Changes in ultrastructure and the occurrence of permeability
transition in mitochondria during rat liver regeneration
Ferruccio Guerrieri
1,
*, Giovanna Pellecchia
1
, Barbara Lopriore
1
, Sergio Papa
1
, Giuseppa Esterina Liquori
2
,
Domenico Ferri
2
, Loredana Moro
3
, Ersilia Marra
3
and Margherita Greco
3
1
Department of Medical Biochemistry and Biology, University of Bari, Italy;
2
Department of Zoology, Laboratory of Histology and
Comparative Anatomy, University of Bari, Italy;
3
Center for the Study of Mitochondria and Energy Metabolism (CNR) Bari, Italy
Mitochondrial bioenergetic impairment has been found in
the organelles isolated from rat liver during the prereplicative
phase of liver regeneration. To gain insight into the mech-

membrane permeability; calcium; cyclosporin-A.
Seventy percent partial hepatectomy (PH) induces cell
proliferation until the original mass of the liver is restored
[1]. The tissue regeneration process consists of two phases:
the prereplicative phase, the duration of which depends on
the age of the animal [2,3] as well as on hormones and
dietary manipulation [2,4] and the replicative phase, during
which a sharp increase in DNA synthesis occurs with active
mitosis [2]. In the light of early changes in ATP concentra-
tion found in liver after PH, before activation of cell
proliferation [5,6], mitochondria were investigated as they
are directly involved in the process of liver regeneration
[4,7–16]. Many mitochondrial functions, including oxidative
phosphorylation [11–13] and generation of reactive oxygen
species [14,15], were investigated in some detail in the
prereplicative phase of liver regeneration. In isolated
mitochondria, a decrease in the respiratory control index
[12], ATP synthesis, probably due to a decrease in the
ATPsynthase complex content [14], and glutathione content
[13] as well as an increase in malondialdehyde production
[14] and oxidant production [15] were found. This suggests
the occurrence in the prereplicative phase of liver regener-
ation of a transient mitochondrial oxidative stress in which
mitochondria can also release proteins from the matrix [16].
Despite this, mitochondria recover their functions in the
replicative phase of liver regeneration [12,14–16].
In this paper, we investigated whether and how the
mitochondrial structure can change in the prereplicative
phase of liver regeneration and whether mitochondrial
permeability properties are somehow affected in this phase

enzyme units.
Enzymes: aspartate aminotransferase (EC 2.6.1.1); glutamate
dehydrogenase (EC 1.4.1.2).
*Note: deceased in November 2000.
(Received 8 February 2002, revised 20 May 2002,
accepted 22 May 2002)
Eur. J. Biochem. 269, 3304–3312 (2002) Ó FEBS 2002 doi:10.1046/j.1432-1033.2002.03010.x
sacrificed. The livers were removed, weighed, and processed
as follow: one-third were cut into sections for electron
microscopy studies and two-thirds were used for the
isolation of mitochondria. Sham-operated rats, obtained
after a small midline abdominal incision without excision of
the liver, were used as a control and killed at 0, 24 and 96 h
after the surgical operation. In all the assays reported, no
difference between sham-operated and rats that did not
receive any surgical operation was observed.
All operations were carried out under sterile conditions.
The animals received humane care and the study was
approved by the State Commission on animal experimen-
tation.
Electron microscopy
Ultrastructural morphology of mitochondria was deter-
mined by electron microscopy. Liver specimens from
control rats and from rats at 24 and 96 h after PH, were
fixed with 4% glutaraldehyde in 0.1
M
sodium cacodylate
buffer pH 7.4 for 4 h at 4 °C. After fixation and an
overnight wash in sodium cacodylate buffer at 4 °C, the
specimens were postfixed with 1% osmium tetroxide in

mitochondria, the supernatant was used for preparation of
cytosol by ultracentrifugation at 105 000 g for 1 h. The final
supernatant was used as cytosolic fraction. In the prepara-
tions used for measurements of mitochondrial Ca
2+
content, 1.6 l
M
ruthenium red and 1 m
M
EDTA were
added in the isolation buffer to restrict Ca
2+
movement
during the subfractionation technique. As preliminary
analyses showed that there was no statistically significant
difference in the Ca
2+
content of mitochondria whether the
buffers used for the subfractionation procedure contained
either 1 m
M
EDTA alone, or 1 m
M
EDTA and 1.6 l
M
ruthenium red or 1 m
M
EGTA, for all subsequent prepa-
rations, 1 m
M

