MC 863
288
KEGELMEYER ET AL.
MOLECULAR CARCINOGENESIS 20:288–297 (1997)
Differential Display Identified Changes in mRNA Levels in
Regenerating Livers From Chloroform-Treated Mice
Alma E. Kegelmeyer, Catherine S. Sprankle, Gregory J. Horesovsky, and Byron E. Butterworth*
Chemical Industry Institute of Toxicology, Research Triangle Park, North Carolina
The nongenotoxic-cytotoxic carcinogen chloroform induces liver necrosis, regenerative cell proliferation,
and, eventually, liver tumors in female B6C3F1 mice when administered by gavage at doses of 238 or 477 mg/
kg/d. Administration of 1800 ppm of chloroform in the drinking water results in similar daily doses but does
not produce liver toxicity or cancer. The differential-display technique was used to compare the expression of
a subset of mRNAs in normal (control) and regenerating liver after chloroform-induced toxicity to define the
proportion of genes whose expression changes under hepatotoxic conditions and to identify the genes that
might play a role in regeneration and perhaps cancer. RNA was purified from the livers of female B6C3F1 mice
after 4 d or 3 wk of gavage treatment with 3, 238, or 477 mg/kg/d of chloroform or treatment with 1800 ppm
chloroform in drinking water. There was a remarkably high degree of consistency of gene expression among
the animals and across dose and treatment groups as visualized by the differential-display technique. Of the
387 bands observed, only four (about 1%) changed expression in regenerating liver. The genes were assigned
locus names by GenBank after sequence submission. The genes with increased mRNA levels as confirmed by
northern blot analysis were MUSTIS21, a mouse primary response gene induced by growth factors and tumor
promoters; MUSMRNAH, a gene highly homologous to a human gene isolated from a prostate carcinoma cell
line; and MUSFRA, a novel gene. The novel gene MUSFRB exhibited decreased mRNA levels. No change in
expression was seen among control mice given the nontoxic regimens of 3 mg/kg/d chloroform or 1800 ppm
chloroform in drinking water, indicating that changes in expression were associated with toxicity and regeneration rather than chloroform per se. These genes and others that may be identified by expanding this approach may play a role in regeneration and perhaps in the process of chloroform-induced carcinogenesis in
rodent liver. Mol. Carcinog. 20:288–297, 1997. © 1997 Wiley-Liss, Inc.
Key words: gene expression; differential display; polymerase chain reaction; cytotoxic chemicals; carcinogenesis; B6C3F1 mouse
INTRODUCTION
Chloroform (CHCl3) is a trace contaminant generated as a by-product of drinking water and swimming
pool chlorination [1,2] and of some paper-bleaching
processes [3,4]. Chloroform induces cancer in rodents
by a nongenotoxic-cytotoxic mode of action [5].
Neither chloroform nor its metabolites react directly
with DNA [6]. Rather, tumor formation is secondary
to events associated with cytolethality and regenerative cell proliferation [7]. For example, when given
by bolus gavage in corn oil at doses of 238 or 477
mg/kg/d to female B6C3F1 mice, chloroform induces
liver necrosis, regenerative cell proliferation, and,
eventually, hepatocellular carcinoma [8,9]. However,
when similar daily doses of chloroform are given in
the drinking water, no increase in liver toxicity or
liver tumors is observed [9,10]. Thus, although chloroform is a rodent carcinogen, it is probably not a
human carcinogen because of its high-dose–only,
nongenotoxic-cytotoxic mode of action and because
it is found only in trace amounts in the environment [5].
Numerous events associated with toxicity and regeneration may be involved in carcinogenesis. Nu© 1997 WILEY-LISS, INC.
cleases can be released that may damage DNA.
Increased cell turnover can increase the rate of conversion of DNA adducts from endogenous and exogenous sources to mutations before DNA repair can
occur. In addition, many genes responsible for
growth control are also oncogenes and may exhibit
increased expression in regenerating tissue [11–14].
Alteration in the sequence or expression of such
genes can be involved in carcinogenesis [15,16]. Furthermore, during transcription the DNA is unwound,
thus making these genes more susceptible to spontaneous mutations [17]. In stimulating regeneration,
increased expression of growth-control genes may
The current address for Alma E. Kegelmeyer is Kaiser Permanente
Medical Group, Genetics Dept., Bldg. 6, 5755 Cottle Rd., San Jose,
CA 95119.
The current address for Gregory J. Horesovsky is Glaxo Wellcome,
5 Moore Dr., Research Triangle Park, NC 27709.
*Correspondence to: CIIT, P.O. Box 12137, Research Triangle Park,
NC 27709.
Received 20 December 1996; Revised 6 May 1997; Accepted 9
May 1997
Abbreviations: PCR, polymerase chain reaction; BrdU, 5-bromo2´-deoxyuridine; SSC, standard saline citrate.
CHLOROFORM-INDUCED CHANGES IN mRNA LEVELS
also provide a selective growth advantage to spontaneously initiated precancerous cells [18,19].
