Medical Research Council Toxicology Unit, University of Leicester,
Leicester, United Kingdom
 |
Introduction |
The aryl hydrocarbon receptor
(AHR), a member of the basic
helix-loop-helix-periodicity/ARNT/simple-minded family of
transcription factors, is the primary target for the environmental
agent 2,3,7,8-tetrachlorodibenzo-p-dioxin (dioxin).
Mechanisms that lead to transcription of a few genes, such as
CYP1A1, have been well studied (Gu et al., 2000
), but the
pathogenesis of many of the large variety of toxic and carcinogenic disturbances in man and experimental systems remain unclear
(Pohjanvirta and Tuomisto, 1994; Gu et al., 2000). In mice,
polymorphism of the Ahr gene with a resulting 10-fold
difference in affinity for ligands is apparently associated with high
susceptibility of the C57BL/6J (b1 allele) and marked
resistance of the DBA/2 strain (d allele) (Poland and
Glover, 1980). The Ahr null mouse seems to be intractably
resistant to environmental chemicals of the dioxin type
(Fernandez-Salguero et al., 1996)). On the other hand, there is
considerable tissue and species variability in response to dioxin that
cannot be ascribed simply to polymorphisms of the Ahr gene
(Pohjanvirta and Tuomisto, 1994; Geyer et al., 1997). This suggests
that other modulating genes have profound effects on AHR-mediated
toxicity. In skin carcinogenesis, the ability of dioxin to act as a
promoter is dependent not only on the
Ahrb1 genotype but also on the
Hr locus (Knutson and Poland, 1982).
One of the actions of dioxin in rodents is to produce a malfunction of
hepatic heme synthesis (Fig. 1) (De
Matteis, 1998) similar to that seen in the human liver disorder
sporadic porphyria cutanea tarda (PCT), which has no known cause
(Elder, 1998; Anderson et al., 2001). In rodents exposed to dioxin and
in human PCT patients, hepatic uroporphyrinogen decarboxylase (UROD)
activity becomes markedly inhibited by an undetermined mechanism,
leading to massive accumulation of uroporphyrin in the liver, and there
is some association with hepatic toxicity (Pohjanvirta and Tuomisto,
1994; De Matteis, 1998; Elder, 1998; Smith et al., 1998). In mice,
CYP1A2 seems essential for the UROD defect, porphyria and aspects of
the liver injury (Smith et al., 2001). Over the years, it has become
clear that iron metabolism is also implicated. First, depletion of
liver iron partly protects against the porphyrinogenic and toxic action of dioxin in the livers of C57BL/6J mice (Sweeney et al., 1979; Jones
et al., 1981), whereas iron overload maximizes the porphyria and liver
injury (Smith et al., 1998). Second, the response to weak AHR ligands
or the marked resistance of some strains with the
Ahrd allele (e.g., SWR but not DBA/2) can
be overcome by elevation of iron stores (Greig et al., 1984; De
Matteis, 1998; Smith et al., 1998). In fact, iron itself will
eventually induce porphyria in some strains of mice (Smith and Francis,
1993; Philips et al., 2001). Thus uroporphyria in mice can be viewed as
a genetically determined disorder induced by iron that is enhanced by
chemicals, most potently by dioxin.
View larger version (22K):
[in this window]
[in a new window]
|
Fig. 1.
Disruption by dioxin of heme synthesis leading to
accumulation and excretion of uroporphyrin. A dual action of dioxin
causing both depressed uroporphyrinogen decarboxylase (UROD) activity
and induced Cyp1a2 expression. The former causes the
accumulation of uroporphyrinogens and some partial decarboxylation
products whereas the latter enhances oxidation of uroporphyrinogens I
and III to unmetabolizable uroporphyrins I and III. Severe hepatic
porphyria is associated with liver injury as is the similar human
disorder PCT (Elder, 1998).
|
|
In both familial and sporadic forms of PCT, there is also considerable
evidence for a crucial role of iron that is potentiated by exogenous
agents (Elder, 1998). For some groups of PCT patients, the C282Y and
H63D hemochromatosis gene (HFE) mutations are significant risk factors (Roberts et al., 1997; Bonkovsky et al., 1998; Sampietro et al., 1998; Bulaj et al., 2000).