Milano, Italy) was added to the reaction medium.
Matrix proteins release assay
For the assay of the in vitro release of matrix proteins,
mitochondria (10 mg proteinÆmL
)1
) were suspended in the
swelling medium, above reported, and incubated at 25 °C
for 8 min. After incubation, the mitochondria were preci-
pitated by centrifugation at 8000 g for 40 s. The superna-
tants were then centrifuged for 10 min at 10 000 g.Five
microliters of the final supernatants were used for SDS/
PAGE analysis with a linear gradient of polyacrylamide
(10–15%) [20]. After the run, the gel was stained with
Coomassie Brilliant Blue. Where indicated, mitochondrial
aspartate-aminotransferase [16] (AAT) or glutamate-dehy-
drogenase (GDH) [21] activities were determined in the final
supernatants. When indicated, CsA (1.7 nmolÆmg
)1
mito-
chondrial proteins) was added. The activities of the two
enzymes were also determined in the mitochondrial and
cytosolic fractions, and in the whole liver homogenate. The
enzyme activity of mitochondrial AAT in the cytosol was
determined as described by Greco et al. [16]. Briefly, two
aliquots of either cytosolic fraction or whole homogenate
were incubated separately at 37 °Cand70°C for 15 min,
then AAT activity in both samples was determined. The
AAT activity of the sample incubated at 37 °Cwastakento
be that of both isoenzymes (mitochondrial and cytosolic
AAT), whereas that of the sample incubated at 70 °Cwas

For determination of the endogenous Ca
2+
content,
mitochondria (0.1 mg proteinÆmL
)1
) were suspended in
0.25
M
sucrose in the presence of 40 l
M
Arsenazo III
(Sigma–Aldrich, Milan, Italy). The absorbance change at
675–685 nm, was monitored by dual wavelength spectro-
photometry. After reading a baseline for 1 min, Triton
X-100 (0.2%) plus 3.3 l
M
SDS were added to disrupt the
mitochondrial membranes [25]. The absorbance change was
calibrated by addition of standard aliquots of Ca
2+
to the
medium. A standard curve was obtained from the pooled
results of five independent series of determinations and used
for analysis of mitochondrial Ca
2+
content, which for the
control was 8 ± 0.2 nmol per mg mitochondrial protein.
No statistically significant differences in Ca
2+
content were

dilated and paled matrix, lack of dense granules (Fig. 1C);
and (c) altered mitochondria (*) with clear vacuolization of
the matrix compartment (Fig. 1D). No evident rupture of
mitochondrial outer membrane integrity was observed in
altered mitochondria. At 96 h after PH (Fig. 1E), mito-
chondria were nearly normal in morphology, cristae-rich,
and with an electron-dense matrix. Quantitation of normal
and altered mitochondria in control liver and in liver at
24 and 96 h after PH was performed. The majority of liver
mitochondria from control rats presented a normal mor-
phology; only a small fraction (3.0 ± 0.6%) belonged to
the altered type. A large proportion (41.0 ± 6.6%) of
mitochondria from liver at 24 h after PH showed alterations
in mitochondrial ultrastructure. At 96 h after PH, only a
small fraction (3.0 ± 0.05%) belonged to the altered type.
The differences between the number of altered mitochon-
dria at 24 h after PH and the number of altered mito-
chondria in control rats were statistically significant
(P < 0.0001). Furthermore, in liver at 24 h after PH the
total number of mitochondria, counted in 10 randomly
selected electron micrographies of a hepatic lobule, was less
than the total number present in either control liver (11%
decrease; P ¼ 0.001) or in liver at 96 h after PH (17%
decrease; P < 0.001). The decrease in the mitochondria
number corresponds to a decrease in the mitochondrial
proportion of the cell volume at 24 h after PH. This was
correlated with a decrease in the activity of the mitochon-
drial marker enzymes GDH and mAAT in the total liver
homogenate at 24 h after PH (15% and 24% decrease
for GDH and mAAT, respectively). Moreover, in the

indicative of the occurrence of permeability transition in
mitochondria during the prereplicative phase of liver
regeneration. Thus we checked whether the isolated mito-
chondria could release matrix proteins into the external
medium. Incubation of rat liver mitochondria, isolated at
24 h after PH, at 25 °C for 8 min in the swelling medium,
resulted in an increased and nonspecific release of mito-
chondrial proteins in the suspension medium (Fig. 3A, lane
c) compared to mitochondria isolated from control rats
(Fig. 3A, lane b) and mitochondria isolated at 96 h after PH
(Fig. 3A, lane d), as revealed by SDS/PAGE of the
supernatants obtained after precipitation of mitochondria
by centrifugation. This release of proteins at 24 h after PH
was associated with the appearance, in the supernatant, of
typical matrix enzyme activity, such as GDH (3.5 ± 0.26-
fold increase vs. control mitochondria; 23 ± 2.5% of the
total mitochondrial activity) and AAT (3.15 ± 0.23-fold
increase vs. control mitochondria; 5.1 ± 0.1% of the total
mitochondrial activity) (Fig. 3B, empty columns b). CsA,
added to the mitochondrial suspensions before incubation,
inhibited the release of enzyme activities (Fig. 3B, filled
columns b). At 96 h after PH, the activities of the enzymes
released in the supernatant (1.8 ± 0.1 and 0.8 ± 0.04% of
the total mitochondrial activity of GDH and AAT,
respectively), were as low as those found in the supernatant
Fig. 2. Absorbance changes at 540 nm of rat liver mitochondria isolated
during liver regeneration. Mitochondria (0.5 mg proteinÆmL
)1
)isolated
at 0, 24, 96 h after PH were suspended in swelling medium and the

observed in cytosols isolated from liver control and liver at
24 and 96 h after PH (Fig. 4B).
Ca
2+
content in mitochondria during liver regeneration
after PH
The occurrence of mitochondrial permeability transition is
due to an increase in mitochondrial Ca
2+
content [27].
Consistently, Ca
2+
pulse to mitochondria isolated before
PH or from sham-operated rats and suspended in an
isotonic sucrose medium supplemented with succinate and
phosphate, caused mitochondrial swelling (Fig. 5A), which
reflects a change in mitochondrial membrane permeability
[19]. Such a mitochondrial swelling was inhibited by the
addition to the mitochondrial suspension of CsA (Fig. 5A),
the specific inhibitor of the permeability transition pore of
mitochondria [26]. This change in permeability of the inner
mitochondrial membrane due to Ca
2+
loading was accom-
panied by a nonspecific release of mitochondrial proteins in
the suspension medium [28] with the appearance, in the
supernatants, of typical matrix enzyme activities, such as
mitochondrial AAT, the release of which was also inhibited
by the addition of CsA (Fig. 5B).
As the mitochondrial permeability transition is dependent