It has been estimated that some mammalian cells
express about 15 000 genes [20]. Thus, it is a daunting task to identify those genes that change expression because of toxicity and regenerative cell
proliferation and that may play a role in carcinogenesis. Differential display is a rapid, sensitive, and reproducible technique that can identify genes that
are expressed at different levels in one population of
cells compared with another [20]. Pairs of primers
are chosen to amplify partial cDNA sequences from
subsets of the total population of mRNAs by reverse
transcription and polymerase chain reaction (PCR).
The 3´ primers include a poly(T) sequence to select
and anchor the primer at the 3´ end of the poly(A)
tail present on most eukaryotic mRNAs. Two additional bases on the 3´ end of the poly(T) sequence
restrict the primer to a subset of mRNAs. An internal
5´ primer is chosen to further restrict amplification
to only cDNA sequences from a small subpopulation of mRNAs.
Our laboratory has described extensive necrosis
and regenerative proliferation in the livers of female
B6C3F1 mice given chloroform by gavage at the carcinogenic doses of 238 and 477 mg/kg/d (but not at
3 mg/kg/d) for periods of 4 d or 3 wk [9]. We also
described the lack of liver toxicity when B6C3F1 mice
were given 1800 ppm of chloroform in drinking water even though that regimen yields a dose of 329
mg/kg/d chloroform when averaged over a 3-wk treatment period [9]. The differential-display technique
was used to analyze frozen liver samples from these
studies to provide an initial estimate of the proportion of genes that change expression under hepatotoxic-carcinogenic conditions. Any genes so
identified may be associated with molecular events
underlying the initial stages of chloroform-induced
hepatocarcinogenesis in female mice.
MATERIALS AND METHODS
Chemicals and DNA Probes
Chloroform, >99.5% pure and stabilized with
0.006% amylenes, was obtained from Fluka Chemical Corp. (Ronkonkoma, NY). The rat albumin probe
(a 1.0-kb PstI fragment) was a gift from Dr. L. Kier
(Monsanto Corp., St. Louis, MO). The mouse TIS21
probe (a 600-bp NotI-SalI fragment) was obtained
from Dr. J. Staudinger (Glaxo-Wellcome, Research
Triangle Park, NC). All reagents for polyacrylamide
gel electrophoresis were from Amresco (Solon, OH).
Radiochemicals were obtained from Amersham Corp.
(Arlington Heights, IL). Hybrisol hybridization solution and 20× standard saline citrate (SSC) stock were
obtained from Oncor (Gaithersburg, MD). All other
chemicals were obtained from Sigma Chemical Co.
(St. Louis, MO), except as otherwise noted. All chemicals were of the highest purity available, and prod-
289
ucts characterized as “ultra pure” or “molecular biology grade” were used when possible.
Animals
This study was conducted under federal guidelines
for the humane care and use of laboratory animals
[21] and was approved by the Chemical Industry
Institute of Toxicology Institutional Animal Care and
Use Committee. The animals were housed in humidity- and temperature-controlled HEPA-filtered, mass
air–displacement rooms in facilities accredited by the
American Association for the Accreditation of Laboratory Animal Care. The animal care procedures used
have been detailed previously [9]. Briefly, female
B6C3F1 mice (B6C3F1/CrlBR) were obtained from
Charles River Breeding Laboratories, Inc. (Raleigh,
NC). The rodents were provided with NIH-07 rodent
chow (Ziegler Bros., Gardener, PA) and deionized,
filtered tap water ad libitum. Lighting was provided
on a 12-h light-dark cycle. Rodents were acclimated
to the animal facility for 2 wk before use and were 9
wk old at the beginning of these studies. This age
was chosen to duplicate the original conditions of
the gavage bioassay [8]. Background cell proliferation rates were low and consistent from animal to
animal at this age (data not shown). The animals
were randomized by weight into treatment groups.
Sentinel animals were housed in the animal facility
as part of an ongoing program to monitor them for
parasitic, bacterial, and viral infections and were
pathogen free throughout the study.
Dosing and Necropsy of Animals
All liver tissue used was from a previously described
study [9] and had been stored frozen at –70°C. No
deterioration of the mRNA in the liver tissues was
noted under these conditions (data not shown). The
mice were given chloroform either daily by gavage
in corn oil or in drinking water ad libitum. Chloroform in corn oil was given at doses of 0, 3, 238, or
477 mg/kg/d for either 4 consecutive days or for 5 d/
wk for 3 wk. Chloroform in drinking water was provided ad libitum at concentrations of 0 or 1800 ppm
for either 4 consecutive days or every day for 3 wk.
At necropsy, the livers to be used for RNA purification were frozen in liquid nitrogen and stored frozen as described above.
The cell replication data from this experiment was
published previously [9]. Groups of concurrently
treated animals separate from those used to analyze
gene expression were treated with 5-bromo-2´deoxyuridine (BrdU) by osmotic pump for 3.5 d before necropsy to label cells in S phase. Regenerative
cell proliferation was then quantitated immunohistochemically from liver sections.
