/collagenase protocol.20,21
Mice were anesthetized deeply, and the liver was perfused via the
portal vein with Hanks’ buffer containing 0.5 mmol/L
ethylenediaminetetraacetic acid, then 1 mg/ml collagenase A ( #1088
793; Boehringer Mannheim, Indianapolis, IN) in Hanks’ buffer with 5
mmol/L CaCl2 and 0.27 mg/ml trypsin inhibitor
(T-9253; Sigma). Each solution was infused for 7 minutes at 5
ml/min. All solutions were kept at 37°C and bubbled continuously with
sterile filtered 100% O2. Livers then were
placed into a Petri dish containing L15 medium (Life Technologies,
Inc., Rockville, MD), and gently teased apart with forceps. The
resulting slurry was passed through a 65-µm nylon mesh (Small Parts,
Inc., Miami Lakes, FL), the filtrate was centrifuged at 4°C for 1
minute at 440 x g, and the pellet was resuspended in
cold L15. The concentration of viable large cells (hepatocytes) was
determined by trypan blue exclusion using a hemacytometer. Cells were
maintained at 4°C until transplantation. Approximately 80% of the
cells isolated using this method are hepatocytes.22
Hepatocytes were transplanted via intrasplenic injection within 4 hours of isolation.10 Recipient mice were anesthetized with Avertin (ICN Biochemicals, Plainview, NY), the spleen was exteriorized through a small left flank incision, a 26-gauge needle on a Hamilton syringe (product #81041) was introduced into the spleen, and 10 µl of the cell suspension was discharged slowly. The spleen was returned to the abdominal cavity and the incision site was closed with suture and wound clips.10 Unless otherwise noted, the total cell number transplanted into each recipient was 1 to 5 x 105. To assure histocompatibility between donor cells and recipient mice, all recipients unless otherwise noted were (C57BL/6xFVB) F1 hybrids, which accept donor cells of either parental genotype.
Detection and Quantitation of Donor Cell Hepatic Repopulation
To identify nLacZ-marked donor cells, recipient mice were administered 0.1 mg/kg cadmium intraperitoneally to induce expression of the MT-nLacZ transgene, then 16 to 20 hours later liver was collected and fixed in 4% paraformaldehyde at 4°C for 1 hour, rinsed for 90 minutes with a detergent cocktail containing 0.1 mol/L sodium phosphate buffer (pH 7.3), 2 mmol/L MgCl2, 0.01% sodium deoxycholate, and 0.0002% Nonidet P-40 (ICN Biomedicals Inc., Irvine, CA), then incubated overnight at 37°C in the dark in the detergent cocktail plus 0.16% ferricyanide, 0.21% ferrocyanide, 0.1% 5-bromo-4-chlor-3-indolyl-ß-D-galactoside (X-gal; United States Biochemical, Cleveland, OH), and 2.5% dimethyl sulfoxide.10 Transgene-expressing cells displayed a blue reaction product. For some studies, livers were frozen in OCT compound (Fisher Scientific, Pittsburgh, PA), and cryostat sections were analyzed on a slide to detect nLacZ-containing cells. To identify hPAP-marked donor cells, recipients of MT-hPAP-marked cells were given an intraperitoneal injection of 0.1 mg/kg cadmium, then 16 to 20 hours later tissue was collected and fixed at 4°C in 4% paraformaldehyde for 2 hours. R26-hPAP mice do not require pretreatment with cadmium. Samples were stained for the presence of hPAP using enzyme histochemistry as described previously for tail blood dots. All samples were stored in 70% ethanol. For some studies, fixed, paraffin-embedded, and sectioned livers were analyzed on a slide using the method described above to detect hPAP-containing cells.
To measure the percentage of hepatic surface area occupied by blue-staining donor cell foci, we used computer-assisted image analysis using a color video camera (Dage-MTI, Inc., Michigan City, IN) and Image-Pro software (Media Cybernetics, Silver Spring, MD). This analysis compared blue-stained liver surface area to total liver surface area, generating a value for percent repopulation by donor cells. Recipient livers analyzed in this way were selected from several experiments to establish a repopulation series, consisting of 14 livers with repopulation ranging from 2 to 80%. These livers were used as standards against which to determine percent surface staining of additional samples. Percent surface staining was estimated independently by two observers for most remaining livers by visual comparison with the liver standards. Estimates then were averaged. Nine samples evaluated in this manner were subsequently re-analyzed via computer-assisted analysis to determine the accuracy of this approach. Visual estimates using the standards and computer measurements differed only by 3.4 ± 2.8% (X ± SD).
