ONCOLOGY REPORTS 21: 283-287, 2009
283
Nuclear translocation of glutathione transferase omega
is a progression marker in Barrett's esophagus
SIMONA PIAGGI1, SANTINO MARCHI2, EUGENIO CIANCIA3, NICOLA DEBORTOLI2,
ALESSANDRA LAZZAROTTI1, MICHELA SAVIOZZI1, CHIARA RAGGI1, VANNA FIERABRACCI1,
ATHANASE VISVIKIS4, HANNE C. BISGAARD5, ALESSANDRO F. CASINI1 and ALDO PAOLICCHI1
1Dipartimento
di Patologia Sperimentale, Biotecnologie Mediche, Infettivologia ed Epidemiologia, Sezione
di Patologia Generale, 2Dipartimento di Medicina Interna, Università di Pisa; 3Azienda Ospedaliera Universitaria
Pisana, Pisa, Italy; 4Centre du Médicament, Université Henri Poincaré, Nancy, France; 5Department of Medical
Biochemistry and Genetics, The Panum Institute, University of Copenhagen, Copenhagen, Denmark
Received June 2, 2008; Accepted September 9, 2008
DOI: 10.3892/or_00000219
Abstract. Barrett's esophagus (BE) represents a major risk
factor for esophageal adenocarcinoma (AC). For this reason,
patients with BE are subjected to a systematic endoscopic
surveillance to detect initial evolution towards non-invasive
neoplasia (NiN) and cancer, that eventually occurs only in
a small fraction of BE patients. This study was aimed to
investigate the possible role of glutathione-S-transferaseomega 1 (GSTO1), a recently discovered member of the
glutathione-S-transferase family, as a progression marker in
the Barrett's disease in order to improve the diagnosis of
NiN in BE and to understand the mechanisms of the
progression from BE to AC. We investigated the expression
and subcellular localization of GSTO1 in biopsies from
patients with BE and in human cancer cell lines subjected to
heath shock treatment. A selective nuclear localisation of
GSTO1 was found in 16/16 biopsies with low- or highgrade NiN, while it appeared in only 4/22 BE biopsies without
signs of NiN (P<0.0001). Among biopsies of BE without
NiN, diffuse (nuclear and cytoplasmic) staining was found in
5/22 cases, while selective cytoplasmic localisation was
found in 13/22. The 6 cases with indefinite grade of NiN
were equally divided between nuclear, cytoplasmic and
diffuse staining (2 each, respectively). Experiments in vitro
_________________________________________
Correspondence to: Dr Simona Piaggi, Dipartimento di Patologia
Sperimentale, BMIE, Sezione Patologia Generale, via Roma 55,
Pisa, Italy
E-mail: s.piaggi@med.unipi.it
Abbreviations: BE, Barrett's esophagus; AC, adenocarcinoma;
NiN, non-invasive neoplasia; LG-NiN, low-grade NiN; HG-NiN,
high-grade NiN; GSTO, glutathione-S-transferase-omega; GERD,
gastroesophageal reflux disease; Hp, Helicobacter pylori
Key words: glutathione transferase, Barrett's esophagus
showed that in human HeLa cancer cells, GSTO1 translocates into the nucleus as a consequence of heath shock.
These findings suggested that the nuclear translocation of
glutathione-S-transferase-omega 1 could be involved in the
stress response of human cells playing a role in the cancer
progression of Barrett's esophagus. Its immunohistochemical
detection could represent a useful tool in the grading of
Barrett's disease.
Introduction
Barrett's esophagus (BE), i.e., the presence of intestinal
metaplasia and goblet cells in biopsies from lower esophagus,
is observed in 5-15% of individuals with chronic gastroesophageal reflux disease (1,2), and represents the main risk
factor for adenocarcinoma (AC) of the esophagus (3,4) AC of
the esophagus may develop through stages from metaplasia to
increasing grade of non-invasive neoplasia (NiN) (low- and
high-grade NiN) (5).
The mechanisms leading to the substitution of squamous
epithelium by specialized intestinal epithelium are poorly
understood, but it is commonly accepted that Barrett esophagus
follows the insufficiency of defence mechanisms of the
esophageal mucosa (luminal secretion of mucus, bicarbonate,
growth factors, etc.) to cope with repeated mucosal injury
(3). Hydrogen ion, pepsin, trypsin and bile acids invade the
esophageal mucosa as a consequence of gastresophageal
reflux, and are thought to act synergistically to harm the
epithelium of the lower part of the esophagus to determine
the intestinal metaplasia (6).
