| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Regular Articles |
From the Department of Anatomy and Cell Biology,*
Division of Medical Cell Biology, University of Heidelberg, Heidelberg,
Germany; the Laboratory of Clinical Chemistry,
Faculty of Pharmacy, Katholieke Universiteit Leuven, Leuven, Belgium;
the Department of Life Sciences,
Janssen
Research Foundation, Beerse, Belgium; and the Department of Molecular
Cell Biology,
Maastricht University,
Maastricht, The Netherlands
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The importance of peroxisomes for human health became evident by the identification of severe disorders of the Zellweger syndrome spectrum (Zellweger syndrome, neonatal adrenoleukodystrophy, and infantile Refsum’s disease), in which functional peroxisomes are absent.12-14 The patients with the cerebro-hepato-renal (Zellweger) syndrome, the most severe form of this disease spectrum, suffer from generalized hypotonia, exhibit hypomyelinization of neurons, and show neuronal migration defects associated with neonatal seizures. In addition, hepatic fibrosis/cirrhosis, adrenal insufficiency, and renal cysts occur in affected children who usually die during the first year of life.15
The molecular defects in the diseases of the Zellweger syndrome spectrum are deletions or mutations in PEX genes, whose gene products—the peroxins (Pex-proteins)—are involved in the biogenesis of peroxisomes.16,17 Malfunctions (or deficiencies) of some peroxins are associated with a plethora of metabolic disorders because of the absence of functional peroxisomes, such as accumulation of very long chain fatty acids and reduction of plasmalogens in plasma and various organs.14,15 Because the pathogenesis of organ malformations and dysfunctions in patients with peroxisomal biogenesis disorders is not known and experimental studies in patients are hampered by medical ethical restrictions, we have generated a knockout-mouse model for Zellweger syndrome by disrupting the PEX5 gene.1 This gene encodes Pex5p, the cytoplasmic shuttle receptor for the import of most peroxisomal matrix proteins.16,17 PEX5-/- mice have a severe peroxisomal import defect, lack functioning peroxisomes, and exhibit all major signs, pathological defects, and biochemical abnormalities of Zellweger patients.1 Therefore, this animal model provides a suitable tool for the study of the pathogenesis of organ dysfunctions in Zellweger syndrome and the investigation of potential therapeutic strategies.18,19
Interestingly, in addition to the absence of peroxisomes, we found marked alterations of mitochondrial fine structure in the liver of PEX5-/- mice.1 Indeed, mitochondrial alterations were also described in the original report by Goldfischer and colleagues12 on the peroxisomal defects in Zellweger syndrome. Since then however, only a few reports have been published concerning ultrastructural changes and alterations of the mitochondrial respiratory chain enzymes in this disease.20-24 Moreover, it is unclear whether those changes in the mitochondrial compartment are directly related to the pathogenesis of Zellweger syndrome.
Because the PEX5 knockout mouse provides an excellent model system for peroxisome biogenesis disorders and alterations of mitochondria are so prominent in the livers of those animals, a thorough analysis of the mitochondrial structure and of the function of the respiratory chain was conducted in different organs of PEX5-/- mice.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
All components for the preparation of Epon 812 were purchased from Fluka (Neu-Ulm, Germany). Paraffin (Paraplast plus) was from Sherwood Medical (St. Louis, MO). LR White, Unicryl, and osmium tetroxide were obtained from Polysciences (Heidelberg, Germany). Glutaraldehyde (25% stock solution), paraformaldehyde, PIPES, 3,3'-diaminobenzidine, cytochrome c, and di-sodium-cytidine-monophosphate were from Sigma (München, Germany). All other chemicals used for morphological experiments were obtained from Merck (Darmstadt, Germany).
Materials for Biochemical Investigations
Lactate dehydrogenase from hog muscle (complex V measurement), pyruvate kinase from rabbit muscle, 3-phosphoglycerate kinase from yeast, glyceraldehyde-3-phosphate dehydrogenase from rabbit skeletal muscle, ß-hydroxybutyrate dehydrogenase grade II from Rhodobacter spheroides, lactate dehydrogenase from rabbit muscle (lactate and pyruvate measurements), glutamate-pyruvate transaminase from pig heart, NAD, and fat-free bovine serum albumin were purchased from Roche (Mannheim, Germany). Coenzyme Q1, decylubiquinone, cytochrome c from horse heart, NADH, n-dodecyl-ß-D-maltoside, phosphoenolpyruvate, rotenone, antimycin A, oligomycin, glycerate-3-phosphate, lactate, pyruvate, and acetoacetate were from Sigma (Bornem, Belgium), ATP from Acros (Geel, Belgium), ß-hydroxybutyrate from Serva (Heidelberg, Germany), PD-10 columns, Hybond-N+ nylon membranes and Protran nitrocellulose transfer membranes from Amersham Pharmacia Biotech Benelux (Roosendaal, The Netherlands), Dulbecco’s modified Eagle’s medium, fetal calf serum, L-glutamine, gentamicin, and Trizol reagent from Life Technologies (Merelbeke, Belgium), primers for polymerase chain reaction were obtained from Eurogentec (Seraing, Belgium).
Primary Antibodies
The monospecificity of the polyclonal rabbit anti-rat liver
catalase antibody was characterized previously.25
Mouse
monoclonal antibodies against the following subunits of the
mitochondrial respiratory chain complexes were purchased from Molecular
Probes (Leiden, The Netherlands): complex I, anti-bovine NADH-ubiquinol
oxidoreductase 39-kd subunit (clone 20C11-B11-B11); complex II,
anti-bovine succinate-ubiquinol oxidoreductase 30-kd subunit (clone
21A11-AE7); complex III, anti-bovine ubiquinol-cytochrome c
oxidoreductase core 1 subunit (clone 13G12AF12-BB11); complex IV,
anti-human cytochrome c oxidase subunit 1 (clone D6-E1-A8);
complex V, anti-bovine mitochondrial ATP synthase 
subunit (clone
7H10-BD4).