To determine whether the percent of surface area occupied by donor cells reflected the total liver volume occupied by donor cells, 40 liver tissue sections from separate lobes of 10 recipient mice were mounted on slides, stained to identify donor cells, then imaged as described above to determine the percent of cross-sectional area occupied by donor-derived parenchyma. For each section this value was compared to the percent of the section perimeter occupied by blue-staining donor cells. The absolute difference between area and surface measures was 7.1 ± 6.7%. However, area values were distributed both higher and lower than surface values, so that mean donor-derived cross-sectional area was 51% and mean surface perimeter occupied by donor cells was 49% in these samples. In general, for the least repopulated quartile, mean perimeter measures tended to overestimate mean area staining (21% versus 18%), whereas for the most repopulated quartile, mean perimeter measures tended to underestimate mean area staining (73% versus 79%). Based on this data, we conclude that percent of liver surface area occupied by donor parenchyma provides a good estimate of total liver repopulation.
The cross-sectional area occupied by single blue-staining donor clones
was determined on the liver surface under low-power magnification by
measuring the major axes (a, b), then
calculating area (of an oval) using the formula A =
[1/4(a + b)]2.
To measure the fraction of transplanted cells that engrafted in liver after splenic transplantation, 11-µm cryostat sections of the left, right, and caudate liver lobes of recipient mice that had received MT-nLacZ-marked donor cells 3 days before sacrifice were stained to detect nuclear ß-galactosidase activity. The number of donor cells visible per cryostat section was counted, this number was reduced by 21% to account for nuclei that bridge adjacent tissue sections (as determined by examination of adjacent 11-µm cryostat sections from four livers), and the corrected number was divided by section volume (11 µm times section area, which was measured using computer-assisted image analysis). This value (donor cells per unit volume) was multiplied by liver mass to give donor cells per liver.
Next, the fraction of transplanted donor cells capable of expanding as
foci was determined using the method of Moore et al,23
which is reported to account specifically for the error introduced by
random cross-sectioning of foci in thin tissue sections. Cryostat
sections of the left, right, and caudate liver lobes of recipient mice
that had received MT-nLacZ-marked donor cells 14 days before sacrifice
were stained to detect ß-galactosidase activity. These sections were
used to count the number of donor foci present per unit cross-sectional
liver area and to measure the radii of donor foci using
r = 1/4(a + b), where
a and b are the major axes. These data were used
to calculate the number of foci per cm3
of liver
using the formula N =
(1/r1 +
1/r2 +
1/r3 + ... +
1/rn)/(SA), where
rn is the radius of the nth donor cell
focus in cm and SA is the surface area of the liver section
examined in cm2. Finally, focus number per
cm3
of liver was multiplied by liver weight in
grams to determine total number of foci per liver.10
This
method could not be used to determine the fraction of engrafted cells
at 3 days after transplantation because the diameters being measured
must be larger than the section thickness.