The same agents that cause metaplasia are thought to act
synergistically with carcinogens to cause cancer progression
that leads to NiN and overt adenocarcinoma of the esophagus
(6). Factors determining the survival of proliferating cells
and their ability to withstand the hostile environment of the
Barrett esophagus are thus likely to play a crucial role in
esophageal carcinogenesis.
Recently, a novel family of proteins has been described in
mammalians, characterized by a series of distinctive features:
due to structural similarities with glutathione transferases
284
PIAGGI et al: GLUTATHIONE TRANSFERASE OMEGA IN BARRETT'S DISEASE
(GST), they have been included in a new family, termed
GST-omega (GSTO) (7). Despite belonging to the GSTs
superfamily these proteins lack any appreciable glutathione
transferase activity, but show a series of properties that
indicate them as stress response proteins potentially involved
in cancer progression: indeed, besides localizing into the
cell nucleus after thermal damage in rodents (8), they i)
induce resistance against anticancer drugs and ionizing
radiations (8); ii) reduce the oxidized form of vitamin C,
dehydroascorbic acid, to the native form, ascorbic acid, thus
involving them in the antioxidant defense mechanisms
(7,9-12), iv) catalyse thioltrasferase reaction (7,11) and may
reduce protein disulfides, thus contributing to the recovery
from oxidative challenge to proteins (13). Polymorphisms of
the two GSTO genes found in humans (GSTO1, GSTO2)
have been associated with the risk of hepatocellular carcinoma,
cholangiocarcinoma, breast cancer and ovarian cancer (14-16).
The present study was aimed to ascertain whether human
GSTO1 retains the stress response properties found in
rodents (8), and whether the activation of the human GSTO1
stress response, i.e., its nuclear translocation might serve as
a progression marker of BE towards NiN and adenocarcinoma.
Histology. All specimens were fixed in 10% formalin,
embedded in paraffin, and cut into serial 4-μm-thick sections
for hematoxylin-eosin, PAS-alcian blue (alcian blue at pH 2.5,
1% periodic acid and Schiff's reagent), Giemsa stain, and for
immunohistochemical evaluation.
Immunohistochemistry. Immunohistochemical studies were
performed on sections placed on positively charged glass
slides; after deparaffinization, the tissue sections were
rehydrated, blocked for endogenous peroxidase activity with
3% hydrogen peroxidase in methanol (10 min), and washed
under running water for 5 min and transferred to phosphatebuffered saline. The sections were then blocked with 2% horse
serum for 15 min. The slides were incubated for 30 min with
a 1:1000 dilution of the primary rabbit anti-GSTO1 antibody.
The sections were then incubated with the secondary antibody (Vectastain Elite ABC Kit, Vector Laboratories Inc.,
Burlington, CA) according to the manufacturer's instructions.
The slides were washed with phosphate-buffered saline
between incubations. The tissue sections were developed using
aminoethyl carbazole as the chromogen and counterstained
with hematoxylin. All immunohistochemical stains were
evaluated for the presence, extent and location of GSTO
expression.
Materials and methods
Patients with diagnosis of Barrett's esophagus were identified
from a database of Department of Medical and Surgical
Gastroenterology spanning of a 2-year period (2004-2005).
This approach was necessary to compare two consecutive
groups with intestinal metaplasia, one without non-invasive
neoplasia (NiN) (22 patients), and the other with different
grade of NiN (22 patients). For the sake of comparison, two
esophageal adenocarcinomas were also included in the study.
Patients with previous history of Barret's disease were
excluded from the study.
Patient evaluation and bioptic procedures. All patients were
evaluated with a detailed history before undergoing upper
endoscopy: global health status, the presence or absence of
typical or atypical GERD symptoms (predominant symptom)
and previous drug consumption were investigated.
All endoscopies were performed by the same operator
using standard diagnostic endoscopes. BE was defined as a
detectable upward displacement of the squamo-columnar
junction at endoscopy, confirmed by intestinal metaplasia
with goblet cells at histology. Systematic 4-quadrant biopsy
specimens at 2-cm intervals along the entire length of the
Barrett's segment starting from the gastroesophageal junction
were obtained with standard-size forceps. Agreement on
the presence and extension of BE mucosa and the degree of
esophagitis was obtained in all cases. The esophagitis, if
present, was classified into four grades (A through D)
according to Los Angeles Classification (3).
All patients underwent gastric biopsies from antrum,
greater curvature (two), and lesser curvature (two) for assessment of Helicobacter pylori (Hp) status; additional targeted
biopsies were taken from erosions, nodules or ulcers. The
esophageal biopsies specimen on Barrett's metaplasia were
considered for this study.
Pathological evaluation of biopsies. All biopsies were
evaluated by the same pathologist. Histological slices were
re-evaluated in a blinded manner by the same pathologist to
reduce the risk of intra-observer variability. According to
Sampliner's criteria (4) BE was classified as the presence of
specialized intestinal-type metaplasia with goblet cells. Noninvasive neoplasia (NiN) was evaluated using the criteria of
Rydell (17) and classified in three different groups: negative,
indefinite and positive. Low-grade NiN and high-grade
NiN represent two different subsets of the positive group:
LG-NiN is defined by the presence of nuclear atypia involving
the mucosal surface, nuclear stratification in the base of
crypts, and preserved crypt architecture; HG-NiN is defined
by the presence of marked nuclear atypia, distorted crypt
architecture and nuclear stratification extending to the luminal
surface.
Cell culture and heat shock experiments. HeLa cells were
grown in D-MEM (Sigma) supplemented with 5% heat inactivated serum, 1% L-Glutamine, in a humidified atmosphere
of 5% CO2, at 37˚C. For the heat shock experiments HeLa
cells were incubated for 60, 120 and 180 min at 43˚C in a
humidified atmosphere of 5% CO2. HeLa cells at various
time after heat shock treatment were washed in PBS, scraped
and centrifuged at 1000 g for 2 min. The cell pellet was
washed with 500 μl of buffer B (HEPES 10 mM pH 7.9, KCl
10 mM, MgCl2 2 mM, EDTA 200 μM), centrifuged at 1000 g
2 min and lysed in buffer B with triton X100 1%, leupeptin
4 μg/ml and aprotinin 1 μg/ml on ice for 30 min. The cell
suspension was then centrifuged at 3000 g at 4˚C for 5 min,
and the nuclear pellet was recovered and further washed with
PBS. The nuclear fraction was stored at -20˚C unitl used.
SDS-PAGE electrophoresis and immunoblot. Sodium
dodecylsulphate-polyacrylamide gel electrophoresis was
ONCOLOGY REPORTS 21: 283-287, 2009
285
performed essentially by the method of Laemmli (18), with 5%
acrylamide stacking gel and 15% separating gel. Immunoblot
was performed by transferring proteins to nitrocellulose sheets
(Bio-Rad Laboratories) following the methods of Towbin et al
(19). Immunodetection was performed employing anti-GSTO1
serum and was revealed by peroxidase-labeled anti-rabbit
IgG (Sigma) and BM Chemiluminescence Blotting Substrate
(POD) (Roche).
Immunofluorescence microscopy. HeLa cells were fixed in
4% paraformaldehyde for 20 min on ice; at the end of
incubation paraformaldehyde was removed and the cells
were washed in PBS. For the blocking step the cells were
incubated in 8% milk with 1% Triton X-100 in PBS for 30 min
at room temperature. Then the cells were incubated for 60 min
at room temperature with anti-GSTO1 serum diluted 1:100
in 0.8% milk and 0.1% Triton X-100. After rinsing twice in
PBS, the cells were incubated for 30 min with the secondary
anti-rabbit fluorescein labelled antibody (Sigma) diluted 1:100
in PBS. Images were captured using a Leica DM microscope
(Leica, Germany).
Protein determination. Protein concentrations were
determined using BCA method (Sigma) following the
manufacturer's instruction.
Statistical analysis. Variables corresponding to the percentage
of GSTO1-positive cells (+, ++ or +++, corresponding to
<25, <50 or >50%), and to the prevalent subcellular
localyzation of GSTO1 staining (nuclear, cytoplasmic or
diffuse) in Barrett's epithelium were analyzed by the ¯2 test
(GraphPad Prizm Software).
Results
Subjects. Forty-six patients, 33 male and 12 female, with a
mean age of 62.8 (±14) were included into the study. The
predominant symptoms recorded before diagnostic endoscopy
were: heartburn in 24/46 (52.2%) patients, regurgitation in
11/46 (23.9%), non-cardiac chest pain in 5/46 (10.9%),
epigastric pain in 3/46 (6.5%), chronic cough in 2/46 (4.3%)
and dysphagia in 1/46 (2.2%).
Endoscopic findings detected in the selected patients at
the time of diagnosis were: absence of esophageal flogosis
in 32/46 (69.6%) patients, from A to D grade of esophageal
flogosis in 12/46 (26.1%) and AC in 2/46 (4.3%).
Short-segment Barrett's esophagus (SSBE) (<3 cm) was
detected in 28/44 (63.6%) and long-segment Barrett's
esophagus (LSBE) (>3 cm) in 16/44 (36.4%). Among the
selected patients with BE (44/46) NiN was absent in 22/44
(50%), low-grade NiN (LG-NiN) was present in 7/44
(15.9%), high-grade NiN (HG-NiN) in 9/44 (20.5%) and
indefinite grade of NiN in 6/44 (13.6%).
Infection with H. pylori was histologically detected in
gastric biopsies in 8/45 (17.8%) of selected patients. No
previous treatment with proton pump inhibitors (PPI) or
histamine H2 receptor antagonists had been performed within
the previous year in all patients.
Immunohistochemistry of GSTO1. In the 46 biopsies of the
lower esophagus studied, immunohistochemically detectable
Figure 1. Immunolocalisation of GSTO1 in esophageal biopsies of subjects
with Barrett's disease. Counterstaining: haematoxylin. (A) Barrett's esophagus
without NiN. GSTO1 stainining involves the cytoplasm of the metaplastic
cells and appears to be more intense in the basal portion of the cells, while
most nuclei stain negative. Original magnification x400. (B) Low-grade
NiN, many dysplastic cells show a selective nuclear staining for GSTO1.
Original magnification x400. (C) High-grade NiN, glandular acinus with
intense and selective nuclear staining in most cells. Original magnification
x1000.
GSTO1 was found to be localized in the surrounding
esophageal epithelium, cardial and fundic mucosa, and
diffuse low-grade staining was found in the mesenchyma
between the glands (Fig. 1A and B). Particularly intense
staining was found in endothelial cells. Among the 44 patients
with BE, absence of NiN was found in 22, while 7 showed
low-grade NiN, and 9 high-grade NiN; 6 were indefinite for
NiN.
286
PIAGGI et al: GLUTATHIONE TRANSFERASE OMEGA IN BARRETT'S DISEASE
Figure 2. Effect of heat shock treatment at 43˚C on subcellular localization of GSTO1 in HeLa cells. In (A) nuclear extracts of Hela cells at different time
intervals after initiation of thermal shock lane 1, control culture; lane 2, 60 min; lane 3, 120 min; lane 4, 180 min; Lane 5 human recombinant GSTO1. In (B)
and (C), immunofluorescence of GSTO1 before and after 180 min of heath shock, respectively.
Among the 22 cases of Barrett's disease without NiN,
5/22 cases had <25% cells (+) stained positive for GSTO,
9/22 had 50% (++) and 8 >50% positive cells (+++). Among
these cases without NiN, 4 showed prevalent nuclear localization, 13 cytoplasmic localization (Fig. 1A), and 5 diffuse
cellular staining; no correlation was found between the
proportion of cells stained and the localization of the staining.
Out of the 22 cases with NiN, all the 7 cases (4+ and 3++)
with low-grade NiN and the 9 cases with high-grade NiN
(5++ and 4+++) showed prevalent nuclear localization of
GSTO1 immunostaining (Fig. 1B and C). The prevalence
of nuclear localization of GSTO1 among the cases with
NiN was significantly higher than among cases without NiN
(P<0.0001). Out of the 6 cases not determined for NiN, 2
showed nuclear, 2 cytoplasmic and 2 diffuse staining. Of the
2 cases of adenocarcinoma of the esophagus, one case (+++)
showed prevalent nuclear localization, and the other showed
prevalent nuclear localization in the more differentiated areas
(+++), and diffuse staining in the less differentiated (++).
Heat shock treatment. To investigate whether in human cells
GSTO1 maintains the shock protein properties found in mice,
HeLa cells, that were found to express GSTO1, were subjected
to thermal stress at 43˚C for different times. The nuclear
translocation of immunoreactive protein to the nucleus started
at the end of the first hour of incubation and reached the
maximum level in 3 h (Fig. 2A). The nuclear translocation of
GSTO1 was confirmed by immunofluorescence studies that
showed the negative image of the nucleus in control cells
(Fig. 2B), and the prevalent nuclear positivity after thermal
shock (Fig. 2C).
Discussion
The glutathione transferases omega (GSTOs) are a new class
of glutathione transferase that show a series of distinctive
features which could make them relevant for cancer
progression. In mice GSTO1 acts as a shock protein, translocating into the nucleus as a consequence of thermal stress,
while in murine leukemia cells its overexpression was found
to associate with increased resistance to ionizing radiations
(8). Incresed expression of GSTO was found also in cisplatinresistant ovarian cancer SKOV3 cells (20) and polymorphisms
of the two GSTO genes found in humans (GSTO1, GSTO2)
have been associated with the risk of hepatocellular carcinoma,
cholangiocarcinoma, breast cancer and ovarian cancer (15,16).
In Barrett's esophagus the progression to invasive AC
may develop through stages from metaplasia to increasing
grade of NiN. Esophageal carcinogenesis is thus a multistep
process of genetic instability due to the accumulation of the
successive mutations in the columnar intestinal epithelial
cells of BE, that lead to NiN and eventually to the cancer
(5).
In the present work we found a selective nuclear localization of GSTO1 in all biopsies with NiN, while diffuse or
selective cytoplasmatic staining was found only in most BE
biopsies without signs NiN. Our main finding was thus that
the selective nuclear translocation of GSTO1 is a marker of
NiN in Barrett's esophagus, suggesting a possible use of
GSTO1 immunostaining for improving the diagnosis of NiN
in BE. In addition we found, for the first time, that human
GSTO1 translocates into the nucleus as a consequence of
heath shock in human cancer cells. Although the biological
consequences of the nuclear translocation of GSTO1 in
Barrett cells are unknown, this finding supports the hypothesis
that the nuclear translocaton of GSTO1 is involved in the
mechanisms of adaptation of dysplastic cells to the hostile
environment of the lower esophagus, thus participating to the
neoplastic progression towards esophageal cancer.
The involvement of GSTO1 translocation could extend
beyond the simple adaptation to dysplastic environment, in
fact among the several functions of GSTO1 ascertained until
now, some could directly influence the mechanisms controlling
cell proliferation and differentiation, in particular the
ONCOLOGY REPORTS 21: 283-287, 2009
thioltransferase activity of GSTO1 (8,11) might be involved
in the control of redox-regulated transcription factors or
signalling elements, such as AP-1 or NF- κ B, a function
already demonstrated for other proteins such as thioredoxin
(21,22). In view of this, understanding the functional effects
of the polymorphism found to be associated with increased
susceptibility to cancer could help in understanding the
precise role of GSTOs in cancer progression.
In conclusion, the nuclear translocation of GSTO1 in
Barrett's NiN suggests a specific function of human GSTO1
in the progression of esophageal cancer. Although further
studies are needed to establish the precise biological
relevance of this phenomenon and the functions of the
translocated protein, the subcellular localization of GSTO1 in
bioptic specimens could provide, in addition to P53 and
Ki67, a useful immunohistochemical diagnostic tool to
increase specificity and sensitivity of the diagnosis of NiN in
the Barrett's disease.
Acknowledgements
This study was supported by a 2007 PRIN grant of the Italian
Ministry of University.
References
1. Winters C Jr, Spurling TJ, Chobanian SJ, et al: Barrett's
esophagus. A prevalent, occult complication of gastroesophageal
reflux disease. Gastroenterology 92: 118-124, 1987.
2. Csendes A, Smok G, Burdiles P, Korn O, Gradiz M, Rojas J
and Recio M: Prevalence of intestinal metaplasia according to
the length of the specialized columnar epithelium lining the
distal esophagus in patients with gastroesophageal reflux. Dis
Esophagus 16: 24-28, 2003.
3. Falk JW: Barrett's esophagus. Gastroenterology 122: 15691591, 2002.
4. Cameron AJ, Lomboy CT, Pera M and Carpenter HA:
Adenocarcinoma of the esophagogastric junction and Barrett's
esophagus. Gastroenterology 109: 1541-1546, 1995.
5. Pera M, Cameron AJ, Trastek VF, et al: Increasing incidence of
adenocarcinoma of the esophagus and esophagogastric junction.
Gastroenterology 104: 510-513, 1993.
6. Cameron AJ, Ott BJ and Payne WS: The incidence of
adenocarcinoma in columnar-lined (Barrett's) esophagus. N
Engl J Med 313: 857-859, 1985.
7. Board PG, Coggan M, Chelvanayagam G, Easteal S, Jermiin LS,
Schlte GK, Danley DE, Hoth LR, Griffor MC, Kamath AV,
Rosner MH, Chrunyk BA, Perreagaux DE, Gabel CA,
Georghean FK and Pandit J: Identification, characterization, and
crystal structures of the Omega Class Glutathione Trasferases. J
Biol Chem 275: 24798-24806, 2000.
287
8. Kodym R, Calins P and Story M: The cloning and characterization of a new stress response protein a mammalian member
of a family of teta class glutathone S-transferase-like protein. J
Biol Chem 274: 5131-5137, 1999.
9. Maellaro E, Del Bello B, Sugherini L, Santucci A, Comporti M
and Casini AF: Purification and caracterization of glutathionedependent dehydroascorbate reductase from rat liver. Biochem J
301: 471-476, 1994.
10. Xu DP, Washburn MP, Sun GP and Wells WW: Purification
and caracterization of a glutathione-dependent dehydroascorbate
reductase from human erithrocytes. Biochem Biophys Res
Commun 221: 117-121, 1996.
11. Schmuck EM, Board PG, Whitbread AK, Tetlow N,
Blackburn AC and Masoumi A: Characterization of the
monomethylarsonate reductase and dehydroascorbate reductase
activity of Omega class glutathione transferase variant.
Pharmcogenet Genomics 15: 493-501, 2005.
12. Ishikawa KA, Casini AF and Nishikimi M: Molecular cloning
and functional expression of rat liver glutathione-dependent
dehydroascorbate reductase. J Biol Chem 273: 28708-28712,
1998.
13. Whitbread AK, Masoumi A, Tetlow N, Schmuck E, Coggan M
and Board PG: Characterizatio of the omega class of glutathione
transferases. Methods Enzymol 401: 78-102, 2005.
14. Whitbread AK, Tetlow N, Eyre HJ, Sutherland GR and Board PG:
Characterizathion of human Omega class gluthathione transferase genes and associated polymorphisms. Pharmacogenetics
13: 131-144, 2003.
15. Marahatta SB, Punyarit P, Bhudisawasdi V, Paupairoj A,
Wongkham S and Petmitr S: Polymorphism of glutathione Stransferase Omega gene and risck of cancer. Cancer Lett 236:
276-281, 2006.
16. Pangstaporn W, Rochanawutanon M, Wilailak S, Linasamita V,
Weerakiat S and Petmitr S: Genetic alteration in chromosome
10q24.3 and glutathione S-transferase Omega 2 polymorphism
in ovarian cancer. J Exp Clin Cancer Res 25: 107-114, 2006.
17. Spechler SJ, Robbins AH, Rubins HB, Vincent ME, Heeren T,
Doos WG, Colton T and Schimmel EM: Adenocarcinoma and
Barrett's esophagus. An overrated risk? Gastroenterology 87: 927933, 1984.
18. Laemmli UK: Cleavage of structural proteins during the
assembly of the head of bacteriophage T4. Nature 227: 680-685,
1970.
19. Towbin H, Stachelin P and Gordon J: Electrophoretic transfer of
protein from polyacrylamide gels to nitrocellulose sheets:
procedure and some applications. Proc Natl Acad Sci USA 76:
4350-4354, 1976.
20. Yan XD, Pan LY, Yuan Y, Lang JH and Mao N: Identification
of platinum-resistance associated proteins through proteomic
analysis of human ovarian cancer cells and their platinumresistant sublines. J Proteome Res 6: 772-780, 2007.
21. Weichsel A, Casdaska JR, Powis G and Montfort WR: Crystal
structures of reduced, oxidized, and mutated human thiredoxin:
evidence for a regulatory homodimer. Structures 4: 735-751,
1996.
22. Arner ES and Holmgren A: Physiological functions of
thioredoxin and thioredoxin reductase. Eur J Biochem 267:
6102-6109, 2000.