Purification of RNA and Preparation of
Northern Blots
Total cellular RNA was purified from approximately
0.5 g of liver tissue; each RNA preparation included
290
KEGELMEYER ET AL.
tissue from several different lobes of the liver. The
RNAzol reagent was used for RNA purification according to the instructions of the manufacturer (TelTest, Friendswood, TX). RNA from at least two
different mice was isolated for each individual data
point. Furthermore, the degree of expression was
compared across dose and treatment groups to confirm a consistent pattern of change. The consistent
lack of response in the 3 mg/kg/d (low dose) and
1800 ppm drinking water groups served as additional
internal controls. One hundred micrograms of total
RNA was stored frozen at –70°C at a concentration of
1 µg/mL for use in the differential-display technique.
Poly(A)+ mRNA was isolated and used to prepare
northern blots as described previously [11]. After electrophoresis to separate the mRNA according to size,
the RNA was transferred to a nylon membrane and
then cross-linked to the membrane with a Stratalinker
UV cross-linker (Stratagene, La Jolla, CA). The membranes were stored in sealed bags at 4°C until used.
Differential Display
Total cellular RNA isolated as described above was
used to synthesize portions of cDNA, and those
cDNAs were subsequently amplified by PCR by using the conditions of Liang and Pardee [20] with
minor modifications. The sequences of the terminal
primers used in the 12 different display reactions were
T11CN, T11GN, or T11AN (where N is A, T, G, or C)
and are collectively referred to as the poly(T) primers. In all cases, the internal primer used was LTK3
(5´-CTTGATTGCC-3´). Reaction mixes of 10× buffer
(100 mM Tris, pH 8.8; 500 mM KCl; and 15 mM
MgCl2; Perkin Elmer, Branchburg, NJ), dNTPs (20 µM
final from Pharmacia Inc., Piscataway, NJ), RNasin
(Promega Corp., Madison, WI), reverse transcriptase
(Gibco/BRL, Gaithersburg, MD), and one of the
poly(T) primers (50 pmol/mL) were prepared. Two
micrograms of total RNA from the liver of a control
or treated mouse was added, and the reaction was
incubated for 10 min at 22°C, 15 min at 37°C, and 5
min at 99°C to make cDNA and inactivate the reverse transcriptase.
The cDNAs were amplified by combining 10×
buffer, dNTPs (2.5 mM final), AmpliTaq (5 U/µL;
Applied Biosystems, Foster City, CA), 2 µL of cDNA,
1 µL of [α-35S]dATP (specific activity 10 mCi/mL), and
50 pmol of both LTK3 and the same poly(T) primer
used previously in a final volume of 20 µL. The PCR
cycling conditions were described previously [20].
Five microliters of amplified product was run on 6%
denaturing sequencing gels for 3 h at 50 W, and each
gel was soaked in water for 15–20 min to remove
urea before drying. Each gel had a molecular-weight
marker lane and lanes for two different animals for
each of the eight gavage and four drinking-water
chloroform treatment conditions, for a total of 25
lanes per gel. The resolution of the final autoradiograms was such that differences in intensity between
bands of equal molecular weight were readily observed visually. Only the cDNA bands that consistently showed changes between control and treated
samples were analyzed further. Reactions that
showed any changes in mRNA levels between control and treated samples were repeated a minimum
of two times to ensure reproducibility.
Cloning of cDNAs
RNA samples exhibiting reproducible differences
between treated and control samples were analyzed
on northern blots prepared as previously described
[11,22]. cDNA bands of interest were cut from the
sequencing gels and eluted in 300 mM sodium acetate, pH 5.2, for 1 h at 37°C and then reamplified
by PCR with the same primer pairs and conditions
used to initially amplify the fragment. The amplified fragments were cloned directly into the pCRII
vector by using the TA Cloning kit (Invitrogen, San
Diego, CA) according to the directions of the manufacturer. Clones were checked for correct insert size
by colony PCR amplification from individual bacterial colonies in 10× buffer (Perkin Elmer), MgCl2 (2
mM final; Perkin Elmer), dNTPs (320 µM final), 5 U
of AmpliTaq, and 50 pmol each of the vector primers SP6 (5´-CCAAGCTATTTAGGTGACACTATAG-3´)
and T7 (5´-AATTGTAATACGACTCACTATAGGG-3´).
The PCR-amplified DNAs were purified for sequencing and probe preparation according to the
manufacturer’s recommendations by using the
Sephaglas BandPrep kit (Pharmacia Inc.). Each clone
was isolated from at least two different colonies of
bacteria and analyzed separately to verify the sequence of the clone. The clones were then used as
probes to hybridize northern blots containing RNA
from the livers of chloroform-treated animals as described above [11,22].
Northern Blot Hybridization and Analysis
Filters containing separated mRNAs were prehybridized with 6 mL of Hybrisol for 4 h at 42°C in a
Robbins hybridization oven. The cDNA clones were
labeled with [32P]dCTP by using the Prime-It II random priming kit (Stratagene) and 50 ng of template.
The probes were labeled to a minimum specific activity of 108 cpm/µg. A total of 1 × 106 incorporated
counts was then added to each prehybridization solution. After hybridization for 16–38 h at 42°C, the
filters were washed as previously described [11,22].
After autoradiography, each cDNA probe was removed from the membrane by two 15-min incubations in stripping solution (1× SSC, 0.25% sodium
dodecyl sulfate, and 50% formamide) at 65°C followed by two brief washes in 2× SSC at room temperature. The same blot was then probed with an
albumin DNA fragment to control for potential variations in mRNA loading. The autoradiographs were
scanned with a laser densitometer, and quantitation
was performed as previously described [11] with an
291
CHLOROFORM-INDUCED CHANGES IN mRNA LEVELS
LKB Ultrascan XL Enhanced Laser Densitometer (LKB
Biotechnology, Uppsala, Sweden). The samples on
each autoradiograph were normalized to the values
on the corresponding albumin autoradiograph. Gene
expression was then calculated relative to levels of
that gene in control animals.
populated by vacuolated and regenerating hepatocytes. There was a dose-related increase in labeling
index over control levels at all 238 and 477 mg/kg/d
time points, with the greatest increase following administration of chloroform for 4 d at 477 mg/kg/d.
Differential Display
Sequencing
DNA sequencing was performed using the Taq
DyeDeoxy Terminator Cycle Sequencing Kit (Applied
Biosystems) according to manufacturer’s directions.
Fifty picomoles of the SP6 and T7 primers was added
to a premix of buffer, dNTPs, DyeDeoxy terminators,
and AmpliTaq DNA polymerase. One microgram of
template DNA was added to the mixture and amplified as follows: 96°C for 30 s, 42°C for 30 s, and 60°C
for 4 min for 25 cycles. The reactions were then purified with a Centri-Sep column (Princeton Separations, Adelphia, NJ) and ethanol precipitated. They
were resuspended in a 5:1 mixture of formamide and
EDTA, loaded onto a 6% denaturing sequencing gel,
and analyzed by using the Applied Biosystems Sequencing System (Applied Biosystems).
RESULTS
Hepatotoxicity and Cellular Changes
The patterns of chloroform-induced necrosis and
regenerative cell proliferation that occurred after
chloroform treatment in this study, have been described previously [9]. Briefly, there were no significant changes from controls in either histology or
BrdU labeling index (percentage of hepatocytes in Sphase) of the livers of animals treated with either
1800 ppm chloroform in drinking water or 3 mg/kg/
d chloroform administered by corn-oil gavage. Histological changes at higher doses of chloroform administered by corn-oil gavage ranged from scattered
centrilobular and subcapsular hepatocyte necrosis at
4 d of 238 mg/kg/d to severe centrilobular hepatocyte necrosis at 3 wk of 5 d/wk 238 mg/kg/d to severe centrilobular coagulative necrosis with some
inflammatory cells at 4 d of 477 mg/kg/d. Finally, at
3 wk of 5 d/wk 477 mg/kg/d, the central zone was
Each of the 12 primer pairs (LTK3 paired with each
poly(T)primer) resulted in amplification of about 30
cDNA fragments of about 100–500 bp. One of the
most remarkable observations from these studies was
the high degree of consistency of gene expression
from animal to animal and across dose and treatment groups, as visualized by the differential-display
technique. Of the 387 bands observed, only nine
showed consistent increases or decreases in band
intensity between at least one chloroform treatment
and controls (Table 1). This consistency confirms the
reliability and reproducibility of this technique. These
clones were named to indicate the internal primer
used, the poly(T)primer used, and the approximate
observed size of the band from the polyacrylamide
gel. For example, a clone of approximately 238 bp
originally amplified in a PCR reaction using the LTK3
and the T11CG primers was named LTK3-T11CG-238.
These fragments ranged in size from approximately
120 to 228 bp. Listed in Table 1 are the genes that
were subsequently confirmed by northern blot analysis as changing expression. No differences in mRNA
levels were observed between control animals and
animals treated with 3 mg/kg/d chloroform in corn
oil or 1800 ppm chloroform in drinking water.
Identification of Chloroform-Responsive Clones
All the fragments that initially demonstrated chloroform-induced changes in expression were isolated
and cloned, and the inserts were used as sequencing
templates. The sequences were checked for homologies to known sequences by using DNA databases at
the National Library of Medicine and the BLAST
server [23].
Four genes appeared to change expression when
band intensities were compared on the differential-
Table 1. cDNAs Responsive to Chloroform
Isolated band
LTK3-T11CG-238
LTK3-T11GG-230
LTK3-T11CG-240
LTK3-T11CG-134
LTK3-T11CG-205
Locus name*
Accession number*
Length (bp)‡
MUSTIS21
MUSMRNAH†
MUSMRNAH†
MUSFRA†
MUSFRB†
M64292
L37047
L37047
L37148
L37149
224
228
228
126
197
*Accession number and locus name were assigned by GenBank after sequence submission. Only genes with confirmed changes in expression by
subsequent northern blot analysis are shown in this table. The clones noted in the text (MUSFN, MUSFRC, MUSHEP, and MMU0443) failed to
exhibit the same treatment-related changes in mRNA levels when hybridized to northern blots as initially observed on the differential-display gels.
We consider these clones to be false positives.
†
These cDNAs are novel clones isolated in this study. LTK3-T11GG-230 and LTK3-T11CG-240 were found to be identical except for a single basepair variation.
‡
Length of the cDNA sequence submitted to GenBank. The total reported length of MUSTIS21 is 4886 bp.
292
KEGELMEYER ET AL.
and accession numbers as noted above and in Table
1. (Note that “FR” in the gene names stands for fragment.) The sequences of these clones are shown in
Figure 2. Clone MUSFRB did not hybridize to RNA
taken from rat liver 30 min after partial hepatectomy
(data not shown) and may be mouse specific.
display gel, but no changes in mRNA levels were seen
with subsequent northern blot analysis. The band
designation, locus name (identification), GenBank
accession number, and length of submitted fragment
for these false positives were LTK3-T11CG-214, MUSFN
(mouse fibronectin), M18194, 217 bp (905 bp reported total length); LTK3-T11CG-126, MMU04443
(mouse non-muscle myosin light chain), U04443,
120 bp (654 bp reported total length); LTK3-T11CG235, MUSFRC (novel), L37150, 191 bp; and LTK3T 11 CG-157, MUSHEP (novel), L37349, 148 bp,
respectively.
Clone LTK3-T11CG-238 was 97.4% homologous to
MUSTIS21, a mouse primary response gene induced
by growth factors and tumor promoters. Clones LTK3T11GG-230 and LTK3-T11CG-240 appeared to be the
same gene fragment except for a single base mismatch and had 97% homology to a portion of the
human gene HSCDN12 and 96% homology to the
human gene M78700 (Figure 1). This sequence was
submitted to the GenBank database and was given
the locus name MUSMRNAH. HSCDN12 is a human
gene originally isolated from LNCaP, a prostate carcinoma cell line [24]. M78700 is a human hippocampus cDNA that has not been well characterized
[25]. Clone LTK3-T 11CG-214 was homologous to
mouse fibronectin, and clone LTK3-T11CG-126 was
homologous to mouse non-muscle myosin light chain.
MUSFRA, MUSFRB, MUSFRC, and MUSHEP had
no significant homology to known sequences in
GenBank and were submitted and given locus names
The isolated clones were used as probes to confirm altered expression by northern blot analysis,
examples of which are shown in Figure 3. Of the eight
bands originally identified from the differential-display gels, only four exhibited significantly changed
expression by northern blot analysis (Table 2). Thus,
of the 387 bands originally observed, just four bands
(about 1%) could be confirmed as exhibiting changes
in mRNA levels in regenerating tissue.
Three clones showed an increase in mRNA levels:
MUSTIS21 (Figure 4a), MUSMRNAH (Figure 4b), and
MUSFRA (Figure 4c). One of the isolated clones, MUSFRB, showed a decrease in mRNA levels (Figure 4d).
All four of the confirmed positive clones were also
used as probes to hybridize northern blots containing mRNA from the livers of female mice treated for
up to 3 wk with 1800 ppm chloroform in drinking
water [9]. No treatment-related changes in mRNA
levels of any of the four messages were observed at
any time point (data not shown).
The clone originally identified as LTK3-T11CG-238
had a high degree of homology to a 228-bp segment
of the MUSTIS21 gene in its 3´ untranslated region.
Figure 1. Sequence homologies of MUSMRNAH to the human genes HSCDN12 and M78700. The entire MUSMRNAH sequence is compared with homologous portions of the human
genes. Sequences that differ are surrounded by boxes. Aster-
isks (*) indicate nucleotides that are deleted. Possible stop codon
sites (TAA) are indicated in bold at positions 145 and 188 of
MUSMRNAH. The first 10 nt from the 5´ end are the sequence
of the LTK3 primer.
Northern Blot Analysis
CHLOROFORM-INDUCED CHANGES IN mRNA LEVELS
293
Figure 2. Nucleotide sequences of MUSFRA (a), MUSFRB (b), MUSFRC (c), MUSHEP (d). The 5´ and 3´
primers are underlined.
However, the message sizes were different (3.6 kb
for MUSTIS21 and 2.5 and 1.05 kb for LTK3-T11CG238). To determine if LTK3-T11CG-238 was in fact
MUSTIS21, a 600-bp fragment (TlS21) of a MUSTIS21
clone was used as a probe to hybridize the same
northern blot on which the differential expression
of LTK3-T11CG-238 was confirmed. The TIS21 probe
hybridized to messages of the same size as LTK3T11CG-238, indicating that LTK3-T11CG-238 was homologous to MUSTIS21 (data not shown).
DISCUSSION
Evidence to date indicates the chloroform produces
cancer in rodents by a nongenotoxic-cytotoxic mode
of action [5,7]. A great deal more of supportive
mechanistic research is needed to identify the specific mechanisms that play a role in this mode of
action. One area of interest is the toxicity-mediated
induction and subsequent mutation or overexpression of genes participating in regenerative
growth. These genes are often oncogenes and may
be more susceptible to spontaneous or chemically
induced mutations during expression or may provide a selective growth advantage to precancerous
cells. Identification of such genes may provide insight into some of the mechanisms involved in chloroform-mediated carcinogenesis.
It has been estimated that some mammalian
cells express 15 000 genes [20,26]. Identifying
genes whose expression changes because of toxicity and regenerative cell proliferation and that
may play a role in the carcinogenic process therefore presents a serious challenge. One goal of our
studies was to determine the proportion of genes
294
KEGELMEYER ET AL.
Figure 3. Typical northern blots showing changes in levels
of two mRNAs identified by differential display of livers from
female mice. The numbers across the top denote the chloroform gavage dose administered (mg/kg/d). The numbers below them denote the time point. (a) MUSMRNAH. Note the
increase in intensity of both the 2.8- and 1.4-kb bands at 3 wk,
477 mg/kg/d. (b) MUSFRB. Note the decrease in the intensity of
the 2.1-kb band at 3 wk, 238 mg/kg/d and 4 d, 477 mg/kg/d. (c)
The albumin probe is a constitutively expressed housekeeping
gene used here to control for sample-to-sample variation.
expressed at different levels in the livers of mice
treated under carcinogenic conditions with chloroform. Using the 12 primer sets described here,
we amplified portions of 387 mRNA species, or
about 2.6% of those that might be expressed in a
cell. Of these 387 bands, only four, or about 1%,
changed mRNA levels under regenerative or carcinogenic conditions. An important observation
from these studies was the high degree of consistency of gene expression from animal to animal
and across dose and treatment groups, as visualized by the differential-display technique. This
consistency confirms the reliability and reproducibility of the technique. The degree of expression
was compared across dose and treatment groups
to confirm the consistent pattern of change. The
consistent lack of response in the 3 mg/kg/d (low
dose) and 1800 ppm drinking water groups served
as additional internal controls. Although the differential display technique appears to be a valu-
Table 2. Differentially Expressed Clones Confirmed by Northern Blot Analysis
Locus name
Observed differences*
MUSTIS21
3 wk, 238 mg/kg/d
4 d, 477 mg/kg/d†
3 wk, 477 mg/kg/d†
4 d, 238 mg/kg/d
3 wk, 477 mg/kg/d
4 d, 477 mg/kg/d
3 wk, 477 mg/kg/d
3 wk, 238 mg/kg/d
4 d, 477 mg/kg/d
MUSMRNAH
MUSFRA
MUSFRB
Intensity difference*
mRNA transcript (kb)
3×
3.5×
6×
4×
9×
3×
7×
0.06×
0.06×
2.5
2.8 and 1.4
3.1‡
2.1
*Observed differences are reported as time point and dose. Intensities were analyzed from the northern blots as compared with controls and
were calculated as described in Materials and Methods.
†
There was an additional 1.2-kb band at each of these doses and time points that was not in any of the other samples.
‡
There was an additional band at 1.8 kb in all the samples, but there was no difference in expression levels.
CHLOROFORM-INDUCED CHANGES IN mRNA LEVELS
295
Figure 4. Expression of isolated genes in the livers of female mice after exposure to chloroform by gavage for 4 d
(shaded bars) or 3 wk (solid bars). Gene expression was assessed
by northern blot hybridization, quantitated by densitometry,
and measured relative to control levels as described in Materials and Methods. Standard error is indicated by the error bars.
(a) MUSTIS21; (b) MUSMRNAH (2.8 kb message); (c) MUSFRA;
and (d) MUSFRB.
able tool for examining changes in expression of the
remainder of the genes, that task remains very large.
Interestingly, changes in expression in the four
genes identified occurred only under regenerative or
carcinogenic conditions. Treatment with 3 mg/kg/d
of chloroform or 1800 ppm chloroform in the drinking water (329 mg/kg/d), which does not produce
tumors, also did not produce changes in mRNA expression, patterns indicating that changes in expression are associated with toxicity and regeneration
rather than chloroform per se.
From the 12 amplification reactions, we initially
isolated nine bands that appeared to be differentially
expressed between control mice and mice treated
with 238 or 477 mg/kg/d of chloroform by gavage.
Confirmation of differential expression by northern
blot analysis revealed that while all of the positive
clones hybridized to at least one distinct mRNA, four
of the initial clones did not show a change in expression between control and treated animals. Such
false positives are common in the use of the differ-
ential-display technique [27,28]. We feel that because
duplicate samples across all dose and treatment
groups showed no change in expression, it is unlikely
that there were many false-negative responses. Of
the five remaining clones, sequencing revealed that
two had the same cDNA. Almost all of our positive
clones were isolated from reactions involving the
primer T11CG, a result also observed by Mou et al.
[29]. It is not known why this occurred. Clearly,
there were hundreds of reproducible bands amplified from the other primers, so this does not
appear to be a weakness of either the amplification or isolation procedures.
MUSTIS21 is a primary response gene induced rapidly and transiently in 3T3 cells by the tumor promoter and mitogen tetradecanoyl phorbol acetate
[30]. This gene’s product may mediate extracellular
signals by modulation of RNA splicing [31] and may
also be involved in programmed cell death [32]. We
observed increases in the levels of MUSTIS21 mRNA
after 3 wk of gavage treatment with 238 mg/kg/d of
296
KEGELMEYER ET AL.
chloroform and 4 d and 3 wk of treatment with 477
mg/kg/d of chloroform. Each of these time points
was characterized by the presence of areas of liver
necrosis [9]. With so many changes occurring in the
liver, it is difficult at this point to assign a probable
function (such as the initiation of regenerative cell
proliferation or a response to cell killing), to this or
any of the other genes. The difference observed between the size of the MUSTIS21 message originally
reported by Fletcher et al. [30] of 3.6 kb and the size
of the transcript homologous to MUSTIS21 in our
study may be due to different splicing or processing
of this message in mouse liver versus 3T3 cells, from
which the MUSTIS21 clone was originally isolated.
The MUSMRNAH clone hybridizes to two mRNA
transcripts expressed in mouse liver, one 2.8 kb in
size and the other 1.4 kb. These mRNAs may be the
products of two different genes with a high degree
of homology, or both transcripts may be alternatively
spliced products of the same gene. We observed a
large increase in the levels of both mRNAs after 3 wk
of gavage treatment with 477 mg/kg/d of chloroform
and smaller increases after 4 d of gavage treatment
with 238 mg/kg/d of chloroform. At these time points
and doses, cell replication levels are moderately elevated, but the widespread necrosis seen at either 4
d of 477 mg/kg or 3 wk of 238 mg/kg/d is not present
[9]. Interestingly, the human gene to which
MUSMRNAH has the greatest homology, HSCDN12,
was also originally isolated by using the differentialdisplay technique with the same primers (LTK3 and
T11GG) [24]. The transcript sizes are similar and suggest that we may have isolated the mouse homologue to this human gene. HSCDN12 was originally
isolated from human LNCaP cells and was found to
exhibit minor but reproducible changes in expression between prostate cancer cell lines [24]. The other
close human homologue, M78700, is an uncharacterized gene isolated from the human hippocampus [25].
We observed elevated levels of the MUSFRA transcript after 4 d and 3 wk of treatment with 477 mg/
kg/d of chloroform. This pattern of expression does
not correlate clearly with either the histopathological changes or the changes in cell proliferation seen
over the course of chloroform treatment. However,
that the increase in mRNA levels was observed exclusively in the animals that received the high dose
indicates that the change in expression may be a response to the extreme toxicity of this treatment.
The decrease in expression of clone MUSFRB at 3
wk, 238 mg/kg/d, and 4 d, 477 mg/kg/d occurred in
the tissues with the highest degree of necrosis [9].
However, there was no clear correlation between the
decrease in MUSFRB expression and the amount of
cell proliferation in these tissues. Differences observed at one time point or one dose but not another emphasize the complex nature of the changes
taking place. It was surprising that no genes related
to the cell cycle or cell proliferation were identified
in these studies.
Acute injury is followed by an orderly process in
which expression of key genes increases as long as
regeneration is necessary and returns to control levels when the tissue lost to toxicity has been replaced.
The constant proliferative stimulus provided by
chronic injury provides an environment in which
these genes may acquire mutations or in which the
mechanisms that keep their expression in check become damaged [7,16]. Studies of the type described
here need to be extended to longer periods of exposure. Comparisons of dose-response relationships in
the target organs of animals ranging from no observed effects to precancerous lesions may reveal the
underlying molecular events associated with tumor
formation. Further examination of the clones isolated in this study and of the many as yet unidentified genes that may be changing expression may
clarify the relationships between chloroform-induced
toxicity and the subsequent events that result in tumor development.
ACKNOWLEDGMENTS
We would like to thank Dr. Jeffrey L. Larson for
his work as principal investigator on the original
chloroform experiments; Drs. Chris Corton, Leslie
Recio, Julian Preston, and Tony Fox for advice and
review of the manuscript; and Ms. Kathy Meyer for
making the primers used in these experiments.
REFERENCES
1. Symons JM, Bellar T, Carswell J, et al. National organics reconnaissance survey for halogenated organics. Journal of the American Water Works Association 67:634–647, 1975.
2. Lahl U, Batjer K, Duszeln JV, Gabel B, Stachel B, Thiemann W.
Distribution and balance of volatile halogenated hydrocarbons
in the water and air of covered swimming pools using chlorine
for water disinfection. Water Research 15:803–814, 1981.
3. Deinzer M, Schaumburg F, Klein E. Environmental health sciences center task force review on halogenated organics in drinking water. Environ Health Perspect 24:209–239, 1978.
4. Leuenberger C, Giger W, Coney R, Craydon JW. Persistent
chemicals in pulp mill effluents. Occurrence and behavior in
an activated sludge treatment plant. Water Research 19:885–
894, 1985.
5. Butterworth BE, Conolly RB, Morgan K. A strategy of establishing mode of action of chemical carcinogens as a guide for approaches to risk assessments. Cancer Lett 93:129–146, 1995.
6. International Programme on Chemical Safety. Chloroform, Environmental Health Criteria 163, World Health Organization,
Geneva, 1994.
7. Butterworth BE, Popp JA, Conolly RB, Goldsworthy TL. Chemically-induced cell proliferation in carcinogenesis. In: Vainio H,
Magee PN, McGregor DB, McMichael AJ (eds), Mechanisms of
Carcinogenesis in Risk Identification. International Agency for
Research on Cancer, Lyon, 1992, pp. 279–305.
8. National Cancer Institute. Carcinogenesis Bioassay of Chloroform
(National Tech. Inform. Service No. PB264018/AS). NCI, Bethesda,
MD, 1976.
9. Larson JL, Wolf DC, Butterworth BE. Induced cytotoxicity and
cell proliferation in the hepatocarcinogenicity of chloroform in
female B6C3F1 mice: Comparison of administration by gavage
in corn oil vs ad libitum in drinking water. Fundam Appl Toxicol
22:90–102, 1994.
10. Jorgenson TA, Meierhenry E, Rushbrook C, Bull R, Robinson M.
Carcinogenicity of chloroform in drinking water to male OsborneMendel rats and female B6C3F1 mice. Fundam Appl Toxicol
5:760–769, 1985.
CHLOROFORM-INDUCED CHANGES IN mRNA LEVELS
11. Butterworth BE, Sprankle CS, Goldsworthy SM, Wilson DM,
Goldsworthy TL. Expression of myc, fos, and Ha-ras in the livers
of furan treated F-344 rats and B6C3F1 mice. Mol Carcinog 9:24–
32, 1994.
12. Goldsworthy TL, Goldsworthy SM, Sprankle CS, Butterworth BE.
Expression of myc, fos, and H-ras associated with chemically
induced cell proliferation in the rat liver. Cell Prolif 27:269–
278, 1994.
13. Sprankle CS, Goldsworthy TL, Goldsworthy SM, Wilson DM,
Butterworth BE. Expression of the hepatocyte growth factor and
c-met genes during furan induced regenerative cell proliferation
in the livers of B6C3F11 mice and F-344 rats. Cell Prolif 27:529–
539, 1994.
14. Sprankle CS, Larson JL, Goldsworthy SM, Butterworth BE. Levels
of myc, fos, Ha-ras, met and hepatocyte growth factor mRNA
during regenerative cell proliferation in female mouse liver and
male rat kidney after a cytotoxic dose of chloroform. Cancer Lett
101:97–106, 1996.
15. Kahn P, Graf T. Oncogenes and Growth Control. Springer-Verlag,
Berlin-Heidelberg, 1986.
16. Antoniades HN. Linking cellular injury to gene expression and
human proliferative disorders: Examples with the PDGF genes.
Mol Carcinog 6:175–181, 1992.
17. Bailey A, Saliba D, Miller DM. A marked increase in sensitivity to
carcinogenesis is associated with induced protooncogene expression without DNA synthesis. Proceedings of the American Association for Cancer Research 30:183, 1989.
18. Stewart TA, Pattengale PK, Leder P. Spontaneous mammary adenocarcinomas in transgenic mice that carry and express MTV/
myc fusion genes. Cell 38:627–637, 1984.
19. Adams JM, Harris AW, Pinkert CA, et al. The c-myc oncogene driven by immunoglobulin enhancers induces lymphoid
malignancy in transgenic mice. Nature 318:533–538, 1985.
20. Liang P, Pardee AB. Differential display of eukaryotic messenger
RNA by means of the polymerase chain reaction. Science
257:967–970, 1992.
21. National Institutes of Health. Guide for the care and use of labo-
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
297
ratory animals, U.S. Department of Health and Human Services,
Public Health Service, NIH Publication, 1985.
Goldsworthy S, Goldsworthy T, Sprankle C, Butterworth B.
Variation in expression of genes used for normalization of Northern blots after induction of cell proliferation. Cell Prolif 26:511–
517, 1993.
Altschul S, Gish W, Miller W, Myers E, Lipman D. Basic local
alignment search tool. J Mol Biol 215:403–410, 1990.
Blok L, Kumar MV, Tindall DJ. Isolation of differentially expressed
cDNAs between androgen-dependent and androgen-independent prostate carcinoma cell lines using differential display PCR.
Prostate 26:213–224, 1995.
Adams MD, Dubnick M, Kerlavage AR, et al. Sequence identification of 2,375 human brain genes. Nature 355:632–634, 1992.
Bauer D, Muller H, Reich J, et al. Identification of differentially
expressed mRNA species by an improved display technique (DDRTPCR). Nucleic Acids Res 21:4272–4280, 1993.
Utans U, Liang P, Wyner LR, Karnovsky MJ, Russell ME. Chronic
cardiac rejection: Identification of five upregulated genes in transplanted hearts by differential mRNA display. Proc Natl Acad Sci
USA 91:6463–6467, 1994.
Zhang L, Medina D. Gene expression screening for specific genes
associated with mouse mammary tumor development. Mol
Carcinog 8:123–126, 1993.
Mou L, Miller H, Li J, Wang E, Chalifour L. Improvements to the
differential display method for gene analysis. Biochem Biophys
Res Commun 199:564–569, 1994.
Fletcher B, Lim R, Varnum B, Kujubu D, Koski R, Herschman H.
Structure and expression of TIS21, a primary response gene induced by growth factors and tumor promoters. J Biol Chem
266:14511–14518, 1991.
Lin W, Gary J, Yang M, Clarke S, Herschmann H. The mammalian immediate-early tis21 protein and the leukemia-associated
BTG1 protein interact with a protein-arginine N-methyltransferase. J Biol Chem 271:15034–15044, 1996.
Mesner P, Epting C, Hegarty J, Green S. A timetable of events
during programmed cell death induced by trophic factor withdrawal from neuronal PC12 cells. J Neurosci 15:7357–7366, 1995.