Flow Cytometric Separation of Hepatocyte Ploidy Subpopulations
Approximately 5 x 106 hepatocytes were isolated as described above, then suspended in 2 ml of L15 medium containing 10 µg/ml Hoechst dye (Sigma), 10 µg/ml propidium iodide (PI) (Sigma), 10% fetal bovine serum, and 1% insulin-transferrin-selenium-X (Life Technologies, Inc., Grand Island, NY). Cells were kept at 37°C for 15 minutes before sorting, then maintained at 4°C while sorting. Cells were sorted by a FACStarPLUS (Becton Dickinson) equipped with an argon ion Innova 90 (Coherent) laser tuned to 488 nm and a Crypton Innova 90 (Coherent) laser tuned to multiline UV. Signals for PI were logarithmically amplified and acquired through a 630/30 optical filter. Hoechst signal was amplified linearly and collected through a 360/60 optical filter. Data were acquired and analyzed using CellQuest software (Becton Dickinson, San Jose, CA). The fluorescent characteristics of the total cell population were displayed on a dot plot as height of the PI signal versus forward scatter (FSC), allowing us to select, or gate, live hepatocytes (PI is excluded from live cells). The live hepatocyte population was displayed on a dot plot of Hoechst fluorescent area, a measure of DNA content, versus Hoechst fluorescent width, a measure of relative cell size. Gates were established around three generally discrete populations of hepatocytes that differed in DNA content, and each gated population was sorted into separate tubes. Cell concentration and viability of sorted cell populations was determined using trypan blue exclusion and a hemacytometer, then cells were transplanted separately into recipient mice as described above. Viability at this stage measured between 50 and 74% for each subpopulation of cells.
Isolated nuclei from nonsorted hepatocytes and splenic lymphocytes were prepared by suspending cells in phosphate-buffered saline (without Ca2+ and Mg2+) at pH 7.4 containing 1 mmol/L ethylenediaminetetraacetic acid and 0.1% bovine serum albumin with 1 mg/ml RNase A, 33 µg/ml PI, and 0.2% Igepal for at least 15 minutes at room temperature. One half of the nonsorted hepatocyte population then was spiked with lymphocyte nuclei to provide a diploid nucleus standard. Nuclear extracts were analyzed by FACScan (Becton Dickinson) and DNA analysis was performed by ModFit LT software, version 2.0 (Verity Software House, Topsham, Maine). In two studies, to determine the purity of the cell populations after the initial sort, a second sort was performed on each population using the methods described above. Additional Hoechst dye and PI were added to each cell population before the second sort. In two studies, cells were sorted directly onto microscope slides to allow determination of the numbers of hepatocyte clumps and nonparenchymal cells present in each sorted population.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsReferences |
|---|
The mouse MUP gene is expressed in
hepatocytes, but expression does not initiate until mice are 2 to 4
weeks old. Expression of this gene also is higher in male animals,
although this difference is less pronounced in animals carrying
transgenes that use this promoter.16
We anticipated that
uPA expression would be delayed in MUP-uPA relative to AL-uPA
transgenic mice, thereby eliminating uPA-mediated neonatal lethality
and permitting more efficient repopulation by donor cells introduced
into older transgenic recipient mice. Injection of the MUP-uPA
construct into fertilized mouse eggs produced 22 FVB founder mice and
nine C57BL/6 founder mice (Table 1)
. At
birth, founder mice were indistinguishable from nontransgenic
littermates. In particular, perinatal hemorrhage was not observed.
Typically, livers in mice in MUP-uPA lineages generated in either
strain displayed a lobular pattern of lesion development by 1 month of
age, with centrilobular cells being affected most severely. Several
lineages, however, displayed diffuse lesions. One lineage of the latter
type generated in the C57BL/6 background, designated 350-2, was used in
all transplantation studies described below.
|
.
Serum total protein and albumin concentrations remained normal in
transgenic mice throughout this period (data not shown). Transgene
expression was not associated with preweaning lethality; 48 out of a
group of 102 weaned offspring of a mating between heterozygous MUP-uPA
mice and nontransgenic mice carried the transgene, close to the
expected 50%. To identify age-related changes in transgene expression,
uPA zymography was performed on serum from line 350-2 transgenic mice
(Figure 2B)
. Serum uPA concentrations were similar in transgenic and
nontransgenic mice at both 2 weeks and 9 weeks of age, but elevated in
transgenic versus nontransgenic mice at 4 weeks of age. By
comparison, uPA levels were increased at birth in AL-uPA transgenic
mice, then gradually decreased to control levels after 1 month of
age.8
|
To determine whether MUP-uPA transgenic mice could support repopulation
by transplanted healthy hepatocytes, 1 to 5 x
105
hPAP- or nLacZ-marked donor hepatocytes were
transplanted into MUP-uPA recipients that were between 2 and 7 weeks of
age. Recipient mice were sacrificed 8 weeks later, livers were stained
histochemically, and the extent of donor cell repopulation was
quantitated. As in AL-uPA mice, repopulation was highly variable among
MUP-uPA recipients regardless of gender, but ranged up to 80% in young
recipients (Figure 2C)
. Mice 5 weeks of age or older at the time of
hepatocyte transplantation displayed less extensive donor cell
repopulation (Figure 2C)
, as expected given the greater extent of
competing endogenous parenchymal repopulation present at these ages.