DISCUSSION
Following PH, the remaining mature hepatocytes enter a
complex process, known as liver regeneration, which after
an initial prereplicative phase reconstitutes the original mass
of the liver [1,2]. The residual hepatocytes re-enter the cell
cycle while the normal homeostatic mechanisms that couple
cell cycle re-entry to cell death are suspended [29,30].
The present study shows that after surgical removal of
two-thirds of the mass of rat liver, mitochondria in the
Fig. 3. Release of matrix proteins from rat liver mitochondria isolated during liver regeneration. (A,B) Mitochondria (10 mg proteinÆmL
)1
)were
suspended in the swelling medium and incubated at 25 °C for 8 min. After incubation, mitochondria were precipitated by centrifugation at 8000 g
for 40 s. The supernatants were, then, centrifuged for 10 min at 10 000 g. (A) Five microliters of the final supernatant was analyzed by SDS/PAGE;
lane a, standard M
r
proteins; lane b, supernatant from control mitochondria; lane c, supernatant from mitochondria isolated 24 h after PH; lane d,
supernatant from mitochondria isolated 96 h after PH. (B) GDH and AAT activities released in the supernatants of control mitochondria (columns
a), mitochondria isolated 24 h after PH (empty columns b), mitochondria isolated 96 h after PH (empty columns c). The enzyme activities in the
presence of 1.7 nmolÆmg
)1
protein CsA added to the incubation medium are reported as filled columns (b and c). The data are the means (± SEM)
of five different mitochondrial preparations. The differences between both GDH and AAT activity at 24 h after PH and the same activities in the
supernatants of control mitochondria are statistically significant (*P<0.001).
3308 F. Guerrieri et al. (Eur. J. Biochem. 269) Ó FEBS 2002
Fig. 4. Glutamate-dehydrogenase, mitochondrial aspartate amino-
transferase activities and cytochrome c content in mitochondria and
cytosol prepared during liver regeneration. (A) Mitochondrial AAT and
GDH activities were measured in mitochondria and cytosol isolated
from liver control (columns a, a¢), at 24 h (columns b, b¢)and96h

proteinÆmL
)1
) were added to the isotonic sucrose medium (swelling
medium) reported in Materials and methods and the absorbance
change at 540 nm at 25 °C was monitored. After 4 min, 150 l
M
CaCl
2
was added. The dotted line shows the same experiment run in the
presence of 1 l
M
CsA added to the suspension medium before mito-
chondria. (B) AAT activity in the supernatant of liver mitochondria
incubated 8 min in the swelling medium (column a) or in the swelling
medium after a Ca
2+
pulse (70 nmolÆmg protein
)1
)(columnb).
Column c: as column b in the presence of CsA (1.7 nmolÆmg pro-
tein
)1
). The data reported are means (± SEM) of five different
experiments. The differences between AAT activity in the presence of
Ca
2+
and AAT activity in the absence of Ca
2+
pulse are statistically
significant (*, P < 0.001).

Ó FEBS 2002 Mitochondria and liver regeneration (Eur. J. Biochem. 269) 3309
remaining hepatocytes undergo, in the first 24 h after
hepatectomy, i.e. in the prereplicative phase, ultrastructural
changes. These are associated with enhancement of the
mitochondrial Ca
2+
content and increase of CsA-sensitive
permeability to sucrose of the mitochondria isolated from
the residual liver mass.
Analysis of the structural and functional state of mito-
chondria in the liver mass which is reconstituted in the
successive 96 h, shows, on the other hand, normal mito-
chondrial ultrastructure, return of mitochondrial Ca
2+
content and CsA-sensitive sucrose permeability to the
normal values observed in the liver before hepatectomy or
in sham-operated rats.
Previous electron microscopy studies [15,31–33] had
revealed changes in the residual hepatocytes after PH but
less attention was paid to elucidating the correlation
between the changes occurring in the ultrastructure of
mitochondria and biochemical parameters during liver
regeneration. The present electron microscopy study shows
that the general organization of the mitochondrial inner
membrane cristae into the typical transverse alignment in
control animals was absent in about 40% of the mitochon-
dria in the hepatocytes at 24 h after PH. These mitochon-
dria were characterized by highly fractured and degenerated
cristae and a clear vacuolation. This suggests that the
decrease in ATP synthesis rate observed in mitochondria

also observed [14,15]. Following PH, an increase in cell
Ca
2+
content has been observed during the prereplicative
phase of liver regeneration [35]. HGF, the most important
in vitro mitogen for primary hepatocytes and whose plasma
level increases within 1 h upon PH [29,36], has been shown
to induce Ca
2+
entry across the hepatocyte plasma
membrane [37]. Furthermore, some hormones, that are
known to modulate liver regeneration acting as mitogens or
comitogens [29,36], raise the liver cytosolic Ca
2+
concen-
tration and cause an increase in the mitochondrial matrix
volume as a consequence of Ca
2+
entry from cytosol into
mitochondria [38].
Both mitochondrial Ca
2+
accumulation and oxida-
tive stress increase the probability that changes in the
mitochondrial membrane permeability occur [25,38,39].
Oxidative stress, Ca
2+
uptake and opening of the transition
pore in mitochondria are signals for cell death [40–42].
However, only a transient small increase in the number of

great proportion of mitochondria undergoing permeability
transition recover in a fully reversible manner. Future
studies will be needed to ascertain the fate of mitochondrial
subpopulations during liver regeneration.
ACKNOWLEDGEMENT
This work was partially supported by a grant within the National
Research Project PRIN: ÔBioenergetics and Membrane TransportÕ of
Murst, Italy.
REFERENCES
1. Michalopoulos, G.K. (1990) Liver regeneration: molecular
mechanisms of growth control. FASEB J. 4, 176–187.
2. Steer, C.J. (1995) Liver regeneration. FASEB J. 9, 1396–1400.
3. Guerrieri,F.,Kalous,M.,Capozza,G.,Muolo,L.,Drahota,Z.&
Papa, S. (1994) Age dependent changes in mitochondrial F
0
F
1
-
ATP synthase in regenerating rat liver. Biochem. Mol. Biol. Int. 33,
117–129.
4. Guerrieri, F., Nicoletti, C., Adorisio, E., Caraccio, G., Leonetti, P.,
Zanotti, F. & Cantatore, P. (2000) Correlation between decreased
expression of mitochondrial F
0
F
1
-ATP synthase and low
regenerating capability of the liver after partial hepatectomy in
hypothyroid rats. J. Bioenerg. Biomembr. 32, 183–191.
3310 F. Guerrieri et al. (Eur. J. Biochem. 269) Ó FEBS 2002

Yanabu, N., Tokuka, A., Kitai, T., Ozawa, K. & Yamaoka, Y.
(1994) Changes in the distribution of the control of the
mitochondrial oxidative phosphorylation in regenerating rabbit
liver. Biochem. Biophys. Acta 1188, 311–317.
12. Guerrieri,F.,Muolo,L.,Cocco,T.,Capozza,G.,Turturro,N.,
Cantatore, P. & Papa, S. (1995) Correlation between rat liver
regeneration and mitochondrial energy metabolism. Biochem.
Biophys. Acta 1272, 95–100.
13. Vendemmiale, G., Guerrieri, F., Grattagliano, I., Didonna, D.,
Muolo, L. & Altomare, E. (1995) Mitochondrial oxidative phos-
phorylation and intracellular glutathione compartmentation dur-
ing rat liver regeneration. Hepatology 21, 1450–1454.
14. Guerrieri, F., Vendemiale, G., Grattagliano, I., Cocco, T.,
Pellecchia,G.&Altomare,E.(1999)Mitochondrial oxidative
alterations following partial hepatectomy. Free Rad. Biol. Med. 26,
34–41.
15. Lee,F.Y.J.,Li,Y.,Zhu,H.,Yang,S.Q.,Lin,H.Z.,Trush,M.&
Diehl,A.M.(1999)Tumornecrosisfactorincreasesmitochondrial
oxidant production and induces expression of uncoupling protein-
2 in the regenerating rat liver. Hepatology 29, 677–687.
16. Greco, M., Moro, L., Pellecchia, G., Di Pede, S. & Guerrieri, F.
(1998) Release of matrix proteins from mitochondria to cytosol
during the prereplicative phase of liver regeneration. FEBS Lett.
427, 179–182.
17. Reynolds, E.S. (1963) The use of lead citrate at high pH as electron
opaque stain in electron microscopy. J. Cell. Biol. 40, 395–414.
18. Bustamante, E., Soper, J.W. & Pedersen, P.L. (1977) High yield
preparative method for isolation of rat liver mitochondria. Anal.
Biochem. 80, 401–408.
19. Petronilli, V., Cola, C., Massari, S., Colonna, R. & Bernardi, P.

27. Gunter, T.E. & Pfeiffer, D.R. (1990) Mechanisms by
which mitochondria transport calcium. Am. J. Physiol. 258,
C755–C786.
28. Igbavboa, V., Zwizinski, C.W. & Pfeiffer, D.R. (1989) Release of
mitochondrial matrix proteins through a Ca
2+
-requiring,
cyclosporin–sensitive pathway. Biochem. Biophys. Res. Commun.
161, 619–623.
29. Michalopoulos, G.K. & De Frances, M.C. (1997) Liver
regeneration. Science 276, 60–66.
30. Diehl, A.M. (1998) Role of CCAAT/enhancer-binding proteins in
regulation of liver regenerative growth. J. Biol. Chem. 273, 30843–
30846.
31. Jordan, S.W. (1964) Electron microscopy of hepatic regeneration.
Exp. Mol. Pathol. 3, 183–200.
32. Becker, F.F. & Lane, B. (press) P. (1966) Regeneration of mam-
malian liver: evidence of role of cytoplasmic alterations in pre-
paration for mitosis. Am. J. Pathol. 49, 227–235.
33. Gutierrez-Salinas, J., Aranda-Fraustro, A., Paredes-Diaz, R. &
Hernandez-Munoz, R. (1996) Sucrose administration to partially
hepatectomized rats: a possible model to study ethanol-induced
inhibition of liver regeneration. Int. J. Biochem. Cell Biol. 28,
1007–1016.
34. Shimizu, S., Kamiike, W., Hatanaka, N., Nishimura, M.,
Miyata, M., Inane, T., Yoshida, Y., Tagawa, K. & Matsuda, H.
(1994) Enzyme release from mitochondria during reoxygenation
of rat liver. Transplantation 57, 144–148.
35. Takahasi, H. & Yamaguchi, M. (1996) Enhancement of plasma
membrane (Ca

177–196.
42. Bernardi, P. (1999) Perspectives on the permeability transition
pore, a mitochondrial channel involved in cell death. In: Frontiers
in Cellular Bioenergetics. (Papa,S.,Guerrieri,F.&Tager,J.M.,
Ó FEBS 2002 Mitochondria and liver regeneration (Eur. J. Biochem. 269) 3311
eds), pp. 773–795. Kluwer Academic/Plenum Publishers, New
York.
43. Zoratti, M. & Szabo, I. (1995) The mitochondrial permeability
transition. Biochem. Biophys. Acta 1241, 139–176.
44. Minamikawa, T., Williams, D.A., Bowser, D.N. & Nagley, P.
(1999) Mitochondrial permeability transition and swelling can
occur reversibly without inducing cell death in intact human cells.
Exp. Cell Res. 246, 26–37.
45. Lim, M.L.R., Minamikawa, T. & Nagley, P. (2001) The proto-
nophore CCCP induces mitochondrial permeability transition
without cytochrome c release in human osteosarcoma cells. FEBS
Lett. 503, 69–74.
46. Mancini, M., Anderson, B.O., Caldwell, E., Sedghinasab, M.,
Paty,P.B.&Hockenbery,D.M.(1997)Mitochondrial
proliferation and paradoxical membrane depolarization during
terminal differentiation and apoptosis in a human colon carcino-
ma cell line. J. Cell. Biol. 138, 449–469.
3312 F. Guerrieri et al. (Eur. J. Biochem. 269) Ó FEBS 2002