Polyclonal rabbit anti-rat manganese-superoxide dismutase (MnSOD) antibodies were purchased from Research Diagnostics Inc. (Flanders, NJ). In addition, for immunohistochemical studies of MnSOD, we used a polyclonal rabbit anti-rat MnSOD antibody, which was a generous gift of Prof. K. Kato (Aichi Human Service Center, Aichi, Japan).26
PEX5 Knockout Mice
Homozygous PEX5-/- mice were obtained by breeding of the heterozygotes and identified by genotyping of tail DNA. They showed all characteristic metabolic changes and pathological alterations as described recently.1 Pregnancy state of heterozygous mothers was calculated according to the appearance of the vaginal plug. Animals of embryonic day 18.5 (E18.5) or newborn mice (P0.5) were used for all investigations. The morning of the appearance of a vaginal plug was considered as E0.5. Pregnant mothers had access to food and water ad libitum and were exposed to a 12-hour light cycle.
Morphological Experiments
Perfusion Fixation of Animals
Newborn mice were anesthetized with ether and the tails cut for the determination of genotypes. After careful opening of the thorax, all animals of three families were perfused via the left ventricle. Before fixation by perfusion, blood cells were flushed out with physiological saline.
The fixative contained for light microscopical investigations, including immunohistochemistry and in situ hybridization, 4% depolymerized paraformaldehyde in phosphate-buffered saline (PBS), pH 7.4 (fixative a); or for electron microscopy, a mixture of 4% depolymerized paraformaldehyde and 0.05% glutaraldehyde with 2% sucrose in PBS, pH 7.4 (fixative b).
For perfusion of fetal (E18.5) mice, the fetuses of three families were delivered by cesarean section under ether anesthesia. To prevent anoxic changes in organs of fetuses, each fetus was perfusion-fixed before the removal of the next fetus from the uterus. After the perfusion, all animals were kept in the same fixatives (for electron microscopy in fixative b without glutaraldehyde) for several hours before further processing.
Light Microscopy
For light microscopic investigations, fixed animals were cut longitudinally into two halves and embedded in paraffin. After cutting of 1- to 3-µm sections with a sliding microtome (Leica, Bensheim, Germany), sections of complete fetuses or newborn animals were mounted on Superfrost Plus slides (Shandon, Frankfurt, Germany), dried overnight at 37°C, and processed either for immunohistochemistry or in situ hybridization.
Immunohistochemistry
Tissue pretreatment conditions and the immunohistochemical procedure for the localization of catalase and MnSOD proteins with polyclonal rabbit antibodies were modified from the method of Grabenbauer and colleagues,27 developed in our laboratory for the localization of peroxisomal proteins and corresponding mRNAs on parallel sections of mouse fetuses. Briefly, before antibody incubation, antigens in tissue sections were either demasked by digestion with trypsin (5 minutes with 0.1% or 10 minutes with 0.01% trypsin) at 37°C in TNB buffer (0.1 mol/L Tris-HCl, pH 7.4, 0.15 mol/L NaCl; 0.5% blocking reagent; Roche Molecular Biochemicals, Mannheim, Germany) or by irradiation for 3 x 5 minutes in a microwave oven (850 W, in citrate buffer, pH 6). After blocking of endogenous peroxidase with 3% H2O2 and of endogenous biotin with an avidin/biotin blocking kit (NEN Life Science, Boston, MA), incubation with the primary antibody was performed overnight at 4°C in a moist chamber. Antigen-binding sites were detected with a peroxidase-coupled biotin/avidin system (rabbit extravidin kit; Sigma, München, Germany) and visualized by histochemical staining for peroxidase using Novared as substrate (Vector Laboratories).
A special mouse-on-mouse detection system (MOM kit, Vector Laboratories, Burlingame, CA) was used for the localization of respiratory chain enzymes with the monoclonal antibodies from mouse, because detection with the regular protocol and the mouse extravidin kit (Sigma, München, Germany) only resulted in nonspecific background staining. The pretreatment conditions for the localization of complexes III to V were the same as mentioned above. However, for the localization of complexes I and II, strong antigen retrieval by microwave irradiation (see above) and additional amplification with a tyramide signal amplification kit was necessary (TSA kit with biotinylated tyramines; NEN, Life Science, Boston, MA).28
In Situ Hybridization
The cDNA for mouse mitochondrial MnSOD was amplified by reverse transcriptase-polymerase chain reaction on mouse liver RNA (for sequences of primers see below in Northern Blot section) and the generated 595-bp fragment cloned into pBluescript SK for in vitro transcription of riboprobes. The 1.6-kbp rat catalase cDNA was a generous gift of Prof. Takashi Hashimoto (Department of Pathology Northwestern University, Chicago, IL) and was subcloned in pGEM-7zf (-).29
After preparation of digoxigenin-labeled cRNA probes by in vitro transcription (Roche Molecular Biochemicals, Mannheim, Germany), nonradioactive in situ hybridization for the localization of catalase or MnSOD mRNA was performed according to the protocols developed by Schad and colleagues29 and recently modified for fetal mouse tissue by Grabenbauer and colleagues.27 Very stringent hybridization (68°C) and washing (72°C) conditions were used to prevent nonspecific hybridization. Hybridization with digoxigenin-labeled sense probes was performed in parallel.
Electron Microscopy
For routine electron microscopy, microslicer sections of the liver (100 µm) and very thin razor blade sections of various other organs were prepared and were fixed for 15 minutes with 1% glutaraldehyde in 0.1 mol/L PIPES buffer, pH 7.4, washed briefly in 0.1 mol/L PIPES buffer, postfixed with reduced osmium,30 and embedded in Epon 812.
Enzyme Cytochemistry
The catalase activity was localized with the alkaline 3,3'-diaminobenzidine (DAB) method31 as modified subsequently by Angermüller and Fahimi.32 Cytochrome c oxidase (mitochondrial complex IV) was localized by the DAB method33 as modified by Angermüller and Fahimi.32 For the localization of acid phosphatase activity the cerium technique34 was used as described.35 After the incubation, tissue sections were postfixed either with reduced osmium (catalase and acid phosphatase) or aqueous osmium (cytochrome c oxidase) and embedded in Epon 812.