Finally, in a separate experiment, five recipient C57BL/6 mice
homozygous for the MUP-uPA transgene displayed 93 ± 3.4%
repopulation by donor cells, suggesting that endogenous cell
competition is reduced in homozygotes (also reported for homozygous
AL-uPA recipient mice25
).
Donor Cell Engraftment and Replication Competence
We used MUP-uPA transgenic mice to examine several biological
characteristics of hepatocyte transplantation. To measure the fraction
of transplanted cells engrafting in host liver parenchyma, 1.6 x
105
nLacZ-marked donor hepatocytes were
transplanted into spleens of 4-week-old MUP-uPA mice. Three days later,
recipients were sacrificed, livers were weighed, and cryostat sections
of liver were incubated with X-gal to identify donor cells in
parenchyma (at this stage present as singlets or adjacent doublets).
The number of blue-stained cells per section was used to determine the
total number of engrafted donor cells per liver, as described in
Methods. As presented in Table 2
, for two
experiments a mean of 21% of the cells introduced into spleen
engrafted into liver parenchyma. To determine the fraction of engrafted
cells capable of replicating and expanding into foci, livers of
additional recipients from these studies were collected at 2 weeks
after transplantation, by which time significant donor focus growth had
occurred. Similar analysis demonstrated that a mean of 6.6% of the
transplanted cells gave rise to multicellular foci, indicating that
approximately one-third of engrafted cells survived and proliferated
(Table 2)
.
|
To assess the rate of donor hepatocyte focal repopulation, we
transplanted hPAP-marked hepatocytes isolated from 2- to 4-month-old
transgenic mice into MUP-uPA recipients, then collected and
histochemically stained recipient livers at 1, 2, 4, and 8 weeks after
transplant. The surface cross-sectional area of 50 donor-derived
hepatic foci was measured for each recipient at each time point. As
illustrated in Figure 3
and Table 3
, focus expansion continued through 8
weeks after transplant. The mean focus cross-sectional area at each
time after transplant was converted into mean focus volume, and this
was used to estimate donor cell-doubling time10
and
cumulative cell doublings (Table 3)
. Mean cell-doubling time was
shortest during the first 2 weeks, corresponding to the maximal rate of
focus expansion. By 8 weeks after transplant donor cells had undergone
a mean of 12.2 cell doublings. Note that focus cross-sectional area was
variable (Figure 3)
. However, this variability could be explained by
relatively small differences in cumulative cell doublings. Also, any
focus initiating from a transplanted hepatocyte doublet could begin
with a growth advantage. Interestingly, transplanted hepatocyte foci
derived from 18- or 23-month-old donor mice displayed a rate of
increase in mean cross-sectional area similar to that measured for
young donor cells (Figure 3)
.
|
|
. In
general, margins between donor and endogenous parenchyma remained
discrete, suggesting that cellular organization in repopulated liver is
highly stable.
|
Beginning in young adulthood, average hepatocyte DNA content
gradually increases because of polyploidization.26
This is
accomplished both by development of polyploid nuclei and binuclearity.
Diploid hepatocytes become a minority. Our finding that 33% of
engrafted cells could mitotically expand as foci in recipient livers
suggested that higher ploidy classes of hepatocytes are likely to
participate in repopulation. However, we could not rule out differences
in cell survival or engraftment between classes after transplantation
of unfractionated liver cell preparations. Therefore, we sought to
determine the ability of cells from three major ploidy classes to
engraft and proliferate after transplantation into MUP-uPA transgenic
recipients. The method of cell subpopulation isolation is illustrated
in Figure 5
. In all cases, the diploid
cell population was numerically the smallest, whereas tetraploid cells
comprised the largest population (Figure 5B)
. Identity of the lowest
peak (M2) with a diploid population was confirmed by spiking a
population of nonsorted hepatocyte nuclei with splenic cell nuclei
(almost all diploid) and observing the increase in nuclei number at the
M2 peak (data not shown). The remaining subpopulations displayed two
and four times greater Hoechst fluorescent-staining intensity, as
expected for tetraploid and octaploid cell populations, respectively.
