PHENOTYPIC REVERSION OF RAT NEOPLASTIC LIVER NODULES IS UNDER
Maria R. DE MIGLIO, Maria M. SIMILE, Maria R. MURONI, Diego F. CALVISI, Patrizia VIRDIS, Giuseppina ASARA, Maddalena FRAU,
Giovanni M. BOSINCO, Maria A. SEDDAIU, Lucia DAINO, Francesco FEO*and Rosa M. PASCALE
Division of Experimental Pathology and Oncology, Department of Biomedical Sciences, University of Sassari, Sassari, Italy
Low DNA synthesis and high redifferentiation (remodel-
ing) characterize neoplastic nodules induced by chemical
carcinogens in hybrid BFF1 rats, generated by crossing the
susceptible F344 and resistant BN strains. We performed
whole-genome scanning of BFF2 rats to identify loci control-
ling remodeling of nodules induced, 32 weeks after initiation
with diethylnitrosamine, by the RH protocol. Remodeling
nodules were identified as areas lacking uniformity of GST-P
immunostaining and with irregular margins. Two loci in sug-
gestive linkage with the percentage of remodeling nodules
were identified on chromosomes 7 and 1 (LOD scores 3.85
and 2.9 at D7Rat25 and D1Mgh14). Significant dosage-nega-
tive effect of the B allele on remodeling and additive inter-
action between these loci were found. Significant epistatic
interactions, showing a recessive, remodeling-enhancing ef-
fect of B alleles, occurred between D1Mit3 and D11Rat11
(corrected p ? 0.0013) and between D6Rat14 and D8Rat46
(corrected p ? 0.028). These data show that remodeling of
neoplastic nodules during rat hepatocarcinogenesis is under
genetic control. Loci affecting remodeling map to chromo-
somal regions syntenic to chromosomal segments of human
HCC showing structural abnormalities.
© 2003 Wiley-Liss, Inc.
Key words: phenotypic reversion; quantitative trait loci; hepatocar-
cinogenesis; neoplastic nodule; linkage analysis; genomic scanning
Individual risk of liver cancer reflects the amount of exposure to
environmental agents, such as hepatitis B and hepatitis C viruses,
aflatoxin B1, ethanol consumption, etc., combined with one’s
genetic predisposition.1–4Owing to the difficulty of studying the
genetic mechanisms of human liver cancer, interspecies compari-
son has been proposed to identify susceptibility and resistance
alleles responsible for polygenic control of the predisposition to
human HCC. Previous studies on a murine model of hepatocarci-
nogenesis have led to the identification of 7 hepatocarcinogenesis
susceptibility loci (Hcs1–7) and 2 resistance loci (Hcr1 and Hcr2)
in crosses between susceptible and resistant strains.5–8Moreover,
2 susceptibility loci (hepatocarcinogenesis in female, Hcf1 and
Hcf2) abrogate the inhibitory effect of ovarian hormones on liver
tumorigenesis in mice.9Study of the genetic control of rat HCC in
the hybrid strain BFF1, generated in our laboratory by crossing the
BN, resistant, to the F344, susceptible, strains,10has shown dom-
inant inheritance of resistance to hepatocarcinogenesis. Linkage
analysis of BFF1 ? F344 backcross and BFF2 intercross rats11,12
revealed the presence of 4 Hcs loci (rat Hcs1–4) and 7 Hcr loci (rat
Hcr1–7). Hcs3 and Hcr2 have also been identified, and named
Dhr1 and Drh2, in intercrosses between the (F344 ? DRH)F1
rats.13A putative suppressor gene (rcc?), mapping to chromo-
some 12p, critical for determining the sensitivity of rats to dieth-
ylnitrosamine-induced liver carcinogenesis, has been identified in
MHC-recombinant rat ACP strains, congenic for the MHC-linked
growth reproduction complex (grc) region.14
Clonal expansion of initiated cells in rat liver carcinogenesis is
characterized by the progressive development of foci of preneo-
plastic hepatocytes and neoplastic nodules exhibiting fast growth
and a number of morphologic, biochemical and molecular com-
monalties with preneoplastic and neoplastic human liver le-
sions.4,15,16Most early preneoplastic and neoplastic liver lesions
undergo progressive decreases in biochemical marker expression
and growth rate and finally regress (remodeling, phenotypic rever-
sion17–22). Relatively few persistent lesions are thought to be
precursors of HCCs. Redifferentiation of neoplastic lesions was
considered a basis for remodeling,18,21and enhanced probability of
apoptosis in the lesions with low expression of biochemical marker
was reported.23Neoplastic liver lesions, induced in rat strains
genetically resistant to hepatocarcinogenesis, such as Wistar, BN,
Copenhagen and DRH, exhibit low growth rate and/or high re-
modeling.11,22,24–26The growth rate of neoplastic hepatocytes is
under the control of Hcs and Hcr loci.24,27However, the genes and
molecular mechanisms responsible for remodeling are unknown.