Immunocytochemistry
After perfusion fixation, 100-µm microslices were directly dehydrated and embedded in LR White (soft grade)36 or in Unicryl at low temperature.37,38 Postembedding immunocytochemistry was performed according to the protocol of Baumgart.39 After blocking nonspecific binding sites with TNB, LR White or Unicryl sections of liver tissue were incubated on drops of rabbit anti-rat MnSOD or anti-rat catalase, washed on a series of TBS drops and the antigen-antibody complexes visualized by incubation with protein A gold (15 nm).39,40 In the case of the proteins of the mitochondrial respiratory chain enzymes, which were localized with the monoclonal mouse antibodies, the antigen-antibody complexes were detected with a secondary gold-labeled goat anti-mouse antibody (Polygold 15 nm; Polysciences, Heidelberg, Germany). All sections were contrasted with uranyl acetate and lead citrate before examination in a Philips 301 electron microscope.
Biochemical Experiments
Tissue Collection
Newborn and embryonic mice were decapitated and the livers and hearts were removed within seconds, snap-frozen in liquid nitrogen and stored at -80°C before use.
Measurement of Mitochondrial Respiratory Chain Complexes and Citrate Synthase Enzyme Activities in Liver and Heart
Individual livers (40 to 80 mg) and pooled hearts (30 to 40 mg) of PEX5-/- and control mice were homogenized in, respectively, 9 or 19 volumes of homogenization buffer (210 mmol/L mannitol, 70 mmol/L sucrose, 5 mmol/L HEPES, 1 mmol/L EGTA, pH 7.2) using a borosilicate glass hand homogenizer in ice. The homogenate was centrifuged at 600 x g, 4°C for 10 minutes to remove cell debris and nuclei. The supernatant with mitochondria was snap-frozen and thawed five times before measuring complex II, -III, -IV, and -V activities. For optimal measurements of complex I and citrate synthase activities, an additional sonication step (4 cycles of 10 seconds at amplitude 4, separated by a 20 second break) with a Soniprep 150 sonicator (MSE Scientific Instruments, Sussex, England) was required. The enzymatic reactions were performed at 30°C as described by Rahman and colleagues.41 The change in absorbance was followed in a Uvikon 810 spectrophotometer (Kontron Instruments, Zurich, Switzerland). Complex III and IV activities are expressed as rate constants (min-1), whereas all other activities are expressed in units (µmol/min). All results were normalized to the activity of citrate synthase to compensate for variation in mitochondrial recovery.
Enzyme Measurements in Fibroblasts
Immortalized as well as primary fibroblasts of PEX5-/- and control mice were grown in Dulbecco’s modified Eagle’s medium with addition of 10% fetal calf serum, 2 mmol/L L-glutamine and 3 mg/L of gentamicin. The cells of eight (immortalized fibroblasts) or five (primary fibroblasts) in 10-cm culture dishes were harvested after trypsinization and washed three times with 10 ml of PBS, washed again with 1 ml homogenization buffer, and finally resuspended in 250 µl of the latter buffer. Before enzyme measurements, the cell suspension was frozen and thawed five times. Complex I and citrate synthase were measured after sonication as described above.
ATP Measurements
Pooled liver tissue (150 to 200 mg) was homogenized in 4 volumes
of 0.9 mol/L perchloric acid. Homogenization was performed as follows:
an appropriate volume of perchloric acid was frozen in a mortar
precooled in liquid N2. The frozen tissue was
added and ground together with the perchloric acid to a powder, which
was transferred to a glass homogenizer for thawing and further
homogenization at 0°C. After centrifugation (3000 x
g at 4°C for 5 minutes) the pH of the supernatant was
adjusted to 
6 with
K2CO3. Subsequently, a
second centrifugation was performed (3000 x g at 4°C
for 5 minutes) to remove the KClO4. ATP was
determined in the supernatant by using the 3-phosphoglycerate
kinase/glyceraldehyde-3-phosphate dehydrogenase method described by
Jaworek and Welsch.42
Lactate, Pyruvate, ß-Hydroxybutyrate, and Acetoacetate Measurements
The quantification of these components was based on methods previously described43-45 but slightly adjusted for fluorometric measurements except for lactate. The perchloric acid extracts prepared for ATP determination (see above) were used for the assays. All reactions were performed at room temperature and fluorescence was measured with a LS50 Luminescence Spectrometer (Perkin Elmer, Norwalk, CT) or absorbance with a Uvikon 810 spectrophotometer (lactate).
Northern Blot
Total RNA from mouse liver was extracted using the Trizol reagent. Thirty µg of RNA were loaded on a 1.2% (w/v) agarose gel containing 2.9% (w/v) formaldehyde. Gels were run in 200 mmol/L MOPS, pH 7.0, 50 mmol/L NaOAc, 10 mmol/L EDTA for 5 hours and RNA was transferred to a Hybond-N+ nylon membrane by capillary blotting. The 595-bp radioactive probe for detection of MnSOD mRNA was synthesized by performing a polymerase chain reaction on adult mouse liver cDNA using the primers 5'-TAG-TAG-GAA-TTC-GCA-CCA-CAG-CAA-GCA-CCA-TGC-GG-3' and 5'-GAT-GAT-GGA-TCC-ATA-AAC-CAG-CCC-GGA-GCC-TGG-CC-3'. After stripping, the same blot was hybridized with a ß-actin probe to check for equal loading of RNA.