In two experiments, the isolated subpopulations were sorted a second
time to assess the accuracy of the initial sort (Figure 5C
and Table 4
). In addition, in four experiments we
microscopically examined sorted cell populations on slides to determine
the number of contaminating nonparenchymal cells or hepatocyte doublets
(Table 5)
. As illustrated in Tables 4 and 5
, subpopulations were highly enriched (>90%) for hepatocytes of the
expected ploidy. Finally, in two of these experiments the isolated
diploid, tetraploid, and octaploid hepatocyte subpopulations were
separately transplanted into MUP-uPA recipient mice (Table 6)
. Although focus number in individual
recipients varied widely, as observed previously after transplantation
of crude cell preparations, each group of sorted cells gave rise to
clonal hepatic foci. Analysis of surface cross-sectional area of
donor-derived hepatic foci revealed similarity among subpopulations for
this measure of donor cell growth potential (Table 6)
, although
nonsorted cells produced slightly larger foci (corresponding to one
additional cell doubling by this stage), perhaps associated with a
delay in initiation of repopulation by sorted cells or the presence of
cell doublets in the nonsorted cell population.
|
|
|
|
Previous studies have shown that freshly isolated or cryopreserved
transplanted hepatocytes can repopulate damaged hepatic
parenchyma10,12-15,25,27
. Our findings now demonstrate
that 1) 21% of transplanted hepatocytes engraft in liver (similar
to previous reports of 5 or 15%10,28
), and one-third of
engrafted cells retain the ability to replicate into foci; 2) cells
isolated from young adults and mice near the end of their life span
display similar repopulation kinetics, undergoing 12 cell doublings
before replication stops; 3) hepatocytes in different ploidy
subpopulations can participate in this process; and 4) the extent of
reconstituted donor parenchyma is stable as recipient mice age
(although the rate of cell turnover in donor parenchyma was not
assessed). Based on these observations, we suggest that a large
fraction of adult hepatocytes can function as clonogens, supporting at
least 12 cell doublings after transplantation into an appropriate host
environment.
The studies using sorted cells are particularly relevant to conclusions
regarding hepatocyte clonogenicity, although sorting of live
hepatocytes by ploidy is not an established methodology. We therefore
took several approaches to evaluate the composition of resulting
subpopulations. First, in these studies, resorting indicated that the
initial populations were >90% pure for the selected ploidy class
(Table 4)
, establishing the extent of reproducibility of this approach.
Second, even after sorting there can be contamination of the isolated
subpopulations by nonparenchymal cells or hepatocyte doublets (which
may be sorted as though they were singlets of higher ploidy). As
indicated in Table 5
, the extent of contamination generally is <10%,
consistent with the value for resort purity. Sorting also requires
chemical treatments, physical manipulations, and time ex
vivo, each of which could affect cell viability or regenerative
capacity.
Despite these limitations, our data support the conclusion that multiple hepatocyte ploidy classes can serve as progenitors for regenerating hepatocyte foci in damaged liver. First, in these experiments, 2 to 6% of transplanted sorted cells gave rise to foci in recipient livers. If contaminating nonparenchymal cells (generally <3%) were the source of all hepatocyte foci, then most to all of these nonparenchymal cells would have to possess stem cell-like characteristics, far higher than the expected frequency of putative nonparenchymal hepatic stem cells.29 Of course, because nonparenchymal cell contamination was present in each population, we cannot exclude the possibility that some foci were derived from these cells. Second, it also is unlikely that contaminating doublets of lower cellular ploidy (2 to 8%) generated all foci. Each subpopulation was highly enriched for cells of the appropriate ploidy, and the repopulation efficiencies of each subpopulation were similar to one another and to nonsorted cells. If only contaminating cell doublets could give rise to foci after transplantation of higher ploidy subpopulations, the resulting donor focus number would have been reduced dramatically (by 12- to 50-fold), in contrast to what we observed. Finally, mean focus cross-sectional areas were comparable for each sorted donor subpopulation, indicating similar cell-doubling capacity of progenitor cells of each type. The question remains open whether the ultimate doubling capacity of cells within each different hepatocyte ploidy subpopulation could attain more than 77 cell doublings, as reported by Overturf et al13 after serial transplantation of unfractionated liver cells, or whether only a subset of liver cells will approach immortality. However, whatever the answer to that question, our findings clearly indicate that transplantation of either crude or fractionated liver cell preparations should be adequate for most proposed therapeutic uses.