Most of the actions of AHR as a regulator of gene expression are
thought to occur by heterodimerization with ARNT, a related transcription factor that can also form a partnership with HIF1
(Gu
et al., 2000). The three transcription factors are members of the
periodicity/ARNT/simple-minded superfamily. The evidence that dioxin in
the liver may act in an oxidative mechanism (Shertzer et al., 1998;
Smith et al., 1998), the role of the pro-oxidant iron, and the
relationship of AHR with ARNT and HIF1, which are implicated in
oxygen signaling pathways, may indicate that these are all
inter-related in the development of porphyria and liver injury. Here,
we have searched for quantitative trait loci (QTL) determining
susceptibility to dioxin toxicity by undertaking genetic analysis of
crosses of C57BL/6J and SWR mice with DBA/2 mice, all of which received
the same doses of iron and dioxin. The identification of genes that
markedly modify the actions of dioxin would be of considerable interest
in understanding the mechanism of how AHR ligands bring about
pathological changes. In addition, because we are dealing with an
important transcription factor family and iron metabolism, elucidation
of such gene interactions and novel polymorphisms might be important
for understanding some human disease processes.
| |
Materials and Methods |
Mice and Treatments.
C57BL/6J, DBA/2, C57BL/6×DBA/2
F1 and SWR mice were purchased at 6 weeks of age
from Harlan UK Ltd (Bicester, Oxfordshire, UK). C57BL/6.D2-Ahr
d mice were obtained from the Jackson
Laboratories (Bar Harbor, ME). F2 crosses of
C57BL/6J or SWR mice with DBA/2 mice were bred in the University of
Leicester. All mice had free access to water and rat and mouse diet no.
3 (Special Diet Services, Witham, UK) and were housed in
negative pressure isolators maintained at 21°C with a 12-h light/dark cycle.
Male mice at approximately 8 weeks of age received iron-dextran (800 mg
iron/kg of body weight) by subcutaneous injection; then, after 1 week,
they received 75 µg/kg dioxin in corn oil by oral administration (40 ml/kg) (Smith et al., 1998). Mice were monitored up to a further 5 weeks then blood was obtained by cardiac puncture under terminal
anesthesia. Liver and thymus were weighed. Where possible, tissue and
blood were obtained from mice that were culled earlier when appearing
morbid. All experiments were conducted under Home Office regulations.
Liver and plasma were stored at 
80°C for analysis.
Phenotypic Analyses.
Plasma alanine aminotransferase (ALT)
and aspartate aminotransferase (AST) activities as indicators of
hepatic damage, were measured using kits (Sigma-Aldrich, Poole, Dorset,
UK). Porphyrin liver contents, expressed as nanomoles of uroporphyrin
per gram, were estimated by spectrofluorometry (Grandchamp et al.,
1980). Uroporphyrin (URO) levels and plasma ALT and AST were converted to log values for analysis. Cytosolic UROD activity was measured as in
(Smith and Francis, 1987) using pentacarboxyporphyrinogen I as
substrate. Hepatic microsomal dealkylations catalyzed by CYP1A isoforms
and Western blotting (antibody from Gentest, Corp. MA) were performed
as described previously (Sinclair et al., 1990; 1998; Smith et al.,
2001). In the presence of both induced isoforms the microsomal
dealkylation of ethoxyresorufin and methoxyresorufin are mediated
predominantly by CYP1A1 and CYP1A2, respectively (Sinclair et al.,
1998; Smith et al., 2001). Blots with chemiluminescence detection were
extensively analyzed using ImageQuant (Molecular Dynamics, Sunnyvale, CA).
Genotyping by Microsatellite Analysis.
Genome-wide scans of
both the C57BL/6JxDBA/2 and the SWR×DBA/2 F2
crosses were performed using polymerase chain reaction (PCR) amplification of microsatellite DNA markers. Initially, high- and
low-responding mice were selected on the basis of uroporphyrin levels.
Marker locations and appropriate polymorphism data for the
C57BL/6J×DBA/2 cross were determined from the Whitehead Institute/MIT (Cambridge, MA; http://www-genome.wi.mit.edu/). Polymorphism data for
the SWR strain is not readily available in the databases; therefore, a
number of markers were tested for polymorphisms. SWR polymorphism data
were also kindly supplied by N. Drinkwater (McArdle Laboratory for
Cancer Research, Madison, WI) and R. A. McIndoe (CuraGen Corp.,
New Haven, CT) (McIndoe et al., 1999). Fluorescently labeled PCR
primers were obtained from Applied Biosystems (Warrington, Cheshire,
United Kingdom). Nonlabeled primers were synthesized by the Protein and
Nucleic Acid Laboratory in the University of Leicester. PCR
reactions contained 200 ng of genomic DNA extracted from tail or liver
samples, 10 pmol/µl each primer, 10 mM Tris, 50 mM KCl, 1.5 mM Mg
Cl2, 0.2 mM dNTPs, and 0.5 U of Taq
polymerase (PerkinElmer Life Sciences, Boston, MA). Samples were cycled
in a PTC-225 DNA Engine Tetrad (MJ Research, Watertown, MA) using a
touchdown program with an annealing temperature of 66 to 56°C with a
final 35 cycles at 55°C annealing temperature. PCR products were
analyzed on an ABI 377 sequencer using GeneScan and Genotyper (Applied
Biosystems). Data from these scans for the high and low responding
groups were analyzed using the 
2 test, which
showed any deviation from the 1:2:1 allelic ratio assumed for random
inheritance and by principle component analysis (PCA). From these we
chose chromosomes with potential QTLs for further analysis.
For the C57BL/6J×DBA/2 F2 cross, all individual
progeny were genotyped for the Trp53 (chromosome 11) and
Ahr (chromosome 12) polymorphisms (Schmidt et al., 1993;
Yang et al., 1999) and for further markers on chromosomes 11, 12, and
14. All mice of the SWR×DBA/2 F2 cross were
further typed with markers on chromosomes 1, 7, 9, and 11. Linkage
analysis was performed on the resultant data using MAPMAKER/EXP and
MAPMAKER/QTL (Lander and Green, 1987; http://www-genome.wi.mit.edu/genome_software/) to obtain peak LOD
scores for each of the selected chromosomes for a number of quantitative traits (i.e., hepatic uroporphyrin levels, plasma ALT and
AST levels, and thymus atrophy). Suggestive linkage was defined as a
LOD score of 2.8 or above and significant linkage as a LOD score of 4.3 or above (Lander and Kruglyak, 1995). Gene positions are those stated
by UniGene (http://www.ncbi.nlm.nih.gov/UniGene/Mm.Home.html). Ranges
of QTLs in centimorgans were estimated from 1 LOD score below the peak.
| |
Results |
Response of Mice to Iron-Dioxin.
The response of parent
strains to treatment with iron plus dioxin is shown in Table
1. Not only were there marked differences in elevation of hepatic porphyrin and plasma ALT levels but also in the
depression of UROD in the liver (Fig. 2).
Previous work has shown that prior administration of iron will maximize
the hepatic response of C57BL/6J (Ahrb1
allele) mice to dioxin observed 5 weeks later. Such mice show massive
uroporphyrin accumulation, moderate to severe inflammation, hepatocyte
loss, and biliary proliferation together with elevated plasma ALT
(Smith et al., 1998). In contrast, DBA/2
(Ahrd allele) mice treated similarly show
no porphyria, only mild lipid accumulation and hepatocyte hypertrophy,
and low elevation of plasma ALT. However, iron will overcome to a
significant degree the porphyric and injury response to dioxin of
another Ahd strain, SWR (Poland and
Glover, 1990; Smith et al., 1998).
View this table:
[in this window]
[in a new window]
|
TABLE 1
Effects of dioxin and iron on porphyria and plasma ALT in mice with
different Ahr genotypes
Mice received a single dose of iron-dextran (800 mg of Fe/kg) by
subcutaneous injection and then after 1 week they received dioxin (75 µg/kg) by oral administration as described under Materials and
Methods. The experiments were terminated after 5 weeks. Values are
means ± S.D. ALT data for SWR taken from Smith et al. (1998).
Values for URO and ALT in untreated mice or in those exposed to iron
alone are in Smith et al. (1998). In general, they resembled those
recorded here for DBA/2 mice.
|
|
View larger version (27K):
[in this window]
[in a new window]
|
Fig. 2.
Depression of UROD activity by dioxin is strain
dependent in mice. Activity was measured in iron-loaded mice 5 weeks
after a single dose of dioxin (75 µg/kg). Activity is expressed as a
percentage of untreated controls as estimated in Smith and Francis
(1987). Results are means ± S.D. (four mice per group).
|
|
To investigate the reasons for susceptibility differences between
strains, F2 intercrosses of C57BL/6J and SWR mice
with DBA/2 mice were given dioxin and iron for determination of QTLs.
Ranges of hepatic URO and plasma ALT levels in these crosses are shown in Table 1. Of the 204 mice of the C57BL/6×DBA/2 cross, 179 survived to 5 weeks. Among those that had to be culled prematurely, a
significant proportion (16 mice) showed signs of peritoneal, pulmonary,
or facial edema. Plasma ALT and AST were measured as markers of liver injury. Each showed the expected wide range of response (Table 1 and
Fig. 3) with a mean AST/ALT ratio of
activity of 3.78. Interestingly, mice recorded as having observable
edema had ALT values in the normal range but highly elevated AST levels
(ratio 24.6). A possible explanation is that this reflected cardiac
injury that may be the cause of dioxin-induced edema (Pohjanvirta and Tuomisto, 1994; Walker and Catron, 2000).
View larger version (24K):
[in this window]
[in a new window]
|
Fig. 3.
Distribution patterns of hepatic uroporphyrin levels,
thymus weights, and plasma ALT and AST responses in iron-loaded
C57BL/6×DBA/2 F2 mice 5 weeks after dioxin. URO and plasma
enzyme levels are plotted as log values because of the wide range of
upper extreme values.
|
|
No morbidity occurred with the SWR×DBA/2 F2
cross and there was a wide range of URO and ALT responses. As expected
(Smith et al., 1998), these were of an overall lower intensity than
seen with the C57BL/6J×DBA/2 cross (Table 1).
Ahr Genotype of C57BL/6J×DBA/2 Cross.
All mice of the
C57BL/6J×DBA/2 cross were genotyped for the
Ahrb1 and
Ahrd alleles. Those observed with edema
were either of the b1/b1 or b1/d
genotype. The extreme cohorts for each of the parameters (Fig. 3) were
analyzed using the 2 test (Table
2) for correlation with Ahr
alleles. Uroporphyrin and AST showed highly significant correlations
with the Ahrb1 alleles but ALT was
equivocal. Thymus atrophy, an established, nonhepatic effect of dioxin
(Poland and Glover, 1980; Pohjanvirta and Tuomisto, 1994) was
highly correlated with the possession of an
Ahrb1 allele. However, for all four
parameters there were approximately as many resistant mice with the
b1/d genotype as there were with the sensitive
phenotype. This finding suggests that although the Ahrb1 gene allele played a major role in
determining the susceptibility to dioxin for porphyric and other
responses, functional polymorphisms in other genes as well as subtle
environmental factors might be as important for these endpoints.
View this table:
[in this window]
[in a new window]
|
TABLE 2
Comparison of Ahr b1/b, b1/d and d/d genotypes
with extreme response cohorts of C57BL/6J (Ahrb1) × DBA/2 (Ahrd) F2 mice by 2
analysis
|
|
The hypothesis that genes other than the Ahr locus
contributing to susceptibility to iron-dioxin treatment was confirmed
by the development of a degree of porphyria in
C57BL/6.D2-Ahrd congenic mice (Fig.
4a). In contrast, the DBA/2 strain is
highly resistant to this regime (Table 1 and Greig et al., 1984; Smith et al., 1998). A similar response of
C57BL/6.D2-Ahrd mice has been observed
with hexachlorobenzene (Hahn et al., 1988). The DBA/2 region on
chromosome 12 in the congenic mice was mapped to approximately 3 centimorgans between D12 Mit153 (12 centimorgans) and D12 Mit222
(15.3cM). With these markers, Ahr lies close to D12 Mit 222 (Fig. 4b). Thus, we cannot currently exclude the possibility that some
other genes close to Ahr within the 3-centimorgan region are
also associated with toxicity.
View larger version (12K):
[in this window]
[in a new window]
|
Fig. 4.
Porphyric response of C57BL/6.
D2-Ahrd congenic mice after dosing
with iron and dioxin as described in Table 1 (a). This demonstrates
that iron pretreatment partially overcame the resistance of mice with
the Ahrd allele to dioxin when present
in a C57BL/6J background. Data are means of five mice per group ± S.D. The congenic region around the Ahr gene on
chromosome 12 was estimated in case other genes of potential influence
are eventually found nearby (b). The Ahr gene lies close
to D12 Mit222 in a region approximately 3 to 5 centimorgans long .
|
|
Mapping of QTLs in C57BL/6J×DBA/2 Cross.
To determine the
location of susceptibility genes, the extreme phenotype cohorts (48 each) were initially genotyped for 44 markers spread across the genome
at an average of about 20 centimorgans apart. As well as chromosome 12 (presumably the Ahr gene), PCA and
2 analyses showed significant
(p < 0.05) associations between URO phenotypes and
markers on other chromosomes, in particular 11 and 14. Consequently,
all mice were genotyped with six to nine markers for each of these
chromosomes. LOD scores by MAPMAKER for URO, plasma enzymes, and thymus
weights are shown in Table 3. Mapping of
significant QTLs for URO on chromosomes 11 and 12 with peak LOD scores
above 4.3 (Lander and Kruglyak, 1995) are shown in Fig.
5. The QTL for porphyria on chromosome 11 at approximately 45 to 60 centimorgans had a peak LOD score similar to
that for the Ahr gene on chromosome 12. It was distal to
Trp53 (39cM) that in preliminary studies showed linkage by
PCA of a combination of all phenotype parameters including body and
liver weight (Yang et al., 1999). A possible QTL for URO on chromosome 14 was also detected at 20 to 30 centimorgans (LOD score, 2.94) (Table
3). No interaction between any of these loci was found using a two-way
general linear model ANOVA. For thymus atrophy and plasma AST, only one
QTL, on chromosome 12 in the region encompassed by D12 Mit242,
Ahr, and D12 Mit88, was detected (Table 3). Although no QTLs
above LOD score 2.8 could be detected on chromosomes 11 and 14 for ALT
and AST, there was significant association by PCA with extreme
phenotypes. Similarly, PCA gave significant linkage between extreme ALT
values and chromosome 12 (D12 Mit242) genotype.
View this table:
[in this window]
[in a new window]
|
TABLE 3
Peak LOD scores for fully mapped chromosomes
A LOD score of 2.8 was taken to be statistically suggestive and 4.3 as
significant (Lander and Kruglyak, 1995). Regression analysis was used
to quantify the percent of variance of URO by the most statistically
significant (as estimated by a one-way ANOVA) genetic markers. The
ratio in the C57BL/6 × DBA/2 F2 cross was estimated as
1:1.75:0.88 for D11Mit179, Ahr, and D14Mit60 respectively. In the
SWR × DBA/2 F2 cross, 1:0.08:1.21 for D1Mit14, D9Mit15,
and D11Mit70 respectively.
|
|
View larger version (17K):
[in this window]
[in a new window]
|
Fig. 5.
QTL plots for porphyria on chromosomes 11 and 12 of
C57BL/6×DBA/2 F2 mice after dioxin and iron. The QTL on
chromosome closely correlated with that for the Ahr
gene.
|
|
Mapping of QTLs in SWR×DBA/2 cross. To investigate further possible
loci other than Ahr, a genome scan (42 markers) was
conducted on the extreme cohorts (35-47 per phenotype) for URO and ALT
of the SWR×DBA/2 cross (Fig. 6).
Significant association by PCA and one-way ANOVA or
2 analysis was observed between response and
potential loci on chromosomes 1, 9, and 11, but not with chromosome 12 (even with D12 Mit88 at 19.7 centimorgans) because the parent strains
are syngenic for the Ahr gene. All mice were subsequently
genotyped for 9 to 12 markers on chromosomes 1,9 and 11. This confirmed the C57BL/6JxDBA/2 result indicating a highly significant QTL for
uroporphyria on chromosome 11 at approximately 40 to 60 centimorgans (Fig. 7; Table 3). In addition,
significant loci for both porphyria (Fig. 7) and plasma ALT were found
on chromosome 1 (Table 3) at approximately 27 to 57 centimorgans.
Two-way ANOVA for URO levels showed a significant interaction between
the loci on chromosomes 1 and 11, although these seemed to control
uroporphyria more than ALT. That is to say, the largest differences in
URO levels among genotypes at the D1 Mit21 locus were seen when there
was an SWR allele at the D11 Mit288 locus. There seemed to be a second
locus on chromosome 1 for uroporphyrin (closest marker, D1 Mit110) with a peak LOD score 4.4 at 86 to 100 centimorgans (Fig. 4), but unlike the
most significant locus on this chromosome, it showed no significant correlation with ALT (LOD 2.1). A possible QTL was also detected on
chromosome 9 (closest marker D9 Mit15 at 60 centimorgans) but distal to
Cyp1a2. Although weak, this locus seemed to be linked particularly with ALT and less so with uroporphyrin but this may just
reflect the limitation of the analyses. In this cross, a QTL was not
detected for thymus weight (Table 3).
View larger version (22K):
[in this window]
[in a new window]
|
Fig. 6.
Distribution of hepatic uroporphyrin levels and
plasma ALT responses in iron loaded SWR×DBA/2 F2 mice 5 weeks after dosing with dioxin. Data are plotted as log values.
|
|
View larger version (21K):
[in this window]
[in a new window]
|
Fig. 7.
QTL plots for porphyria on chromosome 1 and 11 of the
SWR×DBA/2 F2 cross after dioxin and iron using all 242 mice. At least one QTL was present on chromosome 1.
|
|
Expression of CYP1A2.
From studies with Cyp1a2
(/) mice, it is clear that CYP1A2 is essential for the development
of uroporphyria in mice and some aspects of liver injury caused by
dioxin (Sinclair et al., 1998; Smith et al., 2001). Despite
this, there was no evidence that polymorphism of the Cyp1a2
gene (chromosome 9 at 31 centimorgans) accounted for a major proportion
of the susceptibility of C57BL/6J mice compared with the resistant
DBA/2 strain. However, expression of Cyp1a2 under these
conditions is controlled by the AHR so that the polymorphism of the
Ahr gene might contribute to the differences in
susceptibility. The dose of dioxin was greater than that known to
achieve maximum induction of CYP1A isoforms in
Ahrb1 and
Ahrd mice (Pohjanvirta and Tuomisto,
1994). Marked induction of CYP1A2 protein and activity was observed in
both strains and their intercross and did not seem to correlate with
sensitivity (Fig. 8). The SWR and DBA/2
strains are not polymorphic for the Ahr gene (Poland and
Glover, 1990) and we did not find any significant linkage with D9 Mit31
(33 centimorgans) close to the Cyp1a1/2 genes. In addition,
there was no marked difference in CYP1A2 expression in the extreme
porphyria phenotypes of the SWRxDBA/2 cross (Fig. 8).
View larger version (26K):
[in this window]
[in a new window]
|
Fig. 8.
Hepatic microsomal CYP1A2 expressions and activities
in extreme porphyria responses of F2 crosses after dioxin
and iron. a, Western blotting of CYP1A1 and CYP1A2 in C57BL/6J, SWR,
and DBA/2 iron-loaded mice 5 weeks after dioxin and representative
C57BL/6×DBA/2 F2 high- (990-3332 nmol/g) and low-
(0.8-1.7 nmol/g) responding mice and SWR×DBA/2 F2 high-
(297-981 nmol/g) and low- (0.2-0.4 nmol/g) responding mice for
uroporphyrin levels. No differences were detected between high- and
low-responding groups of the SWR×DBA/2 F2 cross after
extensive quantitation of three blots with different samples
(ImageQuant). Levels were reduced in high responding C57BL/6×DBA/2
F2 mice compared with low responders perhaps reflecting
liver damage. b, dealkylations by highest and lowest responding mice
(see Table 2) of C57BL/6×DBA/2 F2 cross (50 per group) and
were statistically different whether b1b1 or
b1d genotypes. Ethoxyresorufin dalkylation (EROD) is
particularly carried out by CYP1A1 and methoxyresorufin dealkylation
(MROD) by CYP1A2. c, dealkylations by highest (282-981 nmol/g) and
lowest (0.2-0.4 nmol/g) responding mice of SWR×DBA/2 F2
mice (16 per group).
|
|
| |
Discussion |
This study has demonstrated QTLs, in addition to Ahr,
that have a marked effect on porphyria and some aspects of liver injury caused by dioxin when interacting with iron. There are at least three
QTLs on chromosomes 1, 11, and 14, and probably one on chromosome 9, besides Ahr, that are responsible for the differences in
susceptibility between DBA/2 mice and the C57BL/6J and SWR strains.
Thus, as far as liver injury is concerned, the view that resistance to dioxin in mice is completely dominated by the Ahr
d allele is misleading, as has been argued
previously (Greig et al., 1984; Pohjanvirta and Tuomisto, 1994; Geyer
et al., 1997). It seems likely that at least one QTL will be linked
with liver iron mobilization and metabolism and that interactions with
expressions mediated by the AHR lead to depression of UROD activity,
porphyria, and some aspects of hepatic injury, which can result in
elevated plasma ALT and AST levels. The SWR×DBA/2 QTL for both ALT and porphyria on chromosome 1 is in the same region as some genes associated with iron metabolism (e.g., ferroportin). In addition, although weak, the QTL on chromosome 9 at approximately 60 centimorgans is also in approximately the same region as some genes for iron metabolism (e.g., lactotransferrin, transferrin, and ceruloplasmin). These iron metabolism genes may not be involved in the porphyria response, but other potential target genes may be associated with them.
Iron has the most marked effect on porphyria development in SWR mice
(Smith et al., 1998). So far the QTL on chromosome 11 seems to have no
strong candidate, yet accounts for much of the susceptibility of
C57BL/6J and SWR strains to iron/dioxin synergism and in the latter may
interact with a gene on chromosome 1 in SWR mice. The QTL found on
chromosome 14 in the C57BL/6×DBA/2 F2 cross is
in a similar region to Hcbip1 found previously in a
C57BL/10ScSn×DBA/2 F2 cross of susceptibility to porphyria induced by
hexachlorobenzene-iron synergism (Akhtar and Smith, 1998) but is
centromeric to the Hr gene at (39cM). No evidence for this locus was found in the SWR×DBA/2 cross. No linkage was detected in
either study between phenotype and markers that are close to Urod on chromosome 4 (D4 Mit352 and D4 Mit57) or known
positions of other heme synthesis genes. However, an important
possibility is that one QTL corresponds to alleles of Alas1,
the controlling gene of heme synthesis (Roberts and Elder, 2001), but
polymorphic alleles and the chromosomal location of this gene in the
mouse have not been reported.
One of the interesting findings from the C57BL/6×DBA2
F2 study is the association of a QTL with the
Ahr gene, yet this did not seem to correlate with CYP1A2
expression that is essential for the porphyric response as shown by
studies with Cyp1a2 (/) mice (Sinclair et al., 1998;
Smith et al., 2001). Although recent studies of caffeine
metabolism in mice have shown a QTL on chromosome 9, close to the
Cyp1a2 gene that might be related to variable CYP1A2 levels
(Casley et al., 1999), this did not seem to be linked in our
experiments. It is possible that genes in addition to Cyp1a2 controlled by the AHR are important for porphyria and some aspects of
liver injury in C57BL/6J mice and are still differentially expressed
when Cyp1 genes are maximally induced. This also illustrates that null mice may demonstrate the necessity of a particular gene to a
pathological process but do not show whether it is a candidate susceptibility gene. Of course as far as levels of CYP1A2 are concerned, we compared activity and protein at the termination of the
experiments. It is possible that levels are critical at earlier times
in the porphyrinogenic process in this model although we have no real
supportive evidence for this from the QTL or previous studies (Smith et
al. 1998).
The development of hepatic porphyria by dioxin has many similarities
with human sporadic PCT (Elder, 1998; Anderson et al., 2001). Both seem
to involve interactions with iron metabolism (Elder, 1998).
Identification of the genes responsible for the QTLs in the present
studies may be important to our understanding the mechanism of human
sporadic PCT as well as being pertinent to other disorders. Of course,
identifying the genes responsible for the QTLs is not an easy
undertaking given the large regions of chromosomes identified, and this
may entail more specific crosses for finer mapping resolution. However,
the number of potential candidate genes is growing rapidly and with the
draft of the mouse genome, progress should be far quicker than might
have been envisaged only a few years ago. The only susceptibility genes
found for some groups of sporadic PCT patients have been identified as
mutants of HFE (Roberts et al., 1997; Bonkovsky et
al., 1998; Sampietro et al., 1998; Bulaj et al., 2000).
Interestingly, penetrance for hemochromatosis by HFE is
thought to depend on alleles of other genes (Whitfield et al., 2000;
Fleming et al., 2001). It is now recognized that many diseases are
complex interactions of susceptibility genes and environmental factors.
Chemically induced porphyria and sporadic PCT are unlikely to be any
different. The influence of modifier genes in other actions of dioxin
also requires investigation.
In summary, although gene expression controlled by variants of the AHR
is usually considered to be the major gene manifesting toxicity of
dioxin, the development of porphyria and elevation of plasma enzymes as
indicators of liver injury have shown the presence and location of
other susceptibility genes. It seems possible that some of the genes
constituting the QTLs are also of significance in the development of
PCT and perhaps other human disorders.
We gratefully acknowledge the advice of Drs. R. A. McIndoe
and N. Drinkwater for polymorphism data; Dr. K. Lilley, C. Travis, and
colleagues for practical advice and assistance; and Professor G. H. Elder and Dr. A. Roberts for reading the manuscript.
Dr. A. G. Smith, MRC
Toxicology Unit, Hodgkin Building, PO Box 138, Leicester University,
Lancaster Rd., Leicester LE1 9HN, UK. E-mail:
ags5{at}le.ac.uk
![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
Top
Abstract
Introduction