Western Blot
Individual livers of newborn mice were homogenized in 4 volumes of 0.25 mol/L sucrose, 5 mmol/L MOPS, pH 7.2, 1 mmol/L ethylenediaminetetraacetic acid, 0.1% ethanol. Thirty or 15 µg of protein were separated on a 10% polyacrylamide gel and transferred to a Protran-nitrocellulose transfer membrane. Subunits of complexes I, II, III, IV, and V were detected using monoclonal mouse antibodies and MnSOD was visualized by a polyclonal rabbit antibody. Immunoreactive bands were stained with NBT/BCIP
Statistical Analysis
The unpaired t-test was used to compare the results between knockout and control mice.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
As previously described,1
regular peroxisomes were
easily identified by histochemical staining for catalase with
3,3'diaminobenzidine in hepatocytes of control mice at gestational day
18.5 (E18.5) (Figure 1A)
, but were absent
in the PEX5 knockouts of the same age (Figure 1B)
. In 5%
of hepatocytes of PEX5-deficient mice peroxisomal
membrane-remnant ghosts consisting of double-membraned loops measuring
100 nm across were observed that were labeled with an antibody to
70-kd peroxisomal membrane protein.1
Similar structures
were also observed in the adrenal glands but not in the kidney tubules
(data not shown).
|
. Such
mitochondrial aggregates were randomly distributed in the liver and
were often found under the sinusoidal and the basolateral surface of
hepatocytes (Figure 1B)
. By electron microscopy different types of
alterations of mitochondria were present, which were more pronounced in
newborn (Figure 1; G to J
) as opposed to E18.5
PEX5-/- mice (Figure 1; C to F
). These
alterations involved all subcompartments of the mitochondria (outer
membrane, inner membrane, and the matrix space). Alterations of the
outer membrane consisted of 1) membrane whorls on the surface of
mitochondria (Figure 1E)
, 2) small blebs with electronlucent contents
(Figure 1G)
, or 3) protrusions of the inner membrane through focal
interruptions of the outer membrane (Figure 1F)
. In mitochondrial
cristae, more obvious morphological changes were observed that
consisted of 1) rarefication of cristae (Figure 1, B and J)
, 2)
arrangements into parallel stacks (Figure 1, B and F)
, 3) curvilinear
and circular forms (Figure 1; C and H to J
), 4) onion-like concentric
rings (Figure 1, D and G)
, 5) dilation with the formation of tubular
cristae similar to those seen in steroid hormone-producing cells
(Figure 1E)
. In hepatocytes of newborn PEX5 knockout mice,
we also observed ghost-like mitochondrial rest structures (Figure 1, H and I
; small arrows), swollen mitochondria with strong rarefication of
cristae, and reduced electron density of their matrix (Figure 1, I and J
; asterisks). Mitochondrial alterations were quite heterogeneous in
different hepatocytes and even within the same cell severely altered
mitochondria were observed adjacent to normal ones (Figure 1; B, C, and
G to J
). The mitochondria in sinusoidal cells (endothelial cells and
Kupffer cells), in contrast to the ones in hepatocytes, appeared normal
as did the mitochondria in fibroblasts of the sparse connective tissue
of the fetal or newborn liver. Mitochondrial Alterations Are also Present in Other Tissues and in Blood Cells
In addition to changes of the mitochondrial compartment in
hepatocytes, structural alterations of the mitochondrial population of
several lipid-metabolizing organs and tissues, which are also
pathologically altered in Zellweger syndrome, were observed. In the
epithelial cells of the adrenal cortex of
PEX5-/- mice, especially in the inner
zona fasciculata, severely altered mitochondria were found (Figure 2, A and B)
, showing either enlarged
lamellated mitochondrial granules (Figure 2A)
or accumulation of
electron-dense material in the intermembrane space of tubular cristae
(Figure 2B)
. Focal proliferation of pleomorphic mitochondria with
different types of lamellated granules were also seen in epithelial
cells of proximal tubules of the kidney, known to be rich in
peroxisomes (Figure 2, C and D)
. In contrast, the mitochondrial
compartment in other parts of the nephron was not altered. In
cardiomyocytes of PEX5-/- mice, routine
electron microscopy revealed an increase in size and number of
mitochondrial granules and the presence of megamitochondria (Figure 2F)
. Pathological alterations in skeletal muscle were dependent on the
region of the body and the specific muscles investigated. Most severe
alterations were observed in the diaphragm, consisting of blebs on the
surface of mitochondria (not shown) as well as rarefication and
curvilinear alterations of mitochondrial cristae. In some myocytes of
leg muscles subsarcolemmal aggregates of mitochondria and
megamitochondria with bizarre forms were present (not shown). On the
other hand, mitochondria in smooth muscle cells of the intestine
displayed only slight alterations, which consisted either of the
enlargement of the intermembrane space of the cristae (Figure 2G)
and
blebs on the surface, or invaginations into their matrix (Figure 2H)
.
In light of the hematopoietic capacity of the liver in newborn mice,
mitochondrial alterations were also evaluated in blood cells present in
this organ. Whereas mitochondria were completely normal in precursors
of red blood cells, such as normoblasts (not shown), they were abnormal
in a small portion of developing neutrophils and lymphocytes (Figure 2I)
. The specific alterations of blood cells were independent of the
severity of mitochondrial changes in neighboring hepatocytes and were
rather cell type-specific. Whereas in neutrophilic granulocytes,
mitochondria with circular cristae were frequently observed (Figure 2I)
, the alterations in lymphocytes were more discrete and resembled
those of smooth muscle cells showing enlargement of the intercristae
space with formation of tubular cristae (not shown).
|
Damaged Mitochondria Are Removed from the Cytoplasm by Increased Autophagocytosis
In addition to ultrastructural changes of the mitochondria, we
observed other signs suggestive of increased cellular toxicity and
damage to mitochondria. Thus, many mitochondria were found in
autophagic vacuoles of adrenal cortical cells, cardiomyocytes, and
hepatocytes of PEX5-/- mice (Figure 2, J to L)
. Indeed, with cytochemical staining for acid phosphatase
activity, large autophagic vacuoles were seen containing many
mitochondria (Figure 2M)
.
Reduction of Complex I Activity Is the Most Consistent Biochemical Finding in Mitochondria of Liver and Heart
Because the five multimeric complexes (complexes I to V) of the
electron transport chain reside in the mitochondrial inner membrane and
severe ultrastructural changes of this compartment were found in
PEX5-/- mice, we investigated the
activities and the protein contents of the respiratory chain enzymes.
Liver, displaying the strongest morphological alterations of
mitochondria, and heart, with more subtle changes, were analyzed. In
Table 1
the activities of mitochondrial
respiratory chain enzymes in liver and heart of newborn mice, expressed
relative to the activity of citrate synthase are summarized. In
comparison to control levels, the activity in liver of newborn
PEX5-/- mice was reduced to 45% for
complex I and to 65% for complex V, whereas complex IV activity was
increased up to 180%. In contrast, no change was observed in the
activities of complexes II and III. These differences in respiratory
chain complex activities were already present at day 18.5 of embryonic
development, except for complex V, which was only reduced by 25%, a
statistically insignificant change (not shown).
|
, which showed only slight ultrastructural
alterations, the differences in activities of the mitochondrial
respiratory chain enzymes between control and
PEX5-/- mice were also less severe than
in liver. Complex I activity of newborn
PEX5-/- mice was reduced to 65% of the
control value, but there was no significant difference in E18.5 animals
(data not shown). On the other hand, complex IV activity was 20%
higher in heart of newborn PEX5-/- mice,
but this difference was not statistically significant. Corresponding to the normal appearance of mitochondria in fibroblasts in primary cultures of PEX5 knockout mice (not shown) or in situ in various organs, complex I activity was not altered in fibroblast cultures of these mice (either immortalized fibroblasts or primary cultures of passage 3 or 4) (data not shown).
Western blot analysis of different subunits of respiratory chain
complexes from liver mitochondria (Figure 3)
revealed an estimated 50% reduction
of the 39-kd subunit of complex I, a less-pronounced reduction of the
core I subunit of complex III and subunit I of complex IV. In contrast
the analyzed subunits of complexes II and V were not significantly
altered (Figure 3)
.
|
In addition to electron microscopy, we used an
extravidin/peroxidase kit to detect mitochondria by light microscopy,
based on their high content of biotinylated proteins (Figure 4; A to D
).46
The staining
pattern corresponded to the intracellular distribution of mitochondria
in electron microscopical preparations. In contrast to the granular
particulate appearance of mitochondria, some of which were localized
around glycogen deposits, in hepatocytes of control mice (Figure 4, A and C)
, aggregates of proliferated mitochondria in subplasmalemmal
regions were observed in E18.5 PEX5 knockout mice (Figure 4B)
. The abundance and proliferation of the mitochondrial aggregates
that filled up some hepatocytes was even more pronounced in newborn
PEX5-/- mice (Figure 4D)
, which in
addition exhibited reduced glycogen deposits.
|
To provide further insight into mitochondrial alterations, the
localization of different mitochondrial respiratory chain complexes was
investigated by immunohistochemistry on paraffin sections (Figure 4; E
to L
). Optimal light microscopical localization of proteins of
respiratory chain enzymes in hepatocytes was observed only after
antigen retrieval by 3 x 5-minute microwave irradiation of
paraffin sections. The localization of complexes III to V was possible
thereafter using our standard procedure.27
However,
localization of complexes I and II was only achieved by means of
additional catalyzed reporter deposition with biotinylated tyramine
(TSA, tyramide signal amplification).28
By applying these
methods, a marked intercellular heterogeneity in the distribution of
the mitochondrial respiratory chain enzymes was observed in hepatocytes
of E18.5 and P0.5 PEX5 knockout mice (shown only
for E18.5 fetuses in Figure 4; F, and H to L
). The strongest reduction
of overall staining intensity was seen in preparations for complex I
(Figure 4, E
versus F), but a reduction, albeit to a lesser
extent, was also present in preparations for other respiratory chain
complexes. Despite aggregation of proliferated mitochondria in most
hepatocytes, the stain intensity was patchy showing heterogeneity in
neighboring cells of PEX5-/- mice (Figure 4; H to L
).
The intracellular heterogeneity of mitochondrial respiratory chain
enzymes was further substantiated by electron microscopical
localization of cytochrome c oxidase enzyme activity in the
liver of PEX5-/- mice (both in E18.5 and
P0.5; Figure 5, B and C
). In contrast,
mitochondria in hepatocytes of control animals exhibited a uniform
distribution of the cytochrome c oxidase activity (Figure 5A)
. Interestingly, the majority of ghost-like or severely altered
mitochondria were either negative or exhibited low-cytochrome
c oxidase activity (Figure 5B)
. On the other hand, the
mitochondria with parallel-stacked cristae in hepatocytes of
PEX5-/- mice showed very strong staining
of their longitudinal cristae, indicating increased cytochrome
c oxidase activity (Figure 5C)
. Interestingly, despite the
relatively subtle changes of mitochondrial morphology in
cardiomyocytes of PEX5 knockout mice, strong heterogeneity
of cytochrome c oxidase staining with negative organelles
next to heavily stained ones was also found in the heart (Figure 5F)
.
On the other hand, a uniform cytochrome c oxidase-staining
pattern was present in the diaphragm, despite the more severe
alterations of the mitochondrial ultrastructure (Figure 5, G and H)
.
|
ATP Levels and Redox State in Hepatocytes of PEX5 Knockout Mice
The question whether changes in the overall activities of complex
I and complex V would lead to reduced ATP levels in hepatocytes of
PEX5 knockout mice, was investigated by measuring the ATP
content in the liver (Table 2)
. The mean
value obtained for PEX5-/- mice
(2.69 ± 0.02 µmol/g liver) was slightly higher than the control
value (2.14 ± 0.18 µmol/g liver), indicating that there is no
deficiency in the steady-state level of hepatic ATP in
peroxisome-deficient mice.
|
). They were all reduced in PEX5
knockout livers in comparison to controls (percentage of control
levels: lactate, 25%; pyruvate, 71%; ß-hydroxybutyrate, 8%; and
acetoacetate, 29%). These low values are probably because of
starvation of the peroxisome-deficient mice as a consequence of the
severe hypotonia and may therefore invalidate the ratios of
lactate/pyruvate and ß-hydroxybutyrate/acetoacetate as indicators for
respiratory chain function. Contrary to what one would expect in case
of a deficiently functioning respiratory chain, both ratios were
strongly reduced in PEX5-/- mice (34 and
27% of control values). Induction of MnSOD in Different Organs of PEX5 Knockout Mice Suggests an Increase of ROS
Because some of the features of mitochondria in
PEX5-/- hepatocytes were reminiscent of
conditions associated with oxidative stress and because complex I
deficiency is known to be accompanied by the production of
ROS,47
we looked for further indications of oxidative
stress in the liver of peroxisome-deficient mice. Because in several
experimental conditions associated with increased oxidative stress
mitochondria express increased amounts of MnSOD protein,47
in situ hybridization and immunocytochemical
procedures were optimized to localize the mRNA and protein of this
enzyme. A strong and reproducible immunohistochemical signal for MnSOD
was obtained with a polyclonal anti-MnSOD antibody received through the
courtesy of Professor Kato.26
When compared to
wild-type littermates, PEX5-/- mice at
E18.5 and older, showed strong increases of MnSOD mRNA and protein
expression in their livers (Figure 6; A
to F
). In addition to the alterations in liver, heterogeneous induction
of the MnSOD mRNA and protein was also found in cardiomyocytes (Figure 6; I to L
).
|
.
Analysis of 200 mitochondria by morphometry revealed that the
immunolabeling for MnSOD was increased in
PEX5-/- mice. Thus, in contrast to 45%
labeling in control mice, 59% of mitochondria were labeled in
PEX5 knockouts. Interestingly, in hepatocytes of
PEX5-/- mice, mostly the
normal-appearing, or ring-shaped mitochondria were labeled with the
MnSOD antibody, whereas severely altered mitochondria were unlabeled.
Despite clear evidence by in situ hybridization and
immunocytochemistry (Figure 6)
the analysis of MnSOD mRNA
and protein by Northern and Western blotting revealed no significant
alterations in liver tissue of newborn peroxisome-deficient mice as
compared to wild-type controls (data not shown). However, a twofold
increase in MnSOD transcripts was found in two
PEX5-deficient mice that exceptionally survived for 2 days
(not shown).
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Severe Mitochondrial Alterations Are Found in PEX5-/- Mice, Corroborating the Fine Structural Changes Reported in Patients with Zellweger Syndrome
In PEX5-/- mice the most prominent
mitochondrial alterations were found in hepatocytes, exhibiting
heterogeneous, focal proliferation of pleomorphic mitochondria in some
cells and involving all mitochondrial subcompartments (outer membrane,
mitochondrial cristae, and matrix space). These alterations are already
present in E18.5 PEX5-/- fetuses but
become more pronounced after birth (P0.5), occasionally leading to
severe and degenerative mitochondrial ghost forms (Figure 1; H to J
).
Alterations of the mitochondrial fine structure in Zellweger patients
were already described in the original report on the absence of
peroxisomes by Goldfischer and colleagues.12
Those
structural changes were, however, less severe in nature and
consisted of irregular and twisted cristae, dilatation of the
intracristae space, as well as occurrence of attenuated organelles,
corresponding to the subtle changes in the livers of our E18.5
PEX5-/- mice. In several subsequent
ultrastructural studies,20-22,24,48
mitochondrial
alterations of varying types and severity were reported, although in a
few cases there also were no changes noted.49,50
In addition to the liver, changes of mitochondria in Zellweger patients have been reported in skeletal muscle.51,52 Consistent with the finding of Sarnat and colleagues,51 the diaphragm of the PEX5-/- mice was more affected than muscles of the limbs, which is interesting, because respiratory failure could be partly responsible for the death of some patients and possibly also of our mice. Furthermore, PEX5-/- mice showed abnormalities of mitochondria with different severity in other tissues known to be altered in Zellweger syndrome, such as epithelial cells of the inner zona fasciculata of the adrenal cortex and the renal proximal tubules, smooth muscle cells of the duodenum, cardiomyocytes, and some blood cells (developing lymphocytes and neutrophils). Although Zellweger patient fibroblasts have been extensively studied, mitochondrial alterations in those cells have not been observed, which is consistent with our observations in PEX5-/- fibroblasts. Finally, despite the depositions of different lipid derivatives in brain,1 no obviously altered mitochondria were found in neurons of the cortical plate of PEX5 knockout mice. Mitochondrial changes were also not observed in neurons of Zellweger patients53-55 but were seen in astrocytes, ranging from severely altered, degenerative forms12 to mitochondrial proliferation.50 Because pathological malformations of the brain occur focally, additional studies of the brain in PEX5 knockout mice are necessary before definitive conclusions on mitochondrial alterations in this organ can be drawn.
Ultrastructural Alterations of Mitochondria Are Associated with Changes in Respiratory Chain Complexes
As shown by a combination of activity measurements and Western
blot analysis of liver homogenates, immunocytochemical localization of
the enzyme proteins, as well as enzyme cytochemical staining for
cytochrome c oxidase, there are profound alterations in the
activity and content of different respiratory chain enzymes in
mitochondria of PEX5-/- mice. The most
consistent finding was a twofold reduction of NADH-ubiquinone
oxidoreductase (complex I) activity that correlated with the reduced
levels of a complex I subunit in liver samples of newborn
PEX5-/- mice (Table 1)
. These results are
in line with the studies of Goldfischer and
colleagues,12
who noted a 70% decrease in
O2 consumption in mitochondrial fractions of
liver of Zellweger patients when the fractions were incubated with
complex I-linked substrates. In contrast to Zellweger patients who also
displayed a reduced activity of complex II, this activity was in the
normal range in liver or heart homogenates of
PEX5-/- mice. For complex III, there was
a marked heterogeneity in immunocytochemical staining despite unchanged
levels of total enzyme activity in homogenates. Importantly, severely
altered mitochondria seemed to contain much less complex III than less
affected mitochondria. Enzyme histochemical staining for cytochrome
c oxidase (complex IV) also revealed a heterogeneous
staining of the mitochondrial cristae. In individual hepatocytes,
ultrastructurally normal mitochondria and mitochondria with parallel
stacks of cristae exhibited a strong cytochrome c oxidase
activity, whereas mitochondria with altered cristae showed a
significantly weaker staining with mitochondrial ghosts exhibiting only
minimal staining. The elevation of complex IV activity observed in
liver homogenates is not in contradiction with the reduction observed
for other respiratory chain enzymes, because this protein is
differently regulated.56
Indeed, in some patients with a
reduced activity of complex I, an elevation of the activity of complex
IV has been reported.57
On the other hand, a significant 35% reduction of complex V was found in liver homogenates of newborn PEX5 knockout mice, and this was also reflected in reduced and heterogeneous staining of complex V in light microscopic preparations. No evidence of uncoupling of the mitochondria in liver and heart of PEX5-/- mice was obtained by enzyme histochemical staining for ATPase activity (M. Borgers, data not shown). This is in agreement with an earlier report on liver mitochondria of Zellweger patients,48 although in skeletal muscle a slight uncoupling of mitochondria has been described.52
Taken together, the ultrastructural alterations of liver mitochondria are accompanied by an altered activity and distribution of the mitochondrial respiratory chain complexes leading to a heterogeneous mitochondrial population with an overall decrease of complex I and complex V activities in liver of newborn PEX5-deficient mice.
Deficiency of the Respiratory Chain Enzymes Does Not Affect ATP Levels
Because the respiratory chain is responsible for the production of
ATP by oxidative phosphorylation and the regeneration of
NAD+, functional defects of respiratory chain
enzymes (mostly complexes I and IV) are often accompanied by low ATP
levels and increased ratios of lactate/pyruvate and
ß-hydroxybutyrate/acetoacetate.58
Unfortunately, because
of starvation of the newborn PEX5-/-
mice, all redox markers were strongly depleted, interfering with the
proper assessment of the redox status. Nevertheless, steady-state ATP
levels were not altered in PEX5-/- mice
as compared to control littermates (Table 2)
, suggesting that despite
the reduction of the activities of complexes I and V, the overall
mitochondrial respiratory function of the whole hepatocyte population
is still sufficient. These results are consistent with the so-called
"threshold effect"59
in patients with complex I
deficiency, who may have normal ATP production despite severe reduction
of complex I activity.47,57
Indications Suggestive of Increased Oxidative Stress in PEX5-/- Knockout Mice
Mitochondrial electron transport has long been recognized as a major source of superoxide (.O2-) and H2O2.60 The superoxide anion radical is generated by direct transfer of electrons from complex I or complex III to oxygen. Under physiological conditions, 2 to 3% of the consumed oxygen is converted to superoxide. To protect the cell from toxic injury, .O2- is converted to the less toxic H2O2 by MnSOD present in the mitochondrial matrix. This H2O2 may either diffuse out of mitochondria into the cytoplasm or may remain in the mitochondrial matrix and is reduced to H2O by glutathione peroxidases of the appropriate cellular compartment. If the balance between the generation and scavenging of these ROS is disturbed, increased oxidative stress may develop in mitochondria,58,61 leading to oxidation of lipids in membranes, of thiol-containing proteins, and of mitochondrial DNA.
The ultrastructural alterations in the mitochondrial populations of different organs of PEX5-/- mice are suggestive of oxidative stress. These alterations, which have also been reported in mitochondrial respiratory chain disorders, include: 1) proliferation of pleomorphic mitochondria, 2) abnormal cristae (eg, parallel stacks, curvilinear forms), 3) fluffy matrix, and 4) swollen mitochondria.62 Similar ultrastructural alterations have also been described in the liver in many other disease states associated with oxidative stress (eg, Reye’s syndrome, acute fatty liver of pregnancy) and intoxication with drugs (eg, peroxisome proliferators, chloramphenicol, nonsteroidal anti-inflammatory drugs) or toxins (ethanol, CCl4, antimycin A, and rotenone).58,63,64 In Wilson’s (copper intoxication), Parkinson’s, or Huntington’s diseases and Friedreich’s ataxia similar mitochondrial alterations have also been described.58,65 The early stage alterations observed in Wilson’s disease are similar to the mitochondrial changes in E18.5 PEX5-/- mice (except for crystalline inclusions). Interestingly, analogous mitochondrial alterations have been noted in iron overload, which is associated with substantial depletion of endogenous antioxidants and moderate lipid peroxidative damage.66 Similar pathological alterations could play a role in the pathogenesis of the secondary mitochondrial changes in Zellweger syndrome, because iron overload and deposition have also been reported in several organs of Zellweger patients.48,52,67
Ultrastructural alterations of mitochondria with circular cristae and dilation of intracristal spaces similar to those reported here were recently described in syncytiotrophoblastic cells of human placenta and it was suggested that they represented an adaptive response to oxidative stress because of elevation of oxygen tension at the end of the first trimester.68 The further intensification of degenerative mitochondrial alterations in newborn (P0.5) PEX5-/- mice in comparison to 18.5-day embryos would also support the notion that oxidative stress, because of elevation of oxygen tension after the birth, is responsible for the worsening of the mitochondrial changes in PEX5 knockouts.
Another sign of oxidative stress in liver mitochondria of
PEX5-/- mice was the significant
up-regulation of MnSOD (Figure 6)
, which is an important enzyme for the
protection against the development of oxidative
stress.69,70
An increase in MnSOD was also reported in
some patients with complex I deficiency.47
As mentioned above, the sites of radical production in the respiratory chain are complexes I and III. A clear correlation between the degree of complex I inhibition, radical production, and cell injury by lipid peroxidation was shown in different experimental models.71 Because in PEX5-/- mice, complex I activity was twofold reduced in liver homogenates and complexes I and III showed a lower immunostaining intensity in liver sections, it is plausible that increased mitochondrial radical production occurs in the peroxisome-deficient mouse livers. Interestingly, it has been demonstrated that radicals produced in mitochondria are themselves effective inhibitors of different respiratory chain enzymes and that the pattern of inactivation is different for distinct oxygen radicals.72 Thus, a vicious cycle could be initiated whereby the reduced activities of complex I and possibly complex III would release ROS that in turn could further inhibit the activities of the respiratory chain enzymes.
Origin and Consequences of Mitochondrial Changes
The fact that the absence of functioning peroxisomes is associated
with an altered mitochondrial structure and function in man as well as
in mice indicates that a conserved mechanism underlies these
subcellular changes. In some Zellweger patients no mitochondrial
alterations were observed, but this might be related to the genetic
heterogeneity of the syndrome that can indeed be caused by defects in
at least 10 different genes.17
In
PEX2-deficient mice, another model of Zellweger
syndrome,73
even though no clear-cut mitochondrial
alterations in hepatocytes were reported, a few morphologically
abnormal mitochondria can be identified in the published electron
micrographs (eg, Figure 6B
in Faust and Hatten73
).
In speculating on a link between the absence of peroxisomes and the mitochondrial alterations two possibilities must be taken into consideration which are not necessarily mutually exclusive: 1) defective rescue function of peroxisomes against ROS and 2) defective detoxification function of peroxisomes.
A physiological function of peroxisomes in defense against oxidative
stress has been well established in plant cells.74
Moreover, the localization in mammalian liver peroxisomes of various
enzymes involved in generation and detoxification of ROS such as
xanthine oxidase,75
superoxide dismutases,76, 77
and glutathione peroxidase78
suggest strongly
that they must be involved in regulation of cellular redox
processes.3
Indeed, Reddy and Lalwani79
were
the first to draw attention to the generation of ROS in hepatic
peroxisomes of rodents treated with peroxisome proliferators and its
possible implications for hepatocarcinogenesis. The same group
described recently the importance of peroxisome proliferation in
regulation of H2O2 levels
in the liver of acyl-CoA oxidase I knockout mice.80
Moreover, the complex interaction between the peroxisomal and
mitochondrial lipid metabolism in the pathogenesis of fatty liver was
shown in mice nullizygous for both peroxisomal acyl-CoA oxidase and
PPAR-.81
In oxidative stress induced by ischemia and
reperfusion of kidney, degenerative changes of regular peroxisomes in
proximal tubules was noted by Singh3
and Gulati and
colleagues82
that is followed by a compensatory
regeneration of smaller particles. Interestingly, in the same
experimental model there is evidence of severe reduction of complex I
activity in renal mitochondria,83
suggesting an
association between peroxisome degeneration and mitochondrial complex I
inactivation, similar to PEX5-/- mice.
Furthermore, fibroblasts from patients with peroxisome biogenesis
disorders have been shown to be much more sensitive than controls to
ROS generated by UV irradiation,84,85
which has been
suggested to be because of reduced plasmalogen levels, which might have
an important anti-oxidative function in biomembranes.86
Interestingly, HepG2 cells with a prominent peroxisome
compartment87
respond to UV irradiation by tubulation and
proliferation of peroxisomes.5
Finally,
H2O2 was recently shown to
induce several PEX genes that are involved in biogenesis of
peroxisomes.4
Taken together there is substantial evidence
that peroxisomes have an important function in cellular rescue against
ROS. How the cellular redox state is affected when both ROS-generating
and -detoxifying enzymes are mislocalized to the cytoplasm such as in
our PEX5 knockout mice, however requires further study.
Nevertheless, it seems highly likely that the absence of peroxisomes
leads to increased ROS levels in cells with prominent peroxisomal
metabolism that in turn could impair the mitochondrial respiratory
chain enzymes triggering a vicious cycle of mitochondrial degeneration
as mentioned above.
An alternative explanation for the damage to the mitochondrial compartment might be the toxic effects of accumulation of compounds that are normally degraded by peroxisomal ß-oxidation eg, very long-chain fatty acids, dicarboxylic acids, bile acid intermediates, and branched chain fatty acids. Notably, it was recently found that bile duct ligation in rats causes impairment of the mitochondrial respiratory chain, possibly by altering the lipid composition of the inner mitochondrial membrane.88 Furthermore, it is well known that free fatty acids act as potent detergents that can damage cellular membranes.89 In this respect it is important to mention that the mitochondrial changes in the PEX5 knockout mice are most prominent in tissues with an active lipid metabolism and a high content of peroxisomes (eg, liver, proximal renal tubules, and adrenals) and that the mitochondrial damage described in this report seems to be a cell autonomous phenomenon. The latter was exemplified in liver where liver parenchymal cells displayed severe mitochondrial changes whereas neighboring endothelial cells, Kupffer cells, and connective tissue cells did not show any abnormalities.
Interestingly, Mathis and colleagues90 have noted that the mitochondrial changes in the liver of Zellweger patients were already present before the development of major hepatic disease. Indeed, in our newborn peroxisome-deficient mice, also no severe liver pathology was observed, suggesting that the mitochondrial alterations could play a role in the pathogenesis of the liver dysfunctions at later stages of the disease. In conclusion, the severely altered mitochondria in various organs of peroxisome-deficient mice that may be induced by oxidative stress could play a major role in the pathogenesis of the multiple organ dysfunctions observed in this devastating disorder.
|
Acknowledgements |
|---|
|
Footnotes |
|---|
Supported by the BioMed II Program of European Community (grant BMH4-98-3569); the Deutsche Forschungsgemeinschaft, Bonn, Germany (grant Ba 1155/1-4 and SFB 601-B1), and the Vlaamse Gemeenschap, Belgium (grant GOA/99/09).
E. B. and I. V. contributed equally to this article.
Present address of E. B.: Department of Biological Chemistry, Johns Hopkins University SOM/725 N. Wolfe/409 Physiology, Baltimore, MD 21205.
Accepted for publication July 11, 2001.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