Donor cell-derived parenchyma also persisted in a panlobular distribution throughout recipient life span, indicating that maintenance of viable parenchyma does not require the contribution of a nonhepatocytic stem cell population. We note, however, that this finding cannot exclude a role for nonhepatocytic stem cells in normal liver cell turnover. There is experimental support for the existence of facultative liver stem cells in adult liver, most likely associated with Canals of Herring.29 Under conditions of hepatic injury coupled with hepatocyte mitoinhibition, stem cell progeny may proliferate and differentiate into new hepatocytes. There also have been suggestions that this cell lineage may give rise to hepatocytes in normal liver.30 However, if activated to proliferate in MUP-uPA mice, cells in this lineage could express the transgene once they initiated the program of hepatocyte differentiation. At that time, they may be eliminated by uPA-mediated toxicity, and any requirement for additional hepatocytes would be met by replication of existing hepatocytes. This outcome would preserve the donor-specific character of parenchyma in repopulated regions of the liver.
In certain ways, focal hepatocyte repopulation after hepatocyte
transplantation resembles the diffuse hepatic proliferation that
follows two-thirds partial hepatectomy. Most or all viable hepatocytes
proliferate in response to the as yet unidentified signal(s) of hepatic
insufficiency, the cell-doubling time is approximately equivalent (29
to 45 hours in this study), and proliferation stops after restoration
of normal liver mass. Similar repopulation kinetics were observed in
the retrorsine/partial hepatectomy rat model.15
There also
are several important differences. Regeneration after partial
hepatectomy requires only one to two cell doublings by remaining
hepatocytes to replace lost liver mass, and cell proliferation occurs
within the microenvironment of existing hepatic lobules. Subsequent
tissue reorganization to restore normal hepatic architecture involves
intralobular proliferation and movement of nonparenchymal cells to
reestablish appropriate spacing between hepatocytes, biliary epithelial
cells, and sinusoids. In contrast, posttransplant repopulation is focal
and, when associated with single progenitor cells, clonal. The process
starts with many fewer proliferating cells (3 x
105
transplanted hepatocytes represents 0.3%
of adult hepatocyte number), and cell proliferation continues through
12 or more cell doublings. Proliferating transplanted cells grow
focally by expansion within parenchyma, so in this sense the process
resembles micronodular hyperplasia associated with
cirrhosis.3,4
Although the consequences of these
differences are not all apparent, it is clear that focal repopulation
will require more time.31
Nevertheless, the long-term
outcome of both focal and diffuse regeneration is the same: restoration
of architecturally and functionally normal liver.
The MUP-uPA transgenic mouse model provides an efficient experimental
system with which to study hepatic repopulation. AL-uPA transgenic mice
were used originally to demonstrate that transplanted healthy donor
liver cells could focally repopulate diseased mouse liver. MUP-uPA
transgenic mice share certain characteristics with their AL-uPA
predecessors. First, hepatocyte-specific expression of uPA is
hepatotoxic. Second, the diseased liver eventually is replaced through
the expansion of small foci composed of normal-appearing parenchyma
which, in AL-uPA transgenic mice, have been shown to lack transgene
expression.9
In AL-uPA transgenic mice, transgene
expression initiates in fetal liver, producing a selective growth
advantage for healthy cells throughout early liver development.
Deletion of the AL-uPA transgene from hepatocytes can occur at any time
during this period, and is followed by clonal replication of the
transgene-deficient cells.9
Consequently, the repopulating
nodules are of widely varied size, ranging from <1 mm to almost 1 cm
in 5-week-old mice. In contrast, the selective growth advantage should
not be present until 3 weeks of age in MUP-uPA transgenic mice.