Mapping of the genes regulating remodeling may contribute to our
understanding of the mechanisms of neoplastic lesion regression
and the genetically transmitted resistance to hepatocarcinogenesis.
In the present work, we performed whole-genome scanning of
BFF2 intercrosses to identify and map the QTL controlling neo-
plastic nodule remodeling and to understand their mechanism of
MATERIAL AND METHODS
Animals and treatments
F344 and BN rats (140–160 g at the beginning of the experi-
ment; Charles-River, Calco, Italy) were crossed to generate BFF1
and BFF2 rats. Animals were fed a standard diet (type 48; Piccioni,
Milan, Italy) and tap water ad libitum and housed individually in
suspended wire-bottomed cages in a room with constant temper-
ature (22°C) and humidity (55%) and with a 12 hr light (6 AM–6
PM)/dark cycle. Neoplastic nodules were induced in 20 F344, 20
BN, 20 BFF1 and 126 intercross rats by the RH model,28which
included initiation with a necrogenic dose of diethylnitrosamine
(150 mg/kg) followed, after repair, by a 15-day feeding of a
hyperprotein diet(type 52,
2-acetylaminofluorene, with partial hepatectomy at the midpoint of
this feeding. All animals received humane care, and the study
protocols were in compliance with our institution’s guidelines for
use of laboratory animals.
Piccioni) containing 0.02%
Abbreviations: BN, brown Norway; F344, Fisher 344; GST-P, glutathi-
one S-transferase, placental form; HCC, hepatocellular carcinoma; Hcr,
hepatocarcinogenesis resistance; Hcs, hepatocarcinogenesis susceptibility;
N, mumber/cm3of lesions; QTL, quantitative trait loci; R, percentage of
remodeling lesions; RH, resistant hepatocyte; TK test, Tukey-Kramer test;
V, mean volume of lesions.
Grant sponsor: Associazione Italiana Ricerche sul Cancro; Grant spon-
sor: Ministero dell’Istruzione; Grant sponsor: Universit` a e Ricerca; Grant
sponsor: Assessorato Igiene e Sanit` a RAS.
*Correspondence to: Division of Experimental Pathology and Oncology,
Department of Biomedical Sciences, University of Sassari, Via P. Manzella
4, 07100 Sassari, Italy. Fax: ?39-079-228485 or ?39-079-228305.
Received 29 July 2002; Revised 12 November 2002; Accepted 3 January
Int. J. Cancer: 105, 70–75 (2003)
© 2003 Wiley-Liss, Inc.
Publication of the International Union Against Cancer
Phenotyping of parental and BFF2 rats
All rats were killed under ether anesthesia 32 weeks after
initiation. Livers were resected, and small pieces of liver were
fixed in buffered formaldehyde (pH 7) and processed for hema-
toxylin and eosin staining and GST-P immunohistochemistry as
reported.10Computer-assisted morphometric analysis was carried
out by 2 independent pathologists (and by a third one in case of
discrepancies higher than 5%) on 13 slices per liver, taken from
each lobe and covering at least a 550 mm2liver surface area, to
calculate the number/cm2of liver and the surface area of nodules,
using QWIN Leica (Cambridge, UK) software. N and V values of
lesions were determined according to Pugh et al.29Transections
with radius ?35 ?m were reliably identified and included in the
analysis. Remodeling lesions were identified in GST-P-immuno-
stained sections as areas lacking uniformity of immunostaining
and exhibiting irregular margins.19,20,22Lesions were classified as
remodeling nodules when at least 20% of their surface was not
immunostained. In dubious cases, a comparison between GST-P-
and hematoxylin and eosin–stained serial sections allowed evalu-
ation of the area of nodule portions negative for GST-P immuno-
staining (and thus of the area of the whole nodule) since loss of
GST-P expression precedes the disappearance of hematoxylin and
Genomic DNA from spleens of BFF2 rats was extracted from
isolated nuclei, purified and precipitated by the Extragen kit (Tal-
ent, Trieste, Italy) in an automated DNA extractor (Talent). Geno-
typing was performed using 179 polymorphic microsatellite mark-
ers (http://ratmap.gen.gu.se). Markers (Roche, Monza, Italy) were
distributed throughout the rat genome (all autosomes and the X
chromosome), leaving gaps of 30.2, 28 and 33 cM on chromo-
somes 1, 3 and 12, respectively. PCRs included 50 ng of genomic
DNA; 20 mM (NH4)2SO4; 75 mM TRIS-HCl (pH 9); 1.5, 2 or 3
mM MgCl2; 3 pmol of each primer; dNTPs (200 ?M, each); and
0.2 U of Taq DNA Polymerase (Roche). Cycling parameters were
5 min at 96°C; 30 sec at 94°C; 30 sec at 53°, 55° or 57°C; 30 sec
at 72°C for 35 cycles; and final extension for 5 min at 72°C in a
Robocycler Gradient 96 PCR System (Eppendorf, Milan, Italy).
PCR products were run on 3.5% agarose gels to distinguish ho-
mozygous from heterozygous BFF2 rats. When markers polymor-
phic for only 2–8 bases were used, the PCR system was supple-
mented with 1–4 ?M fluorescently labeled primers. Aliquots of
PCR products were loaded on a 4.8% polyacrylamide–8 M urea
gel and run in an AlphExpress Automated Sequencer (Amersham,
Milan, Italy). Data were automatically collected and analyzed by
the AlleleLinks software (Amersham).
Linkage maps were constructed using the MAPMAKER/EXP
3.0 program (Whitehead Institute for Medical Research, Washing-
ton, DC). Associations of phenotypic parameters (percentage, N
and V of remodeling and nonremodeling lesions) with alleles of rat
microsatellite markers were evaluated by LOD score, calculated
from the recombination frequencies observed between phenotype
and genotype. Phenotypic variables were analyzed as such or after
rank transformation. p values in the range of 10–4were considered
a statistically significant threshold for linkage.31Threshold LOD
score values at 2.8 and 4.3 were considered for “suggestive” and
“significant” linkage, respectively.32The proportion of total BFF2
variability explained by the association between the marker and
the trait (R2) was taken as an index of the importance of each
locus. QTL analysis was carried out with MAPMAKER/QTL 1.9.
Mean values of phenotypic parameters of BFF2 rats were deter-
mined for each genotype class and analyzed by ANOVA (SAS
Institute, Cary, NC). Potential interactions between genetic loci
were analyzed by ANOVA, and p values were corrected for
multiple comparisons using a formula proposed by Lander and
Schork.31Differences between parental strains and between ge-
netic subpopulations of BFF2 rats for phenotypic parameters were
analyzed by ANOVA, and multiple comparisons were made by TK
test using GraphPad (San Diego, CA) InStat 3 (www.graphpad.
Nodule remodeling in BFF2 rats
The absence of uniformity of GST-P immunostaining and the
irregularity of lesion margins represent reliable criteria to distin-
guish remodeling lesions of both F344 (Fig. 1b) and BFF1 (Fig.
1c,d) rats from nonremodeling nodules (Fig. 1a,b,d, asterisks).
Remodeling lesions are generally small in BFF1 rats (Fig. 1c),
though some large lesions are occasionally found (Fig. 1d). The
boundaries of nodules with the surrounding parenchyma may still
be recognizable even in nodules with pronounced remodeling (Fig.
Evaluation of the percentage of remodeling lesions (Fig. 2, R),
calculated on the basis of the number of lesions/cm3of liver,
showed about a 3.5-fold increase in the percentage of remodeling
FIGURE 1 – Neoplastic nodules 32 weeks after initiation in the liver
of F344 (a,b,e) and BFF1 (c,d) rats subjected to the RH model. (a)
Nonremodeling nodules showing almost uniform GST-P immunohis-
tochemistry and regular margins. (b) Two remodeling nodules with
nonuniform pattern of GST-P immunostaining, contrasting with the
uniformity of a nonremodeling nodule (asterisk). (c,d) Numerous
small remodeling nodules (c), a large remodeling nodule (d) and a
nonremodeling nodule (d, asterisk) in liver of BFF1 rats. (e) Nodule
with pronounced remodeling showing large peripheral GST-P-nega-
tive portions; note that the margins of the lesion (arrows) may be
distinguished from the compressed surrounding parenchyma. Magni-
fication ?13 (a–d), ?86 (e).
GENETIC CONTROL OF PHENOTYPIC REVERSION
lesions in BN and BFF1 rats with respect to F344 rats (p ?
0.0001). Nodular volume (Fig. 2, V) was much higher in F344 than
in BN and BFF1 rats (p ? 0.0001). It was 34% higher in nonre-
modeling than in remodeling lesions of F344 rats (p ? 0.001),
without any difference between the 2 types of lesion in BN and
BFF1 rats. As expected, BFF2 rats showed a distribution of per-
centage and volume of remodeling lesions in the range of parental
strain values (not shown). Apoptosis does not appear to be a major
determinant of remodeling in nodules of both susceptible and
resistant rats. Indeed, the percentages of apoptotic bodies, stained
with hematoxylin and eosin, with respect to whole neoplastic
hepatocytes were 1.25 ? 0.11% and 1.51 ? 0.16% (means ? SD,
n ? 5) in nonremodeling and remodeling nodules, respectively, of
F344 rats and 1.55 ? 0.09% and 2.08 ? 0.18% in the same lesions
of BN rats.
Linkage mapping of loci affecting remodeling
Evaluation of phenotypic parameters in 10 different arbitrary
phenotypic classes, covering the entire data range, showed a
normal distribution for N and percentage of lesions and a
non-normal distribution for V. Therefore, V values were rank-
transformed to obtain an improved normality.33QTL analysis
revealed the presence in BFF2 rats of 2 loci in suggestive
linkage with the percentage of remodeling lesions (Fig. 3). The
first locus, on chromosome 7, had a LOD score peak value of
3.85, at D7Rat25, positioned 53.2 cM from the first marker used
on the chromosome. The second locus, on chromosome 1, had
a LOD score peak value of 2.9, at D1Mgh14, positioned 172.1
cM from the first marker used. These QTL, named Hcrem1 and
Hcrem2 (hepatocarcinogenesis remodeling), respectively, were
also found, though with slightly lower LOD scores, when the
absolute N was considered, whereas they were absent when the
percentage (Fig. 3) and the number of nonremodeling lesions
were analyzed (LOD scores lower than 1). QTL analysis of
nodule V led to the identification, in both remodeling and
nonremodeling nodules, of 2 loci on chromosome 8, with LOD
score peaks of 3.8 and 2.9 for V, at D8Rat36 and D8Rat18,
respectively, positioned 42.1 and 78.8 cM from the first marker
used on the chromosome (not shown). These loci overlapped
with previously discovered Hcr3 and Hcr5, controlling V of the
whole nodule.11,12This indicates a common genetic control of
V in remodeling and nonremodeling nodules.
Genetic linkage between Hcrem1 and Hcrem2 and pheno-
typic traits was confirmed by ANOVA of average pheno-
typic values and allelic distribution patterns in homozygous
and heterozygous BFF2 rats, showing a significant dosage-
negative effect of homozygous B alleles on remodeling inci-
dence (Table I). Intermediate remodeling values in heterozy-
gous FB rats suggest incomplete penetrance of the B allele for
Hcrem1, whereas a recessive effect of the B allele was seen for
Table I includes the proportion of total intercross variability
explained by the association between the marker and the trait (R2).
This was 15.3% and 11.2%, for Hcrem1 and Hcrem2, respectively,
and could be further increased by additive interaction between the
2 loci. Indeed, significantly lower remodeling occurred in BFF2
rats, double homozygous for the B allele at D1Mgh14 and
D7Rat27 marker loci, corresponding to the LOD score peaks of
Hcrem1 and Hcrem2, respectively. Rats inheriting 3 B alleles
(FB/BB or BB/FB at D1Mgh14/D7Rat25) showed intermediate
values not significantly different from those of double BB/BB
homozygous rats at these markers. Significantly higher phenotypic
values (at least p ? 0.05) were found in rats bearing 1–2 B alleles
at only 1 of the 2 markers or in double FF homozygous rats, in
which no additive interactions were possible. A higher phenotypic
value with respect to double BB/BB homozygous rats also occurred
in double FB/FB heterozygous rats (p ? 0.001), which is consis-
FIGURE 2 – Percentage of remodeling nodules (R) and mean volume
(V, cm3? 104) of nonremodeling (hatched bars) and remodeling
(dotted bars) nodules in F344, BN and BFF1 rats 32 weeks after
initiation with diethylnitrosamine and selection according to the RH
model. Results are means ? SD, n ? 20. ANOVA showed significant
differences among groups for R and V in parental strains; TK test:
F344 vs. BN and BFF1 p ? 0.0001 for R and V; remodeling vs.
nonremodeling, p ? 0.001 for V of F344.
FIGURE 3 – LOD score for frequency of remodeling (thick line) and
nonremodeling (thin line) neoplastic liver nodules in BFF2 rats. BFF2
rats were genotyped by PCR of polymorphic genetic markers spanning
the length of chromosomes 7 and 1. QTL analysis was carried out with
MAPMAKER/QTL 1.9. On the abscissa, distance from the first mic-
rosatellite marker to the markers used on each chromosome. Dashed
vertical lines, Lander-Kruglyak thresholds for “suggestive” (LOD
score 2.8 in intercross) linkage.
DE MIGLIO ET AL.
tent with the presence of incompletely dominant (at D7Rat25) and
recessive (at D1Mgh14) B alleles.
Additional QTL without main phenotypic effects on remodeling
may be detected by evaluation of epistatic interactions inducing
phenotypic effects not predictable on the basis of the sum of their
separate effects. Thus, we performed 2?2 ANOVA with all of the
markers against all of the other markers. Among interactions,
calculated on the basis of remodeling, only those having p ? 0.05
(values corrected for genomewide comparisons32) were consid-
ered. Significant reciprocal interaction occurred between D1Mit3
and D11Rat11 (p ? 0.00002, corrected p ? 0.0013) and between
D6Rat14 and D8Rat46 (p ? 0.00041, corrected p ? 0.028),
though neither showed a significant individual effect, as shown by
insignificant LOD score values at these loci, ranging between
0.009 and 1.6. In both instances, significantly higher remodeling
occurred in rats homozygous for the B allele at both marker loci
(Fig. 4), with respect to all other allelic combinations (at least p ?
0.05), indicating the presence of recessive B alleles with a dosage-
positive effect on remodeling.
Our data indicate that the incidence of remodeling neoplastic rat
liver lesions is under the control of 2 major loci, Hcrem1 and
Hcrem2, mapping to chromosomes 7 and 1, respectively. These
loci do not overlap with the previously described Hcs1, on chro-
mosome 7, and Hcs2 and Hcs3, on chromosome 1.11,12Although
Hcrem1 and Hcrem2 do not reach the threshold of significance
(LOD score peak 4.3 in intercross rats32), their LOD scores were
above the suggestivity threshold (2.8 in intercross rats) and con-
sequently are worthy of report.32In addition, linkage of these loci
to remodeling was confirmed by ANOVA, which showed a sig-
nificant dosage-negative effect of the B allele on remodeling and
additive interaction between the 2 loci for their phenotypic effect.
Due to the inhibitory effect of B alleles, these QTL may be
considered susceptibility loci, inactive in the resistant parental
BFF1 rats, probably as a consequence of inhibition by modifier
genes. A similar situation was seen for rat Hcs1, Hcs3 and Hcs4,
located on chromosomes 7, 1 and 16, respectively, whose positive
effect on nodule growth depends on B alleles.11,12A possible
explanation of this behavior is based on the observation that,
according to the phylogenetic tree of various rat strains, including
F344 and BN,34numerous genetic events occurred during the
evolution of susceptible F344 rats from an ancestor common to BN
rats. This suggests selective mutation of resistance alleles, includ-
ing those responsible for remodeling, during the evolution of
susceptible rats from the common ancestor. Maintenance of unal-
tered resistance alleles in BN rats may result in inactivation (mod-
ifier effect) of susceptibility alleles, including those inhibiting
remodeling. Alleles contributed by BFF1 rats, associated with a
susceptible, nonremodeling phenotype in the intercross subpopu-
lation, are identified as susceptibility B alleles during linkage
analysis.3The proportion of total variability explained by the
association between the marker at LOD score peak and the char-
acter (R2) was about 15% and 11% for Hcrem1 and Hcrem2,
respectively, indicating consistent penetrance of the trait. Additive
interaction between these loci brought the total variability of the
trait up to about 26%. However, this effect could be overestimated
owing to the contribution of environmental variance not recog-
nized as additional loci in our BFF2 population.35This is in
keeping with the hypothesis that several not yet identified low-
penetrance genes regulate nodule remodeling, as also suggested by
the epistatic interactions discovered in the present study.
Remodeling of preneoplastic liver lesions occurs, in susceptible
rat strains, in early stages of the process,19,20,22reaching a maxi-
mum, in F344 rats, around 15 weeks after initiation.10Thereafter,
it progressively decreases and, at 32 weeks, 85–90% of lesions are
persistent (neoplastic) nodules. It is not clear if remodeling nodules
derive from distinct precursor cells or if they reflect the absence of
accumulating genetic events necessary for nodule progression.4In
the resistant rat strains, remodeling is higher in early stages of
hepatocarcinogenesis and progressively increases at least up to 32
weeks.10,22At this time, about 50% of lesions are still remodeling
in BN and BFF1 rats (see Pascale et al.10and the present report).
TABLE I – SUMMARY OF LINKAGE ANALYSIS IN BFF2 RATS
Percentage of remodeling nodules in BFF2 rats with genotype
19.50 ? 1.00 (28)
17.33 ? 1.02 (29)
16.34 ? 0.76 (55)
16.93 ? 0.88 (48)
12.34 ? 0.78 (24)4
12.06 ? 1.12 (34)4
1Distance from the first marker on the chromosome.–2F, F344 allele; B, BN allele. Means ? SE, with number of rats in parentheses.–
3Proportion of total intercross variance explained by the marker-genotype classes. R2and LOD scores are derived from Mapmaker/QTL 1.9
program.–4p calculated by ANOVA, and multiple comparisons made by TK test: BB vs. FF and BB vs. FB, at least p ? 0.05 for both QTL.
FIGURE 4 – Two-locus comparison of genotypes for 2 markers
showing epistatic interactions for percentage of remodeling nodules
(R) in BFF2 rats. Data are means ? SD of R for allelic combinations
at each locus on the abscissa. Statistical analysis: p values determined
by ANOVA were corrected according to Lander and Schork:31
D1Mit3/D11Rat11, corrected p ? 0.0013; D6Rat14/D8Rat46, cor-
rected p ? 0.028; multiple comparisons among allelic combinations,
performed by TK test, showed BB/BB significantly different (at least
p ? 0.05) from all other allelic combinations.
GENETIC CONTROL OF PHENOTYPIC REVERSION
The map position of the loci involved in remodeling is outside
previously identified Hcs and Hcr loci ,11,12suggesting that Hcrem
loci influence selectively nodule progression and do not directly
affect HCC development. However, the great increase in the
percentage of remodeling nodules in resistant rats reduces the
number of nodules prone to progress to HCC, in part explaining
the decrease in HCC incidence and multiplicity in these rats.10,11
The genes responsible for phenotypic reversion of neoplastic
liver lesions are presently unknown. The existence of inverse
relationships between remodeling and growth rate22suggests a
connection between some growth-regulatory mechanisms and re-
modeling. Various genes potentially involved in cell production,
mapping to the Hcrem1 (c-myc, Igf1, Ppp1cc, Avpr1a) and
Hcrem2 (Cyp17, Adra2a, Adrb1, Prkg1) loci (database http://
ratmap.gen.gu.se), are deregulated in rat and human liver carcino-
genesis. c-myc and c-myc target genes, including Cyclin E, E2F1
and Odc,36are upregulated in neoplastic liver lesions of rodents
and humans, and this is influenced by genetic predisposition to
hepatocarcinogenesis.24,27,37Genes involved in cell growth regu-
lation located at the Hcrem2 locus include Adra2a and Adrb1,
encoding adrenergic receptors ?2a and ?1, respectively, and Prkg1
and Cyp17, which are deregulated in rat and/or human HCC.38,39
Cyp17 encodes cytochrome P-450-17?, which mediates 17?-hy-
droxylase and 17,20-lyase activities in the androgen biosynthesis
pathway. It is a susceptibility gene for prostate cancer in men40and
interacts with steroid 5?-reductase type II and androgen receptor
genes, resulting in increased risk of HCC in humans.41Frequent
imbalance, suggestive of allelic gain, occurs in correspondence of
Hcrem1 and Hcrem2 loci in HCCs of susceptible rats42,43but not
in less advanced lesions of resistant BFF1 rats.44This suggests that
susceptibility loci Hcrem1 and Hcrem2 may be active in carcino-
gen-treated susceptible rats but not in resistant rats, probably as a
consequence of the modifier effect of resistance genes. The chro-
mosomal segment where Hcrem1 is located is syntenic to human
chromosomes 8q24, and the segment where Hcrem2 is located is
syntenic to human 10q11-21 and 10q24. Frequent allelic loss or
gain occurs in HCCs at 8q and, much less frequently, at chromo-
some 10q,45,46suggesting that some common mechanisms may
control growth and differentiation of rat and human neoplastic
nodules. All of the genes at Hcrem1 and Hcrem2 may be candi-
dates or targets of candidate genes. However, their relationships
with cell remodeling remain unknown. It would be interesting to
evaluate their expression in remodeling and nonremodeling lesions
and eventual interstrain polymorphisms. Generation of recombi-
nant congenic rats is under way in our laboratory, to obtain more
precise positioning of QTL and better characterization of the
phenotypic effects of the B and F alleles in each QTL.
Our previous and present data indicate that genetic predisposi-
tion to rat hepatocarcinogenesis depends on the interplay of several
genetic factors controlling growth and differentiation of neoplastic
liver lesions. This model, implicating the action of various low-
penetrance susceptibility genes controlled by modifier resistance
genes, could be extended to human hepatocarcinogenesis, taking
into account that assortative mating of genetically predisposed,
high-risk individuals leads to low penetrance of the trait in the
progeny. This is consistent with the observation of a slight risk
increase in the relatives of HCC patients and with the rarity of
familial clusters of HCCs even in high-risk areas.47
1. Dragani TA, Canzian F, Pierotti MA. A polygenic model of inherited
predisposition to cancer. FASEB J 1996;10:865–70.
Feo F, Pascale RM. Polygenic control of susceptibility and resistance
to hepatocarcinogenesis. In: Dragani TA, ed. Human polygenic dis-
eases. Animal models. Amsterdam: Harwood, 1998. 189–216.
Wood GA, Korkola JE, Archer MC. Tissue-specific resistance to
cancer development in the rat: phenotypes of tumor-modifier genes.
Feo F, Pascale RM, Simile MM, De Miglio MR, Muroni MR, Calvisi
DF. Genetic alterations in liver carcinogenesis: implications for new
preventive and therapeutic strategies. Crit Rev Oncog 2000;11:19–
Drinkwater NR, Ginsler JJ. Genetic control of hepatocarcinogenesis
in C57BL/6J and C3H/HeJ inbred mice. Carcinogenesis 1986;7:
Gariboldi M, Manenti G, Canzian F, Favella FS, Pienotti MA, Della
Porta G, Binelli G, Dragani TA. Chromosome mapping of murine
susceptibility loci to liver carcinogenesis. Cancer Res 1993;15:209–
Manenti G, Binelli G, Gariboldi M, Canzian F, De Gregorio L,
Favella FS, Dragani TA, Pienotti MA. Multiple loci affect genetic
predisposition to hepatocarcinogenesis in mice. Genomics 1994;23:
Lee GH, Drinkwater NR. The Hcr (hepatocarcinogen resistance) loci
of DBA/2J mice partially suppress phenotype expression of the Hcs
(hepatocarcinogen sensitivity) loci of C3H/HeJ mice. Carcinogenesis
Poole TM, Drinkwater NR. Two genes abrogate the inhibition of
murine hepatocarcinogenesis by ovarian hormones. Proc Natl Acad
Sci USA 1996;93:5848–53.
10. Pascale RM, Simile MM, De Miglio MR, Muroni MR, Gaspa L,
Dragani TA, Feo F. The BN rat strain carries dominant hepatocar-
cinogen resistance loci. Carcinogenesis 1996;17:1765–8.
11. De Miglio MR, Canzian F, Pascale RM, Simile MM, Muroni MR,
Calvisi D, Romeo G, Feo F. Identification of genetic loci controlling
hepatocarcinogenesis on rat chromosomes 7 and 10. Cancer Res
12. De Miglio MR, Pascale RM, Simile MM, Muroni MR, Calvisi DF,
Virdis P, Bosinco GM, Frau M, Seddaiu MA, Ladu S, Feo F. Chro-
mosome mapping of multiple loci affecting the genetic predisposition
to rat liver carcinogenesis. Cancer Res 2002;62:4459–63.
13. Zeng Z-Z, Higashi S, Kitayama W, Denda A, Yan Y, Matsuo K,
Konishi Y, Hiai H, Higashi K. Genetic resistance to chemical carcin-
ogen-induced preneoplastic hepatic lesions in DRH strain rats. Cancer
14. Melhem MF, Kunz HW, Gill TJ III. A major histocompatibility
complex–linked locus in the rat critically influences resistance to
diethylnitrosamine carcinogenesis. Proc Natl Acad Sci USA 1993;90:
15. Farber E, Sarma DSR. Hepatocarcinogenesis: a dynamic cellular
perspective. Lab Invest 1987;56:4–22.
16. Su Q, Benner A, Hofmann WJ, Otto G, Pichlmayr R, Bannasch P.
Human hepatic preneoplasia: phenotypes and proliferation kinetics of
foci and nodules of altered hepatocytes and their relationship to liver
cell dysplasia. Virchows Arch 1997;431:391–406.
17. Teebor GW, Becker FF. Regression and persistence of hyperplastic
hepatic nodules induced by N-2-fluorenylacetamide and their relation-
ship with hepatocarcinogenesis. Cancer Res 1971;31:1–3.
18. Kitagawa T. Sequential phenotypic changes in hyperplastic areas
during hepatocarcinogenesis in the rat. Cancer Res 1976;36:2534–9.
19. Enomoto K, Farber E. Kinetics of phenotypic maturation of remod-
eling of hyperplastic nodules during liver carcinogenesis. Cancer Res
20. Moore MA, Hacker J, Bannasch P. Phenotypic instability in focal and
nodular lesions induced by a short term system in rat liver. Carcino-
21. Tatematsu M, Nagamine Y, Farber E. Redifferentiation as a basis for
remodeling of carcinogen-induced hepatocyte nodules to normally
appearing liver. Cancer Res 1983;43:5049–58.
22. Garcea R, Daino L, Pascale RM, Simile MM, Puddu M, Frassetto S,
Cozzolino P, Seddaiu MA, Gaspa L, Feo F. Inhibition of promotion
and persistent nodule growth by S-adenosyl-L-methionine in rat liver
carcinogenesis: role of remodeling and apoptosis. Cancer Res 1989;
23. Schulte-Hermann R, Timmermann-Trosiener I, Barthel G, Bursch W.
DNA synthesis, apoptosis, and phenotypic expression as determinants
of growth of altered foci in rat liver during phenobarbital promotion.
Cancer Res 1990;50:5127–35.
24. De Miglio MR, Simile MM, Muroni MR, Pusceddu S, Calvisi D,
Carru A, Seddaiu MA, Daino L, Deiana L, Pascale RM, Feo F. c-myc
overexpression and amplification correlate with progression of pre-
neoplastic liver lesions to malignancy in the poorly susceptible Wistar
rat strain. Mol Carcinogen 1999;25:21–9.
25. Denda A, Kitayama W, Konishi Y, Yan Y, Fykamachi Y, Miura M,
Gotoh S, Ikemura K, Abe T, Higashi T, Higashi K. Genetic properties
for the suppression of development of putative preneoplastic gluta-
thione S-transferase placental form-positive foci in the liver of car-
cinogen-resistant DRH strain rats Cancer Lett 1999;140:59–67.
26. Wood GA, Korkola JE, Lee VM, Sarma DRS, Archer MC. Resistance
DE MIGLIO ET AL.
of Copenhagen rats to chemical induction of glutathione S-transferase
7-7-positive liver foci. Carcinogenesis 1997;18:1745–50.
27. Pascale RM, Simile MM, De Miglio MR, Muroni MR, Calvisi DF,
Asara G, Casabona D, Frau M, Seddaiu MA, Feo F. Cell cycle
deregulation in liver lesions of rats with and without genetic predis-
position to hepatocarcinogenesis. Hepatology 2002;35:1341–50.
28. Solt DB, Medline A, Farber E. Rapid emergence of carcinogen-
induced hyperplastic lesions in a new model of sequential analysis of
liver carcinogenesis Am J Pathol 1977;88:595–618.
29. Pugh TD, King JH, Koen H, Nychka D, Chover J, Wahba G, He Y-H,
Goldfarb S. Reliable stereological method for estimating the number
of microscopic hepatocellular foci from their transections. Cancer Res
30. Pascale RM, Marras V, Simile MM, Daino L, Pinna G, Bennati S,
Carta M, Seddaiu MA, Massarelli G, Feo F. Chemoprevention of rat
liver carcinogenesis by S-adenosyl-L-methionine: a long-term study.
Cancer Res 1992;52:4979–86.
31. Lander ES, Schork NJ. Genetic dissection of complex traits. Science
32. Lander ES, Kruglyak L. Genetic dissection of complex traits: guide-
lines for interpreting and reporting linkage results. Nat Genet 1995;
33. Conover WJ, Iman RL. Rank transformations as a bridge between
parametric and nonparametric statistics. Am Stat 1981;35:124–9.
34. Canzian F, Ushijima T, Pascale RM, Sugimura T, Dragani TA, Nagao
M. Construction of a phylogenetic tree for inbred strains of rat by
arbitrarily primed polymerase chain reaction (AP-PCR). Mamm Ge-
35. Barton NH, Keightley PD. Understanding quantitative genetic varia-
tion. Nat Rev Genet 2002;3:11–21.
36. Nasi S, Ciarapica R, Jucker R, Rosati J, Souceck L. Making decisions
through Myc. FEBS Lett 2001;490:153–62.
37. Pascale RM, Simile MM, Gaspa L, Daino L, Seddaiu MA, Pinna G,
Carta M, Zolo P, Feo F. Alterations of ornithine decarboxylase gene
during the progression of rat liver carcinogenesis. Carcinogenesis
38. Sanae F, Kohei K, Nomura M, Miyamoto K. Studies on responsive-
ness of hepatoma cells to catecholamines. VI. Characteristics of
adrenoceptors and adenylate cyclase response in rat ascites hepatoma
cells and human hepatoma cells. J Pharmacobiodyn 1992;15:303–9.
39. Kang-Park S, Lee JH, Shin JH, Lee YI. Activation of the IGF-II gene
by HBV-X protein requires PKC and p44/p42 MAP kinase signaling.
Biochem Biophys Res Commun 2001;283:303–7.
40. Chang B, Zheng SL, Isaacs SD, Wiley KE, Carpten JD, Hawkins GA,
Cleecker ER, Walsh PC, Trent JM, Meyers DA, Isaacs WB, Xu J.
Linkage and association of CYP17 gene in hereditary and sporadic
prostate cancer. Int J Cancer 2001;95:354–9.
41. Yu MW, Yang YC, Yang SY, Cheng SW, Liaw YF, Lin SM, Vhen
CJ. Hormonal markers and hepatitis B virus–related hepatocellular
carcinoma risk: a nested case-control study among men. J Natl Cancer
42. De Miglio MR, Muroni MR, Simile MM, Virdis P, Asara G, Frau M,
Calvisi DF, Seddaiu MA, Pascale RM, Feo F. Frequent loss of
heterozygosity at the Hcr1 (hepatocarcinogenesis resistance) locus on
chromosome 10 in primary hepatocellular carcinomas from LFF21 rat
strain. Hepatology 2001;33:1110–7.
43. Teerguarden JG, Newton MA, Dragan YP, Pitot HC. Genome-wide
loss of heterozygosity analysis of chemically induced rat hepatocel-
lular carcinomas reveals elevated frequency of allelic imbalances on
chromosomes 1, 6, 8, 11. 17, and 20. Mol Carcinog 2000;28:51–61.
44. Gariboldi M, Pascale RM, Manenti G, De Miglio MR, Calvisi D,
Carru A, Dragani TA, Feo F. Analysis of loss of heterozygosity in
neoplastic nodules induced by diethylnitrosamine in the resistant
BFF1 rat strain. Carcinogenesis 1999;20:1363–8.
45. Marchio A, Medeb M, Pineau P, Danglot G, Tiollais P, Bernheim A,
Dejean A. Recurrent chromosomal abnormalities in hepatocellular
carcinoma detected by comparative genomic hybridization. Genes
Chromosomes Cancer 1997;18:59–65.
46. Wong N, Lai P, Lee S-W, Fan S, Pang E, Liew C-T, Sheng Z, Lay J
W-Y, Johnson PJ. Assessment of genetic changes in hepatocellular
carcinoma by comparative genomic hybridization analysis. Relation-
ship to disease stage, tumor size, and cirrhosis. Am J Pathol 1999;
47. Fernandez E, La Vecchia C, D’Avanzo B, Negri E, Franceschi S.
Family history and the risk of liver, gallbladder, and pancreatic
cancer. Biomarkers Prev 1994;3:209–12.
GENETIC CONTROL OF PHENOTYPIC REVERSION