Surprisingly, this later initiation of lesion development in MUP-uPA
transgenic mice was not associated with a delay in repopulation of
parenchyma by normal-appearing hepatocytes. Instead, at this age we
observed simultaneous appearance of multiple small, uniformly sized
foci of normal-appearing hepatocytes that rapidly became confluent.
This is consistent with the presence in 3-week-old mice of a subset of
hepatocytes that lack the ability to express the transgene, and which
then begin to proliferate at the onset of transgene expression in
neighboring cells. The net effect is parenchymal replacement by
lesion-free cells that no longer produce uPA at 2 to 3 months of age in
mice carrying either transgene construct. Third, both transgenic mouse
models can support extensive although highly variable parenchymal
repopulation by healthy donor hepatocytes. There are several likely
causes of the variation in repopulation. The rate of replacement of
damaged parenchyma by endogenous healthy cells is not identical among
recipients carrying uPA transgenes. Thus, variation in the extent of
competing endogenous regenerating parenchyma can influence the extent
of donor cell repopulation. Also, during cell delivery, there can be
leakage of transplant solution from the spleen, resulting in loss of
donor cells. Injecting a small volume of solution minimizes splenic
damage and cell leakage. Other aspects of the process, including
transit of cells from spleen to liver, engraftment, and subsequent
focal expansion also may vary among recipients. However, in both
models, homozygous recipients support up to 100% repopulation by donor
cells. The MUP-uPA model does provide several important advantages
relative to AL-uPA mice. First, neonatal lethality is not observed in
MUP-uPA transgenic mice because of the later onset of hepatic transgene
expression. Second, donor hepatocytes can be transplanted into 3- to
4-week-old MUP-uPA mice, providing easier surgical manipulation and
better recipient survival (cells transplanted into younger recipients
likely remain mitotically quiescent until the initiation of transgene
expression at 3 weeks). Finally, in contrast to AL-uPA transgenic mice,
MUP-uPA transgenic mice can be maintained on an inbred background. This
expands the opportunity to examine the biology of chimeric livers under
conditions in which both host and donor genotypes can be controlled.
|
Acknowledgements |
|---|
|
Footnotes |
|---|
Supported by National Institutes of Health grants RO1-ES07671, RO1-DK49787 (to E. P. S.), and RO1-HL47826 (to J. L. D.) and grant 2 T32 ES7015-21 from the National Institute of Environmental Health Sciences (to T. C. W.).
Accepted for publication August 22, 2000.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  F. B. Engel, M. Schebesta, M. T. Duong, G. Lu, S. Ren, J. B. Madwed, H. Jiang, Y. Wang, and M. T. Keating p38 MAP kinase inhibition enables proliferation of adult mammalian cardiomyocytes Genes & Dev., May 15, 2005; 19(10): 1175 - 1187. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Strick-Marchand, S. Morosan, P. Charneau, D. Kremsdorf, and M. C. Weiss Bipotential mouse embryonic liver stem cell lines contribute to liver regeneration and differentiate as bile ducts and hepatocytes PNAS, June 1, 2004; 101(22): 8360 - 8365. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 W. C. Kisseberth and E. P. Sandgren Polyclonal Development of Mouse Mammary Preneoplastic Nodules Cancer Res., February 1, 2004; 64(3): 857 - 863. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. C. Weglarz and E. P. Sandgren Cell cross-talk mediates PPAR{alpha} null hepatocyte proliferation after peroxisome proliferator exposure Carcinogenesis, January 1, 2004; 25(1): 107 - 112. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. M. Braun, A. W. Thompson, and E. P. Sandgren Hepatic Microenvironment Affects Oval Cell Localization in Albumin-Urokinase-Type Plasminogen Activator Transgenic Mice Am. J. Pathol., January 1, 2003; 162(1): 195 - 202. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. R. Gorla, H. Malhi, and S. Gupta Polyploidy associated with oxidative injury attenuates proliferative potential of cells J. Cell Sci., March 10, 2002; 114(16): 2943 - 2951. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS |
