- Email: [email protected]
Light microscopic methods to visualize mitochondria on tissue sections
Methods 46 (2008) 274–280
Contents lists available at ScienceDirect
Methods journal homepage: www.elsevier.com/locate/ymeth
Review Article
Light microscopic methods to visualize mitochondria on tissue sections Kurenai Tanji a, Eduardo Bonilla a,b,* a b
Department of Pathology, Columbia University, New York, NY 10032, USA Department of Neurology, Columbia University, 1150 St. Nicholas Avenue, Room 3-318, New York, NY 10032, USA
a r t i c l e
i n f o
Article history: Accepted 30 September 2008 Available online 16 October 2008 Keywords: Histochemistry Immunohistochemistry Tissue sections Light microscope Mitochondrial dysfunction
a b s t r a c t Mitochondria are cytoplasmic, double-membrane organelles, a main role of which is to synthesize ATP, the universal energy ‘supply’ of cells. In the last three decades, molecular genetic, biochemical, immunological and cell biological techniques have been applied in a coordinated fashion to unveil the pathogenesis of known mitochondrial disorders, as well as to explore the role of mitochondria in aging and neurodegenerative diseases. Once to be thought to be rare, it is now clear that mitochondrial dysfunction is an important cause of neurological and cardiac diseases, and age-related disorders such as cancer. Here, we review, illustrate, and provide updated protocols of two histochemical, and three immunohistochemical methods that in our opinion are the most reliable tools to visualize mitochondria on tissue sections from normal and disease specimens. Ó 2008 Elsevier Inc. All rights reserved.
1. Introduction Mitochondria are the primary ATP-generating organelles in most mammalian cells and they contain their own DNA (mtDNA) which is maternally inherited [2,15]. ATP is produced via oxidative phosphorylation through five respiratory complexes located in the inner mitochondrial membrane. The human mitochondrial genome is a 16,569-bp double-stranded DNA. It is highly compact, and contains only 37 genes: 2 genes encode ribosomal RNAs (rRNAs), 22 encode transfer RNAs (tRNAs), and 13 encode polypeptides. All 13 polypeptides are components of the respiratory chain, including 7 subunits of complex I or NADH dehydrogenase-ubiquinone oxidoreductase, 1 subunit of complex III or ubiquinone-cytochrome c oxidoreductase, 3 subunits of complex IV—cytochrome c oxidase (COX), and 2 subunits of complex V or ATP synthase [4]. The respiratory complexes also contain nuclear DNA (nDNA) encoded sububits, which are imported into the organelle from the cytosol and assembled, together, with the mtDNA-encoded subunits, into the respective holoenzymes in the mitochondrial inner membrane ([29]; Smeitink et al. [40]; Medeiros et al. this volume). Complex II or succinate dehydrogenase-ubiquinone oxidoreductase contains only nDNA-encoded subunits. The mitochondrial disorders encompass a heterogeneous group of diseases in which mitochondrial dysfunction produces clinical manifestations. Because of the dual genetic make up of mitochondria, these diseases are typically caused by genetic errors either in mtDNA or nDNA ([54]; Calvaruso et al. this volume).
* Corresponding author. Address: Department of Neurology, Columbia University, 1150 St. Nicholas Avenue, Room 3-318, New York, NY 10032, USA. E-mail address: [email protected] (E. Bonilla). 1046-2023/$ - see front matter Ó 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.ymeth.2008.09.027
Pathogenic mtDNA mutations have been identified in three of the ‘‘prototypes” of the mitochondrial disorders. First, large scale mtDNA rearrangements (i.e. deletions [D-mtDNA] and/or duplications [dup-mtDNA]) have been associated with sporadic KearnsSayre syndrome (KSS), and are often seen in patients with isolated ocular myopathy (OM) [53,26]. Second, myoclonus epilepsy with ragged-red fibers (MERRF) has been frequently associated with two different point mutations, both in the tRNALys gene [36,39]. Third, mitochondrial encephalopathy, lactic acidosis and strokelike episodes (MELAS) has been mainly associated with two different point mutations, both in the tRNALeu(UUR) gene [16,17]. Other point mutations in tRNAs and protein-coding genes of the mitochondrial genome as well as disorders leading to generation of multiple D-mtDNAs or to depletion of mtDNA but due to primary nDNA defects have now been described [54]. Because mitochondria are inherited only from the mother [15], pedigrees harboring defects in mtDNA genes should exhibit maternal inheritance. Moreover, because there are hundreds or even thousands of mitochondria in each cell, with an average of 2–10 mtDNAs per organelle [6,33], mutations in mtDNA result in two populations of mtDNAs—mutated and wild-type—a condition known as heteroplasmy. The phenotypic expression of a mtDNA mutation is regulated by the threshold effect, that is, the mutant phenotype is expressed in the heteroplasmic cells only when the relative proportion of mutant mtDNAs exceeds a minimum value [51]. Research on mitochondrial diseases has also uncovered an increasing number of disorders that are caused by mutations in nuclear genes encoding the subunits of the respiratory chain, or other proteins that are essential for the biosynthesis of specific cofactors or assembly of the complexes. Interestingly, complex I and II defects are associated with mutations in genes encoding the subunits of the complexes, while complex III, IV, and V defects are caused by
K. Tanji, E. Bonilla / Methods 46 (2008) 274–280
275
mutations in specific assembly proteins or biosynthetic factors [38,54]. Nerve and muscle, whose function is highly dependent on oxidative metabolism, are the most severely affected tissues in the mitochondrial disorders [12]. Consequently, genetic as well as morphologic studies of muscle and brain from patients with mitochondrial diseases have been proven fundamental to understand the pathogenesis of mitochondrial dysfunction at the level of individual muscle fibers and in different neuronal nuclei [8,41,45]. The purpose of this chapter is to present the histochemical and immunohistochemical, methods that, in our experience, appear to be the most reliable for the correct identification of mitochondria on frozen or paraffin-embedded tissue sections, to illustrate their potential using specific pathological samples, and to provide an updated version of the methods. While the described protocols refer to skeletal muscle or brain mitochondria, the methods described can be applied to any cell type [42,46]. It is not our intention to cover every study or method related to morphological aspects of mitochondria, but rather to provide enough information to allow investigators to apply these selected light microscopy tools to a particular scientific or diagnostic question.
inhibitor, sodium malonate (0.01 M), is added to the incubation medium. Using this method for detecting SDH activity in normal muscle sections, two populations of fibers are seen resulting in a checkerboard pattern. Type II fibers, which rely on glycolytic metabolism, show a light blue network-like stain. Type I fibers, whose metabolism is highly oxidative and therefore contain more mitochondria, show a more elaborate and darker mitochondrial network (Fig. 1a). In samples with pathological proliferation of mitochondria (RRF), the RRF show an intense blue SDH reaction corresponding to the distribution of the mitochondria within the fiber (Fig. 1c). This proliferation of mitochondria is associated with most mtDNA defects (deletions and tRNA point mutations), but RRF can also be observed in other disorders that are thought to be due to defects of nDNA, such as the depletion of muscle mtDNA, and the fatal and benign COX-deficient myopathies of infancy [13]. SDH histochemistry is also useful for the diagnosis of complex II deficiency. Several patients with myopathy and complex II deficiency have been reported. In agreement with the biochemical observations, SDH histochemistry showed complete lack of reaction in muscle [18,47].
1.1. Histochemistry
1.1.2. COX COX, the last component of the respiratory chain, catalyzes the transfer of reducing equivalents from cytochrome c to molecular oxygen. The holoenzyme contains two heme a moieties (a and a3) and three copper atoms (two in the CuA site and one in the and CuB site) bound to a multisubunit protein frame embedded in the mitochondrial inner membrane [11,44]. In mammals, the apoenzyme is composed of thirteen different subunits. The three
The visualization of normal and pathological mitochondria on frozen tissue sections can be carried out using a number of cytochemical techniques. These include the modified Gomori trichrome and hematoxylin-eosin stains, and histochemical methods for the demonstration of oxidative enzyme activity. The most informative histochemical alteration of mitochondria in skeletal muscle is the ragged-red fiber (RRF), observed on frozen sections stained with the trichrome method of Engel and Cunningham [14]. The name derives from the reddish appearance of the trichrome-stained muscle fiber as a result of subsarcolemmal and/or intermyofibrillar proliferation of the mitochondria. The fibers harboring abnormal deposits of mitochondria are most often type I myofibers, and they may also contain increased numbers of lipid droplets. Since accumulations of materials other than mitochondria may simulate RRF formation, the identification of deposits suspected of being mitochondrial proliferation should be confirmed histochemically by the application of oxidative enzyme stains. In our experience, enzyme histochemistry for the activity of succinate dehydrogenase (SDH) and COX has proven to be the most reliable methods for the correct visualization of normal mitochondria, and for the interpretation and ultimately the diagnosis of some of the mitochondrial disorders affecting skeletal muscle [13]. 1.1.1. SDH Succinate dehydrogenase is the enzyme that catalyzes the conversion of succinate to fumarate in the tricarboxylic acid cycle. It consists of two large subunits (a 70-kDa flavoprotein and a 30kDa iron-sulfur containing protein) which form complex II of the mitochondrial respiratory chain along with two smaller subunits, responsible for attaching SDH to the inner mitochondrial membrane [1]. Because complex II is the only component of the respiratory chain whose subunits are all encoded by the nDNA, SDH histochemistry is extremely useful for detecting any variation in the fiber distribution of mitochondria, independently of any alteration affecting the mtDNA. The histochemical method for the microscopic demonstration of SDH activity on frozen tissue sections is based on the use of a tetrazolium salt (nitro blue tetrazolium, NBT) as electron acceptor with phenazine methosulfate (PMS) serving as intermediate electron donor to NBT [34,32]. The specificity of the method may be tested by performing control experiments in which an SDH
Fig. 1. Histochemical stains for SDH and COX activities on serial muscle sections from a normal and a KSS patient. The normal shows a checkerboard pattern with both enzymes (a and b). The KSS samples show in one section one RRF (white asterisk) by SDH stain (c), and the same fiber on the serial section (black asterisk) shows lack of COX activity (d). Bars = 50 l.
276
K. Tanji, E. Bonilla / Methods 46 (2008) 274–280
largest polypeptides (I, II, and III), which are encoded by mtDNA and synthesized within the mitochondria, confer the catalytic and proton pumping activities to the enzyme. The ten smaller subunits are synthesized in the cytoplasm under the control of nuclear genes and are presumed to confer tissue specificity, thus adjusting the enzymatic activity to the metabolic demands of different tissues [24,10]. Additional nDNA-encoded factors are required for the assembly of COX, including those involved in the synthesis of heme a and a3, transport and insertion of copper atoms, and proper co-assembly of the mtDNA- and nDNA-encoded subunits. Several COX assembly genes have been identified in yeast [25,31], and recently pathogenic mutations in the human homologues of two of these genes, SURF1 and SCO2, have been discovered in patients with COX-deficient Leigh syndrome [55,48], and in patients with a cardioencephalomyopathy characterized by COX deficiency [30]. The dual genetic make up of COX and the availability of a reliable histochemical method to visualize its activity have made COX one of the ideal tools for basic investigations of mitochondrial biogenesis, nDNA–mtDNA interactions, and for the study of human mitochondrial disorders at both light and electron microscopic levels [7,22,52]. The histochemical method to visualize COX activity is based on the use of 3,30 -diaminobenzidine (DAB) as electron donor for cytochrome c [35]. The reaction product on oxidation of DAB occurs in the form of a brown pigmentation corresponding to the distribution of mitochondria in the tissue. The specificity of the method may also be tested by performing control experiments in which the COX inhibitor, potassium cyanide (0.01 M), is added to the incubation medium. As in the case of SDH, staining of normal muscle for COX activity also shows a checkerboard pattern. Type I fibers stain darker due to their mainly oxidative metabolism and more abundant mitochondria content and type II fibers show a finer and less intensely stained mitochondria network (Fig. 1b). The application of COX histochemistry to the investigation of KSS, MERRF, and MELAS has revealed one of the most important clues for the study of pathogenesis in these disorders. Muscle from KSS and MERRF patients shows a mosaic expression of COX consisting of a variable number of COX-deficient and COX-positive fibers [22] (Fig. 1d). Before the advent of molecular genetics, it was difficult to understand the reason for the appearance of this mosaic, but when it was discovered that these patients harbored mutations of mtDNA in their muscles, it became evident that the mosaic was an indicator of the heteroplasmic nature of the genetic defects. The mosaic pattern of COX expression in mitochondrial disorders is now considered the ‘‘histochemical signature” of a heteroplasmic mtDNA mutation affecting the expression of mtDNAencoded genes in skeletal muscle [8,37]. Muscle biopsies from patients with MELAS also show COX-deficient fibers. RRF are COX-positive, but the activity is decreased in the center of the fibers and largely preserved in the subsarcolemmal regions. It should also be noted that COX deficiency, but showing a more generalized pattern, as well as COX-negative RRF, are also observed in infants with depletion of muscle mtDNA or with either the fatal or the benign COX-deficient myopathies of infancy [13]. In addition, a generalized pattern of COX deficiency including not only extrafusal muscle fibers, but also intrafusal fibers and arterial walls of blood vessels, is seen in children affected with COX-deficient Leigh syndrome resulting from mutations in either SURF1 or SCO2 genes [55,48,30]. These observations indicate that histochemical studies in mitochondrial disorders provide significant information about both the nature and the pathogenesis of mitochondrial disorders. Moreover, they provide useful clues as to which molecular testing is needed to provide a specific diagnosis.
1.2. Immunohistochemistry The unique ability of immunohistochemistry to allow for the detection of specific proteins in single cells makes it a method of choice to study the expression of both mtDNA and nDNA genes in mitochondria of small and heterogeneous tissue samples. Recent technical advances have greatly increased the scope of immunohistochemistry and made it accessible to a variety of investigators with minimal expertise in immunology. Several immunological probes are presently available to perform immunohistochemical studies of mitochondria on frozen tissue sections. These include antibodies directed against mtDNAand nDNA-encoded subunits of the respiratory chain complexes other mitochondrial proteins, and antibodies against DNA that allow the detection of mtDNA [3,8,30]. Because the entire mitochondrial genome has been sequenced, any mtDNA-encoded respiratory chain subunit is potentially available for immunocytochemical studies, and it is anticipated that the same will soon be true for all the nDNA-encoded subunits of the respiratory chain [43]. There are several immunohistochemical methods for the study of mitochondria on tissue sections. These include enzyme-linked methods (peroxidase, alkaline phosphatase and glucose oxidase) and methods based on the application of fluorochromes. For studies on frozen tissue sections, we favor the use of fluorochromes because they allow for the direct visualization of the antigen– antibody binding sites, and because they are more flexible for double-labeling experiments. For studies of mitochondria on formalinfixed and paraffin-embedded samples, we routinely employ the avidin–biotin-peroxidase complex (ABC) method [21,5]. 1.2.1. Localization of nDNA and mtDNA-encoded subunits of the respiratory chain on frozen samples As mentioned earlier, we prefer immunolocalization via immunofluorescence on serial frozen sections, in particular methods using different fluorochromes for double-labeling studies. The main advantage of this approach is that it allows for the visualization of two different probes in the same mitochondria and in the same plane of section. These methods also eliminate the inferences that must be made with studies on serial sections, and they are particularly indicated in immunocytochemical investigations of mitochondria in non-syncitial tissues such as heart, kidney, and brain. In our laboratory, we routinely use a monoclonal antibody against COX IV as probe for a nDNA-encoded mitochondrial protein (www.invitrogen.com) and a polyclonal antibody against COX II as probe for a mtDNA-encoded protein (www.milipore.com/). For these studies, the sections are first incubated with both the polyclonal and the monoclonal antibodies at optimal dilution. That is the lowest concentration of the antibody giving a clear particulate immunostain corresponding to the localization of the mitochondria in normal muscle fibers. Subsequently, the sections are incubated with goat anti-rabbit IgG-fluorescein (to visualize the mtDNA probe in ‘‘green”) and goat anti-mouse IgG-Texas red (to visualize the nDNA probe in ‘‘red”). We carry out these studies with unfixed frozen sections, but with some antibodies it may be required to permeabilize the mitochondrial membranes to uncover the antigenic sequences or to facilitate the penetration of the probes into the inner mitochondrial compartment. In agreement with Jonhson et al. [23], we have also found that fixation of fresh frozen sections with 4% formaldehyde in 0.1 M CaCl2 pH 7 followed by dehydration in serial alcohols (outlined in the final section of this chapter) provides the most reproducible and successful results. Using unfixed muscle sections from normal samples, a checkerboard pattern resembling the one described for histochemistry is
K. Tanji, E. Bonilla / Methods 46 (2008) 274–280
usually observed, type I fibers appearing brighter due to their higher mitochondria content (Fig. 2a and b). In muscle sections from patients with KSS harboring a documented D-mtDNA, COX-deficient RRF typically show lack or marked reduction of COX II, whereas immunostain is typically normal in both COX-positive and COX-deficient fibers using antibodies against COX IV (Fig. 2c and d). Presumably, this is due to the fact that even the smallest deletion eliminates essential tRNAs that are required for translation of the mitochondrial genome [27]. Cell culture studies have confirmed this hypothesis: transfer of DmtDNA to mtDNA-less human cell lines produces a severe defect in the synthesis of the mtDNA-encoded subunits of the respiratory chain in recipient cells containing predominantly D-mtDNAs [20]. 1.2.2. Localization of mtDNA on frozen samples Immunohistochemistry using anti-DNA antibodies has been applied as an alternative method to in situ hybridization for the studies of localization and distribution of mtDNA in normal and pathological conditions [3,49]. The advantages of this method are that both mitochondrial and nuclear DNA is detected simultaneously, at the single cell level, and that the nuclear signal can be used as an internal control. For detection of mtDNA using immunological probes, we also carry out double-labeling experiments with different fluorochromes. We utilize polyclonal antibodies against COX IV for immunolabeling of mitochondria in one color (green), and a monoclonal antibody against DNA (www.milipore.com/) for immunostain of mtDNA and nDNA in another color (red).
277
In frozen muscle sections from normal controls and from patients without depletion of mtDNA, these antibodies show an intense staining of both the nuclei and a cytoplasmic network correlating with mitochondrial localization (Fig. 3a and b). Conversely, when muscle biopsies from patients with mtDNA depletion are analyzed, the particulate immunostaining of mtDNA is not detectable or it is only present in a small number of muscle fibers (Fig. 3c and d). The intensity of nDNA immunostaining shows no alteration compared with non-depleted controls. Immunohistochemistry utilizing antibodies against DNA is a useful method for the rapid evaluation of the distribution of mtDNA in normal cells and for the detection of depletion of muscle mtDNA. The method is particularly precise for the diagnosis of mtDNA depletion when it is confined to only a sub-population of fibers [50]. 1.2.3. Localization of mitochondrial proteins on paraffin-embedded samples As mentioned earlier, we prefer immunoperoxidase for the localization of mitochondria on parrafin-embedded brain samples, in particular the ABC method [21,5]. This method is based on the high affinity of avidin, an egg white protein for the vitamin biotin. In this technique, avidin can be view as an antibody against the biotin-labeled peroxidase. A reliable ABC kit can be obtained from a commercial source (www.vectorlabs.com/); alternatively, the ABC reagents can be prepared according to a previously published method [21]. Studies of the mitochondrial respiratory chain on paraffinembedded samples of brain, using the ABC method have provided significant information regarding the pathogenesis of neuronal dysfunction in mitochondrial disorders, and of the role of mitochondria in neurodegenerative disorders of the CNS including Parkinson and Alzheimer’s disease ([19,41,45,9]; Vermulst et al. this volume). For example, in patients with MELAS, studies of the mitochondrial respiratory chain in the hippocampal formation (HF) have shown a severe defect in the expression of the mtDNA-encoded subunit II of COX in neurons of the dentate gyrus (Fig. 4). Because the HF plays a critical role the processing of memory, it is possible that COX deficieny in the HF may be a contributing factor in the pathogenesis of the global cognitive deficit observed in patients with MELAS [28]. 1.3. Histochemical methods 1.3.1. SDH Collect 8 lm-thick cryostat sections on poly-L-lysine coated (0.1%) coverslips. Dissolve the following in 10 ml of 5 mM phosphate buffer, pH 7.4: -
Fig. 2. Immunolocalization of COX II and COX IV on muscle sections from a normal and a KSS patient using different fluorochromes. The normal muscle shows an almost identical mitochondrial network for COX II (a) and for COX IV (b). The KSS sample show two muscle fibers (white asterisks) that lacks COX II immunostain (d), and the same fibers (white asterisks) shows enhanced stain for COX IV (d). Bar = 50 l.
5 mM Ethylenediaminetetraacetic (EDTA). 1 mM Potassium cyanide (KCN). 0.2 mM Phenazine methosulfate (PMS), (www.fishersci.com). 50 mM Succinic acid. 1.5 mM Nitro blue tetrazolium (NBT), (www.fishersci.com). Adjust pH to 7.6. Filter solution with filter paper n 1. Incubate sections for 20 min at 37 °C. For control sections, sodium malonate (0.01 M) is added to the incubation medium. Rinse 5 min 3 times in distilled water, at room temperature (RT). Mount on glass slides with warm glycerin gel.
1.3.2. COX Collect 8 lm-thick cryostat sections on poly-L-lysine coated (0.1%) coverslips. Dissolve the following in 10 ml of 5 mM phosphate buffer, pH 7.4:
278
K. Tanji, E. Bonilla / Methods 46 (2008) 274–280
Fig. 3. Immunolocalization of COX IV and mtDNA on muscle sections from a normal and a patient with depletion of mtDNA using different fluorochromes. The normal muscle shows an almost identical mitochondrial network (white arrows) for COX IV (a) and for mtDNA (b). The mtDNA depleted sample shows two clearly defined RRF (white asterisks) with enhanced stain for COX IV (c), and the same fibers (white asterisks) show lack of stain for mtDNA (d). Bars = 50 l.
- 0.1% (10 mg) 3,30 -Diaminobenzidine (DAB), (www.sigmaaldrich. com/). - 0.1% (10 mg) Cytochrome c (from horse heart), (www.sigmaaldrich. com/). - 0.02% (2 mg) Catalase (www.sigmaaldrich.com/). Adjust pH to 7.4. Do not expose solution to light. Filter solution with filter paper n 1. Incubate sections for 1 h at 37 °C. For control sections, potassium cyanide (0.01 M) is added to the incubation medium. Rinse with distilled water, 3 5 min at RT. Mount on glass slides with warm glycerin gel. 2. Immunohistochemical methods 2.1. Double-labeling for the simultaneous visualization of mtDNA and nDNA-encoded subunits of the respiratory chain using different fluorochromes Collect 4 lm-thick cryostat sections on poly-L-lysine coated (0.1%) coverslips.
Fig. 4. Immunostaining of sections of the HF from a control (a and b) and from a MELAS patient (c and d) for the localization of the mtDNA-encoded COX II subunit of Complex IV (a and c) and the nDNA-encoded Fe S subunit of Complex III (b and d). The patient shows a marked decrease of immunostain in neurons of the dentate gyrus for COX II (arrows), but normal stain for Fe S. Bar = 50 l.
Incubate the sections for 2 h at RT (in a wet chamber) with antiCOX II polyclonal antibody and with anti-COX IV monoclonal antibody at optimal dilutions (1:100 to 1:500) in phosphate buffer saline containing 1% bovine serum albumin (PBS/BSA). Control sections are incubated without the primary antibodies. Rinse the samples with PBS 3 5 min at RT. Incubate the sections for 1 h at RT (in a wet chamber) with anti-rabbit IgG-fluorescein (www.invitrogen.com) and with anti-mouse IgGTexas red (www.invitrogen.com) diluted 1:100 in 1% BSA/PBS. Rinse the samples with PBS, 3 5 min at RT. Mount on slides with 50% glycerol in PBS.
K. Tanji, E. Bonilla / Methods 46 (2008) 274–280
2.2. Double-labeling for the simultaneous visualization of mitochondria and mtDNA using different fluorochromes Collect 4 lm-thick cryostat sections on poly-L-lysine coated (0.1%) coverslips. Fix the sections in 4% formaldehyde in 0.1 M CaCl2 pH 7 for 1 h at RT. Dehydrate the sections in 70%, 805, 90% ethyl alcohol, 5 min each, and in 100% ethyl alcohol, 15 min. Rinse the samples with PBS, 3 5 min at RT. Incubate the sections for 2 h at RT (in a wet chamber) with antiDNA monoclonal antibody (1:100) and with anti-COX IV polyclonal antibody (1:500) in 1% BSA/PBS. Control sections are incubated without the primary antibodies. Rinse the samples with PBS, 3 5 min. at RT. Incubate the sections for 30 min at RT (wet chamber) with biotynilated anti-mouse IgG (www.invitrogen.com) diluted 1:100 in 1% BSA/PBS. Rinse the samples with PBS, 3 5 min at RT. Incubate the sections for 30 min at RT (wet chamber) with Streptavidin-Texas Red (www.invitrogen.com) diluted 1:250), and with anti-rabbit IgG FITC (www.invitrogen.com) diluted 1:100) in 1% BSA/PBS. Rinse the samples with PBS, 3 5 min at RT. Mount on slides with glycerol-PBS 1:1.
2.3. Immunolocalization of mitochondrial proteins on paraffinembedded samples using the ABC method Collect 4 lm paraffin-embedded sections on poly (L-lysine)coated (0.1%) slides. Deparaffinize the sections through xylene and descending ethanol series (100%, 95%, 80%, and 75%). Incubate the sections in methanol containing 5% H2O2 for 30 min at RT. Place the slides in PBS, and then incubate the samples with 5% normal serum (from the same species as the host of the second antibody) for 1 h at RT. Incubate the slides with the primary antibody (1:1000–1:2000) overnight at 4 °C. Rinse the slides with PBS, 3 5 min at RT. Incubate the slides with the biotinilated second antibody at the optimal conditions (1:100–1:300) for 1 h at RT. Rinse the slides with PBS, 3 5 min at RT. Incubate the slides with ABC complex (prepare 30 min–1 h prior to use). Rinse the slides with PBS, 3 5 min at RT. Incubate the slides with DAB-H2O2 solution (40 mg of 3,30 diaminobenzidine tetrahydrochloride dissolved in 100 ml of PBS or 0.05 M Tris–HCl buffer (pH 7.6) containing 0.005% H2O2) for 1–3 min at RT. Rinse the slides with distilled water (dH2O) several times. Counterstain the slides briefly with hematoxylin. Rinse the slides with dH2O, dehydrate through ascending ethanol series, and clear in xylene. Mount the slides with synthetic resin (Permount).
Acknowledgment This work was supported by grants from the National Institutes of Health (NS11766 and PO1HD32062).
279
References [1] B.A.C. Ackrell, Mol. Aspects Med. 23 (2002) 369–384. [2] S. Anderson, A.T. Bankier, B.G. Barrel, N.H.L. de Bruin, A.R. Coulson, J. Drouin, I.C. Operon, D.P. Nierlich, B.A. Roe, F. Sanger, et al., Nature 290 (1981) 457–465. [3] F. Andreetta, H.J. Tritschler, E.A. Schon, S. DiMauro, E. Bonilla, J. Neurol. Sci. 105 (1991) 88–92. [4] G. Attardi, G. Schatz, Annu. Rev. Cell Biol. 4 (1988) 289–333. [5] C.D. Bedetti, J. Histochem. Cytochem. 33 (1985) 446–452. [6] D. Bogenhagen, D.A. Clayton, J. Biol. Chem. 249 (1984) 7991–7995. [7] E. Bonilla, D.L. Schotland, S. DiMauro, B. Aldover, J. Ultrastruct. Res. 51 (1975) 404–408. [8] E. Bonilla, M. Sciacco, K. Tanji, M. Sparaco, V. Petruzzella, C.T. Moraes, Brain Pathol. 2 (1992) 113–119. [9] E. Bonilla, K. Tanji, M. Hirano, T.H. Vu, S. DiMauro, E.A. Schon, Biochem. Biophys. Acta 1410 (1999) 171–182. [10] R.A. Capaldi, Arch. Biochem. Biophys. 280 (1990) 252–262. [11] C.E. Cooper, P. Nicholls, J.A. Freedman, Biochem. Cell Biol. 69 (1991) 596–607. [12] S. DiMauro, E.A. Schon, N. Engl. J. Med. 348 (2003) 2656–2668. [13] S. DiMauro, E. Bonilla, in: A.G. Engel, C. Franzini-Armstrong (Eds.), Myology Third Edition, McGraw-Hill, New York, 2004, pp. 1623–1662. [14] W.K. Engel, G.G. Cunningham, Neurology 13 (1963) 919–926. [15] R.E. Giles, H. Blanc, H.M. Cann, D.C. Wallace, Proc. Natl. Acad. Sci. USA 77 (1980) 6715–6719. [16] Y.I. Goto, I. Nonaka, S. Horai, Nature 348 (1990) 651–653. [17] Y.I. Goto, I. Nonaka, S. Horai, Biochem. Biophys. Acta 1097 (1991) 238–240. [18] R. Haller et al., J. Clin. Invest. 88 (1991) 1197–1206. [19] N. Hattori, M. Tanaka, T. Ozawa, Y. Mizuno, Ann. Neurol. 30 (1990) 563–571. [20] J.-I. Hayashi, S. Ohta, A. Kikuchi, M. Takemitsu, Y.I. Goto, I. Nonaka, Proc. Natl. Acad. Sci. USA 88 (1991) 10614–10618. [21] S.M. Hsu, L. Raine, H. Fanger, J. Histochem. Cytochem. 29 (1981) 577–580. [22] M.A. Johnson, D.M. Turnbull, D.J. Dick, H.S. Sherratt, J. Neurol. Sci. 60 (1983) 31–53. [23] M.A. Jonhson, B. Kadenbach, M. Droste, S.L. Old, D.M. Turnbull, J. Neurol. Sci. 87 (1988) 75–90. [24] B. Kadenbach, L. Kuhn-Nentwig, U. Buge, Curr. Top. Bioenerg. 15 (1987) 113– 161. [25] B. Kloeckener-Gruissem, J.E. McEwen, R.O. Poyton, Curr. Genet. 12 (1987) 311– 322. [26] C.T. Moraes, S. DiMauro, M. Zeviani, A. Lombes, S. Shanske, A.F. Miranda, H. Nakase, E. Bonilla, L.C. Werneck, S. Servidei, et al., N. Engl. J. Med. 320 (1989) 1293–1299. [27] H. Nakase, C.T. Moraes, R. Rizzuto, A. Lombes, S. DiMauoro, E.A. Schon, Am. J. Hum. Genet. 46 (1990) 418–427. [28] S.A. Neargarder, M.P. Murtagh, B. Wong, I.K. Hill, Cogn. Behav. Neurol. 2 (2007) 83–92. [29] W. Neupert, Ann. Rev. Biochem. 66 (1997) 863–917. [30] L.C. Papadopoulou, C.M. Sue, M. Davidson, K. Tanji, I. Nishino, J. Sadlock, J. Selby, D. Moira Glerum, R. Van Coster, G. Lyon, et al., Nat. Genet. 23 (1999) 333–337. [31] H.J. Pel, A. Tzagoloff, L.A. Grivell, Curr. Genet. 21 (1992) 139–146. [32] D. Pette, Histochem. J. 13 (1981) 319–327. [33] M. Satoh, T. Kuroiwa, Exp. Cell Res. 196 (1991) 137–140. [34] A.M. Seligman, R.M. Rutenburg, Science 113 (1951) 317–320. [35] A.M. Seligman, M.J. Karnovsky, H.L. Wasserkrug, J.S. Hanker, J. Cell Biol. 38 (1968) 1–15. [36] J.M. Shoffner, M.T. Lott, A.M.S. Lezza, P. Seibel, S.W. Ballinger, D.C. Wallace, Cell 61 (1990) 931–937. [37] E.A. Shoubridge, in: S. DiMauro, D.C. Wallace (Eds.), Mitochondrial DNA in Human Pathology, Raven Press, New York, 1993, pp. 109–123. [38] E.A. Shoubridge, Hum. Mol. Genet. 10 (2001) 2277–2284. [39] G. Silvestri, C.T. Moraes, S. Shanske, S.J. Oh, S. DiMauro, Am. J. Hum. Genet. 51 (1992) 1213–1217. [40] J.A. Smeitink, M. Zeviani, D.M. Turnbull, H.T. Jacobs, Cell Metb. 3 (2006) 9–13. [41] M. Sparaco, E.A. Schon, S. DiMauro, E. Bonilla, Brain Pathol. 5 (1995) 125–133. [42] M.J. Szabolics, R. Seigle, S. Shanske, E. Bonilla, S. DiMauro, V. D’Agati, Kidney Int. 45 (1994) 1388–1396. [43] J.W. Taanman, M.D. Burton, M.F. Marusich, N.G. Kennaway, R.A. Capaldi, Biochem. Biophys. Acta 1315 (1996) 199–207. [44] J.W. Taanman, J. Bioenerg. Biomembr. 29 (1997) 151–163. [45] K. Tanji, T.H. Vu, E.A. Schon, S. DiMauro, E. Bonilla, Ann. Neurol. 45 (1999) 377– 383. [46] K. Tanji, J. Gamez, C. Cervera, F. Mearin, A. Ortega, J. Torre, J. Montoya, A.L. Andreu, S. DiMauro, E. Bonilla, Acta Neuropathol. 105 (2003) 69–75. [47] R.W. Taylor, M.A. Birch-Machin, J. Schaefer, L. Taylor, R. Shakir, B.A.C. Ackrell, B. Cochran, L.A. Bindoff, M.J. Jackson, P. Griffiths, et al., Ann. Neurol. 39 (1996) 224–232. [48] V. Tiranti, K. Hoertnagel, R. Carrozo, C. Galimberti, M. Munaro, M. Granatiero, L. Zelante, P. Gasparini, R. Marzella, M. Rocchi, et al., Am. J. Hum. Genet. 63 (1998) 1609–1621.
280
K. Tanji, E. Bonilla / Methods 46 (2008) 274–280
[49] H.J. Tritschler, F. Andreetta, C.T. Moraes, E. Bonilla, E. Arnaudo, M.J. Danon, S. Glass, B.M. Zalaya, E. Vamos, S. Shanske, et al., Neurology 42 (1992) 209–217. [50] T.H. Vu, K. Tanji, H. Valsamis, S. DiMauro, E. Bonilla, Neurology 51 (1998) 1190–1193. [51] D.C. Wallace, Annu. Rev. Biochem. 61 (1992) 1175–1212.
[52] M.T.T. Wong-Riley, Trends Neurosci. 12 (1989) 94–101. [53] M. Zeviani, C.T. Moraes, S. DiMauro, H. Nakase, E. Bonilla, E.A. Schon, L.P. Rowland, Neurology 38 (1988) 1339–1346. [54] M. Zeviani, S. Di Donato, Brain 127 (2004) 2153–2172. [55] Z. Zhu, J. Yao, T. Johns, K. Fu, I. De Bie, C. Macmillan, A.P. Cuthber, R.F. Newbold, J. Wang, M. Chevrette, et al., Nat. Genet. 20 (1998) 337–343.
Contents lists available at ScienceDirect
Methods journal homepage: www.elsevier.com/locate/ymeth
Review Article
Light microscopic methods to visualize mitochondria on tissue sections Kurenai Tanji a, Eduardo Bonilla a,b,* a b
Department of Pathology, Columbia University, New York, NY 10032, USA Department of Neurology, Columbia University, 1150 St. Nicholas Avenue, Room 3-318, New York, NY 10032, USA
a r t i c l e
i n f o
Article history: Accepted 30 September 2008 Available online 16 October 2008 Keywords: Histochemistry Immunohistochemistry Tissue sections Light microscope Mitochondrial dysfunction
a b s t r a c t Mitochondria are cytoplasmic, double-membrane organelles, a main role of which is to synthesize ATP, the universal energy ‘supply’ of cells. In the last three decades, molecular genetic, biochemical, immunological and cell biological techniques have been applied in a coordinated fashion to unveil the pathogenesis of known mitochondrial disorders, as well as to explore the role of mitochondria in aging and neurodegenerative diseases. Once to be thought to be rare, it is now clear that mitochondrial dysfunction is an important cause of neurological and cardiac diseases, and age-related disorders such as cancer. Here, we review, illustrate, and provide updated protocols of two histochemical, and three immunohistochemical methods that in our opinion are the most reliable tools to visualize mitochondria on tissue sections from normal and disease specimens. Ó 2008 Elsevier Inc. All rights reserved.
1. Introduction Mitochondria are the primary ATP-generating organelles in most mammalian cells and they contain their own DNA (mtDNA) which is maternally inherited [2,15]. ATP is produced via oxidative phosphorylation through five respiratory complexes located in the inner mitochondrial membrane. The human mitochondrial genome is a 16,569-bp double-stranded DNA. It is highly compact, and contains only 37 genes: 2 genes encode ribosomal RNAs (rRNAs), 22 encode transfer RNAs (tRNAs), and 13 encode polypeptides. All 13 polypeptides are components of the respiratory chain, including 7 subunits of complex I or NADH dehydrogenase-ubiquinone oxidoreductase, 1 subunit of complex III or ubiquinone-cytochrome c oxidoreductase, 3 subunits of complex IV—cytochrome c oxidase (COX), and 2 subunits of complex V or ATP synthase [4]. The respiratory complexes also contain nuclear DNA (nDNA) encoded sububits, which are imported into the organelle from the cytosol and assembled, together, with the mtDNA-encoded subunits, into the respective holoenzymes in the mitochondrial inner membrane ([29]; Smeitink et al. [40]; Medeiros et al. this volume). Complex II or succinate dehydrogenase-ubiquinone oxidoreductase contains only nDNA-encoded subunits. The mitochondrial disorders encompass a heterogeneous group of diseases in which mitochondrial dysfunction produces clinical manifestations. Because of the dual genetic make up of mitochondria, these diseases are typically caused by genetic errors either in mtDNA or nDNA ([54]; Calvaruso et al. this volume).
* Corresponding author. Address: Department of Neurology, Columbia University, 1150 St. Nicholas Avenue, Room 3-318, New York, NY 10032, USA. E-mail address: [email protected] (E. Bonilla). 1046-2023/$ - see front matter Ó 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.ymeth.2008.09.027
Pathogenic mtDNA mutations have been identified in three of the ‘‘prototypes” of the mitochondrial disorders. First, large scale mtDNA rearrangements (i.e. deletions [D-mtDNA] and/or duplications [dup-mtDNA]) have been associated with sporadic KearnsSayre syndrome (KSS), and are often seen in patients with isolated ocular myopathy (OM) [53,26]. Second, myoclonus epilepsy with ragged-red fibers (MERRF) has been frequently associated with two different point mutations, both in the tRNALys gene [36,39]. Third, mitochondrial encephalopathy, lactic acidosis and strokelike episodes (MELAS) has been mainly associated with two different point mutations, both in the tRNALeu(UUR) gene [16,17]. Other point mutations in tRNAs and protein-coding genes of the mitochondrial genome as well as disorders leading to generation of multiple D-mtDNAs or to depletion of mtDNA but due to primary nDNA defects have now been described [54]. Because mitochondria are inherited only from the mother [15], pedigrees harboring defects in mtDNA genes should exhibit maternal inheritance. Moreover, because there are hundreds or even thousands of mitochondria in each cell, with an average of 2–10 mtDNAs per organelle [6,33], mutations in mtDNA result in two populations of mtDNAs—mutated and wild-type—a condition known as heteroplasmy. The phenotypic expression of a mtDNA mutation is regulated by the threshold effect, that is, the mutant phenotype is expressed in the heteroplasmic cells only when the relative proportion of mutant mtDNAs exceeds a minimum value [51]. Research on mitochondrial diseases has also uncovered an increasing number of disorders that are caused by mutations in nuclear genes encoding the subunits of the respiratory chain, or other proteins that are essential for the biosynthesis of specific cofactors or assembly of the complexes. Interestingly, complex I and II defects are associated with mutations in genes encoding the subunits of the complexes, while complex III, IV, and V defects are caused by
K. Tanji, E. Bonilla / Methods 46 (2008) 274–280
275
mutations in specific assembly proteins or biosynthetic factors [38,54]. Nerve and muscle, whose function is highly dependent on oxidative metabolism, are the most severely affected tissues in the mitochondrial disorders [12]. Consequently, genetic as well as morphologic studies of muscle and brain from patients with mitochondrial diseases have been proven fundamental to understand the pathogenesis of mitochondrial dysfunction at the level of individual muscle fibers and in different neuronal nuclei [8,41,45]. The purpose of this chapter is to present the histochemical and immunohistochemical, methods that, in our experience, appear to be the most reliable for the correct identification of mitochondria on frozen or paraffin-embedded tissue sections, to illustrate their potential using specific pathological samples, and to provide an updated version of the methods. While the described protocols refer to skeletal muscle or brain mitochondria, the methods described can be applied to any cell type [42,46]. It is not our intention to cover every study or method related to morphological aspects of mitochondria, but rather to provide enough information to allow investigators to apply these selected light microscopy tools to a particular scientific or diagnostic question.
inhibitor, sodium malonate (0.01 M), is added to the incubation medium. Using this method for detecting SDH activity in normal muscle sections, two populations of fibers are seen resulting in a checkerboard pattern. Type II fibers, which rely on glycolytic metabolism, show a light blue network-like stain. Type I fibers, whose metabolism is highly oxidative and therefore contain more mitochondria, show a more elaborate and darker mitochondrial network (Fig. 1a). In samples with pathological proliferation of mitochondria (RRF), the RRF show an intense blue SDH reaction corresponding to the distribution of the mitochondria within the fiber (Fig. 1c). This proliferation of mitochondria is associated with most mtDNA defects (deletions and tRNA point mutations), but RRF can also be observed in other disorders that are thought to be due to defects of nDNA, such as the depletion of muscle mtDNA, and the fatal and benign COX-deficient myopathies of infancy [13]. SDH histochemistry is also useful for the diagnosis of complex II deficiency. Several patients with myopathy and complex II deficiency have been reported. In agreement with the biochemical observations, SDH histochemistry showed complete lack of reaction in muscle [18,47].
1.1. Histochemistry
1.1.2. COX COX, the last component of the respiratory chain, catalyzes the transfer of reducing equivalents from cytochrome c to molecular oxygen. The holoenzyme contains two heme a moieties (a and a3) and three copper atoms (two in the CuA site and one in the and CuB site) bound to a multisubunit protein frame embedded in the mitochondrial inner membrane [11,44]. In mammals, the apoenzyme is composed of thirteen different subunits. The three
The visualization of normal and pathological mitochondria on frozen tissue sections can be carried out using a number of cytochemical techniques. These include the modified Gomori trichrome and hematoxylin-eosin stains, and histochemical methods for the demonstration of oxidative enzyme activity. The most informative histochemical alteration of mitochondria in skeletal muscle is the ragged-red fiber (RRF), observed on frozen sections stained with the trichrome method of Engel and Cunningham [14]. The name derives from the reddish appearance of the trichrome-stained muscle fiber as a result of subsarcolemmal and/or intermyofibrillar proliferation of the mitochondria. The fibers harboring abnormal deposits of mitochondria are most often type I myofibers, and they may also contain increased numbers of lipid droplets. Since accumulations of materials other than mitochondria may simulate RRF formation, the identification of deposits suspected of being mitochondrial proliferation should be confirmed histochemically by the application of oxidative enzyme stains. In our experience, enzyme histochemistry for the activity of succinate dehydrogenase (SDH) and COX has proven to be the most reliable methods for the correct visualization of normal mitochondria, and for the interpretation and ultimately the diagnosis of some of the mitochondrial disorders affecting skeletal muscle [13]. 1.1.1. SDH Succinate dehydrogenase is the enzyme that catalyzes the conversion of succinate to fumarate in the tricarboxylic acid cycle. It consists of two large subunits (a 70-kDa flavoprotein and a 30kDa iron-sulfur containing protein) which form complex II of the mitochondrial respiratory chain along with two smaller subunits, responsible for attaching SDH to the inner mitochondrial membrane [1]. Because complex II is the only component of the respiratory chain whose subunits are all encoded by the nDNA, SDH histochemistry is extremely useful for detecting any variation in the fiber distribution of mitochondria, independently of any alteration affecting the mtDNA. The histochemical method for the microscopic demonstration of SDH activity on frozen tissue sections is based on the use of a tetrazolium salt (nitro blue tetrazolium, NBT) as electron acceptor with phenazine methosulfate (PMS) serving as intermediate electron donor to NBT [34,32]. The specificity of the method may be tested by performing control experiments in which an SDH
Fig. 1. Histochemical stains for SDH and COX activities on serial muscle sections from a normal and a KSS patient. The normal shows a checkerboard pattern with both enzymes (a and b). The KSS samples show in one section one RRF (white asterisk) by SDH stain (c), and the same fiber on the serial section (black asterisk) shows lack of COX activity (d). Bars = 50 l.
276
K. Tanji, E. Bonilla / Methods 46 (2008) 274–280
largest polypeptides (I, II, and III), which are encoded by mtDNA and synthesized within the mitochondria, confer the catalytic and proton pumping activities to the enzyme. The ten smaller subunits are synthesized in the cytoplasm under the control of nuclear genes and are presumed to confer tissue specificity, thus adjusting the enzymatic activity to the metabolic demands of different tissues [24,10]. Additional nDNA-encoded factors are required for the assembly of COX, including those involved in the synthesis of heme a and a3, transport and insertion of copper atoms, and proper co-assembly of the mtDNA- and nDNA-encoded subunits. Several COX assembly genes have been identified in yeast [25,31], and recently pathogenic mutations in the human homologues of two of these genes, SURF1 and SCO2, have been discovered in patients with COX-deficient Leigh syndrome [55,48], and in patients with a cardioencephalomyopathy characterized by COX deficiency [30]. The dual genetic make up of COX and the availability of a reliable histochemical method to visualize its activity have made COX one of the ideal tools for basic investigations of mitochondrial biogenesis, nDNA–mtDNA interactions, and for the study of human mitochondrial disorders at both light and electron microscopic levels [7,22,52]. The histochemical method to visualize COX activity is based on the use of 3,30 -diaminobenzidine (DAB) as electron donor for cytochrome c [35]. The reaction product on oxidation of DAB occurs in the form of a brown pigmentation corresponding to the distribution of mitochondria in the tissue. The specificity of the method may also be tested by performing control experiments in which the COX inhibitor, potassium cyanide (0.01 M), is added to the incubation medium. As in the case of SDH, staining of normal muscle for COX activity also shows a checkerboard pattern. Type I fibers stain darker due to their mainly oxidative metabolism and more abundant mitochondria content and type II fibers show a finer and less intensely stained mitochondria network (Fig. 1b). The application of COX histochemistry to the investigation of KSS, MERRF, and MELAS has revealed one of the most important clues for the study of pathogenesis in these disorders. Muscle from KSS and MERRF patients shows a mosaic expression of COX consisting of a variable number of COX-deficient and COX-positive fibers [22] (Fig. 1d). Before the advent of molecular genetics, it was difficult to understand the reason for the appearance of this mosaic, but when it was discovered that these patients harbored mutations of mtDNA in their muscles, it became evident that the mosaic was an indicator of the heteroplasmic nature of the genetic defects. The mosaic pattern of COX expression in mitochondrial disorders is now considered the ‘‘histochemical signature” of a heteroplasmic mtDNA mutation affecting the expression of mtDNAencoded genes in skeletal muscle [8,37]. Muscle biopsies from patients with MELAS also show COX-deficient fibers. RRF are COX-positive, but the activity is decreased in the center of the fibers and largely preserved in the subsarcolemmal regions. It should also be noted that COX deficiency, but showing a more generalized pattern, as well as COX-negative RRF, are also observed in infants with depletion of muscle mtDNA or with either the fatal or the benign COX-deficient myopathies of infancy [13]. In addition, a generalized pattern of COX deficiency including not only extrafusal muscle fibers, but also intrafusal fibers and arterial walls of blood vessels, is seen in children affected with COX-deficient Leigh syndrome resulting from mutations in either SURF1 or SCO2 genes [55,48,30]. These observations indicate that histochemical studies in mitochondrial disorders provide significant information about both the nature and the pathogenesis of mitochondrial disorders. Moreover, they provide useful clues as to which molecular testing is needed to provide a specific diagnosis.
1.2. Immunohistochemistry The unique ability of immunohistochemistry to allow for the detection of specific proteins in single cells makes it a method of choice to study the expression of both mtDNA and nDNA genes in mitochondria of small and heterogeneous tissue samples. Recent technical advances have greatly increased the scope of immunohistochemistry and made it accessible to a variety of investigators with minimal expertise in immunology. Several immunological probes are presently available to perform immunohistochemical studies of mitochondria on frozen tissue sections. These include antibodies directed against mtDNAand nDNA-encoded subunits of the respiratory chain complexes other mitochondrial proteins, and antibodies against DNA that allow the detection of mtDNA [3,8,30]. Because the entire mitochondrial genome has been sequenced, any mtDNA-encoded respiratory chain subunit is potentially available for immunocytochemical studies, and it is anticipated that the same will soon be true for all the nDNA-encoded subunits of the respiratory chain [43]. There are several immunohistochemical methods for the study of mitochondria on tissue sections. These include enzyme-linked methods (peroxidase, alkaline phosphatase and glucose oxidase) and methods based on the application of fluorochromes. For studies on frozen tissue sections, we favor the use of fluorochromes because they allow for the direct visualization of the antigen– antibody binding sites, and because they are more flexible for double-labeling experiments. For studies of mitochondria on formalinfixed and paraffin-embedded samples, we routinely employ the avidin–biotin-peroxidase complex (ABC) method [21,5]. 1.2.1. Localization of nDNA and mtDNA-encoded subunits of the respiratory chain on frozen samples As mentioned earlier, we prefer immunolocalization via immunofluorescence on serial frozen sections, in particular methods using different fluorochromes for double-labeling studies. The main advantage of this approach is that it allows for the visualization of two different probes in the same mitochondria and in the same plane of section. These methods also eliminate the inferences that must be made with studies on serial sections, and they are particularly indicated in immunocytochemical investigations of mitochondria in non-syncitial tissues such as heart, kidney, and brain. In our laboratory, we routinely use a monoclonal antibody against COX IV as probe for a nDNA-encoded mitochondrial protein (www.invitrogen.com) and a polyclonal antibody against COX II as probe for a mtDNA-encoded protein (www.milipore.com/). For these studies, the sections are first incubated with both the polyclonal and the monoclonal antibodies at optimal dilution. That is the lowest concentration of the antibody giving a clear particulate immunostain corresponding to the localization of the mitochondria in normal muscle fibers. Subsequently, the sections are incubated with goat anti-rabbit IgG-fluorescein (to visualize the mtDNA probe in ‘‘green”) and goat anti-mouse IgG-Texas red (to visualize the nDNA probe in ‘‘red”). We carry out these studies with unfixed frozen sections, but with some antibodies it may be required to permeabilize the mitochondrial membranes to uncover the antigenic sequences or to facilitate the penetration of the probes into the inner mitochondrial compartment. In agreement with Jonhson et al. [23], we have also found that fixation of fresh frozen sections with 4% formaldehyde in 0.1 M CaCl2 pH 7 followed by dehydration in serial alcohols (outlined in the final section of this chapter) provides the most reproducible and successful results. Using unfixed muscle sections from normal samples, a checkerboard pattern resembling the one described for histochemistry is
K. Tanji, E. Bonilla / Methods 46 (2008) 274–280
usually observed, type I fibers appearing brighter due to their higher mitochondria content (Fig. 2a and b). In muscle sections from patients with KSS harboring a documented D-mtDNA, COX-deficient RRF typically show lack or marked reduction of COX II, whereas immunostain is typically normal in both COX-positive and COX-deficient fibers using antibodies against COX IV (Fig. 2c and d). Presumably, this is due to the fact that even the smallest deletion eliminates essential tRNAs that are required for translation of the mitochondrial genome [27]. Cell culture studies have confirmed this hypothesis: transfer of DmtDNA to mtDNA-less human cell lines produces a severe defect in the synthesis of the mtDNA-encoded subunits of the respiratory chain in recipient cells containing predominantly D-mtDNAs [20]. 1.2.2. Localization of mtDNA on frozen samples Immunohistochemistry using anti-DNA antibodies has been applied as an alternative method to in situ hybridization for the studies of localization and distribution of mtDNA in normal and pathological conditions [3,49]. The advantages of this method are that both mitochondrial and nuclear DNA is detected simultaneously, at the single cell level, and that the nuclear signal can be used as an internal control. For detection of mtDNA using immunological probes, we also carry out double-labeling experiments with different fluorochromes. We utilize polyclonal antibodies against COX IV for immunolabeling of mitochondria in one color (green), and a monoclonal antibody against DNA (www.milipore.com/) for immunostain of mtDNA and nDNA in another color (red).
277
In frozen muscle sections from normal controls and from patients without depletion of mtDNA, these antibodies show an intense staining of both the nuclei and a cytoplasmic network correlating with mitochondrial localization (Fig. 3a and b). Conversely, when muscle biopsies from patients with mtDNA depletion are analyzed, the particulate immunostaining of mtDNA is not detectable or it is only present in a small number of muscle fibers (Fig. 3c and d). The intensity of nDNA immunostaining shows no alteration compared with non-depleted controls. Immunohistochemistry utilizing antibodies against DNA is a useful method for the rapid evaluation of the distribution of mtDNA in normal cells and for the detection of depletion of muscle mtDNA. The method is particularly precise for the diagnosis of mtDNA depletion when it is confined to only a sub-population of fibers [50]. 1.2.3. Localization of mitochondrial proteins on paraffin-embedded samples As mentioned earlier, we prefer immunoperoxidase for the localization of mitochondria on parrafin-embedded brain samples, in particular the ABC method [21,5]. This method is based on the high affinity of avidin, an egg white protein for the vitamin biotin. In this technique, avidin can be view as an antibody against the biotin-labeled peroxidase. A reliable ABC kit can be obtained from a commercial source (www.vectorlabs.com/); alternatively, the ABC reagents can be prepared according to a previously published method [21]. Studies of the mitochondrial respiratory chain on paraffinembedded samples of brain, using the ABC method have provided significant information regarding the pathogenesis of neuronal dysfunction in mitochondrial disorders, and of the role of mitochondria in neurodegenerative disorders of the CNS including Parkinson and Alzheimer’s disease ([19,41,45,9]; Vermulst et al. this volume). For example, in patients with MELAS, studies of the mitochondrial respiratory chain in the hippocampal formation (HF) have shown a severe defect in the expression of the mtDNA-encoded subunit II of COX in neurons of the dentate gyrus (Fig. 4). Because the HF plays a critical role the processing of memory, it is possible that COX deficieny in the HF may be a contributing factor in the pathogenesis of the global cognitive deficit observed in patients with MELAS [28]. 1.3. Histochemical methods 1.3.1. SDH Collect 8 lm-thick cryostat sections on poly-L-lysine coated (0.1%) coverslips. Dissolve the following in 10 ml of 5 mM phosphate buffer, pH 7.4: -
Fig. 2. Immunolocalization of COX II and COX IV on muscle sections from a normal and a KSS patient using different fluorochromes. The normal muscle shows an almost identical mitochondrial network for COX II (a) and for COX IV (b). The KSS sample show two muscle fibers (white asterisks) that lacks COX II immunostain (d), and the same fibers (white asterisks) shows enhanced stain for COX IV (d). Bar = 50 l.
5 mM Ethylenediaminetetraacetic (EDTA). 1 mM Potassium cyanide (KCN). 0.2 mM Phenazine methosulfate (PMS), (www.fishersci.com). 50 mM Succinic acid. 1.5 mM Nitro blue tetrazolium (NBT), (www.fishersci.com). Adjust pH to 7.6. Filter solution with filter paper n 1. Incubate sections for 20 min at 37 °C. For control sections, sodium malonate (0.01 M) is added to the incubation medium. Rinse 5 min 3 times in distilled water, at room temperature (RT). Mount on glass slides with warm glycerin gel.
1.3.2. COX Collect 8 lm-thick cryostat sections on poly-L-lysine coated (0.1%) coverslips. Dissolve the following in 10 ml of 5 mM phosphate buffer, pH 7.4:
278
K. Tanji, E. Bonilla / Methods 46 (2008) 274–280
Fig. 3. Immunolocalization of COX IV and mtDNA on muscle sections from a normal and a patient with depletion of mtDNA using different fluorochromes. The normal muscle shows an almost identical mitochondrial network (white arrows) for COX IV (a) and for mtDNA (b). The mtDNA depleted sample shows two clearly defined RRF (white asterisks) with enhanced stain for COX IV (c), and the same fibers (white asterisks) show lack of stain for mtDNA (d). Bars = 50 l.
- 0.1% (10 mg) 3,30 -Diaminobenzidine (DAB), (www.sigmaaldrich. com/). - 0.1% (10 mg) Cytochrome c (from horse heart), (www.sigmaaldrich. com/). - 0.02% (2 mg) Catalase (www.sigmaaldrich.com/). Adjust pH to 7.4. Do not expose solution to light. Filter solution with filter paper n 1. Incubate sections for 1 h at 37 °C. For control sections, potassium cyanide (0.01 M) is added to the incubation medium. Rinse with distilled water, 3 5 min at RT. Mount on glass slides with warm glycerin gel. 2. Immunohistochemical methods 2.1. Double-labeling for the simultaneous visualization of mtDNA and nDNA-encoded subunits of the respiratory chain using different fluorochromes Collect 4 lm-thick cryostat sections on poly-L-lysine coated (0.1%) coverslips.
Fig. 4. Immunostaining of sections of the HF from a control (a and b) and from a MELAS patient (c and d) for the localization of the mtDNA-encoded COX II subunit of Complex IV (a and c) and the nDNA-encoded Fe S subunit of Complex III (b and d). The patient shows a marked decrease of immunostain in neurons of the dentate gyrus for COX II (arrows), but normal stain for Fe S. Bar = 50 l.
Incubate the sections for 2 h at RT (in a wet chamber) with antiCOX II polyclonal antibody and with anti-COX IV monoclonal antibody at optimal dilutions (1:100 to 1:500) in phosphate buffer saline containing 1% bovine serum albumin (PBS/BSA). Control sections are incubated without the primary antibodies. Rinse the samples with PBS 3 5 min at RT. Incubate the sections for 1 h at RT (in a wet chamber) with anti-rabbit IgG-fluorescein (www.invitrogen.com) and with anti-mouse IgGTexas red (www.invitrogen.com) diluted 1:100 in 1% BSA/PBS. Rinse the samples with PBS, 3 5 min at RT. Mount on slides with 50% glycerol in PBS.
K. Tanji, E. Bonilla / Methods 46 (2008) 274–280
2.2. Double-labeling for the simultaneous visualization of mitochondria and mtDNA using different fluorochromes Collect 4 lm-thick cryostat sections on poly-L-lysine coated (0.1%) coverslips. Fix the sections in 4% formaldehyde in 0.1 M CaCl2 pH 7 for 1 h at RT. Dehydrate the sections in 70%, 805, 90% ethyl alcohol, 5 min each, and in 100% ethyl alcohol, 15 min. Rinse the samples with PBS, 3 5 min at RT. Incubate the sections for 2 h at RT (in a wet chamber) with antiDNA monoclonal antibody (1:100) and with anti-COX IV polyclonal antibody (1:500) in 1% BSA/PBS. Control sections are incubated without the primary antibodies. Rinse the samples with PBS, 3 5 min. at RT. Incubate the sections for 30 min at RT (wet chamber) with biotynilated anti-mouse IgG (www.invitrogen.com) diluted 1:100 in 1% BSA/PBS. Rinse the samples with PBS, 3 5 min at RT. Incubate the sections for 30 min at RT (wet chamber) with Streptavidin-Texas Red (www.invitrogen.com) diluted 1:250), and with anti-rabbit IgG FITC (www.invitrogen.com) diluted 1:100) in 1% BSA/PBS. Rinse the samples with PBS, 3 5 min at RT. Mount on slides with glycerol-PBS 1:1.
2.3. Immunolocalization of mitochondrial proteins on paraffinembedded samples using the ABC method Collect 4 lm paraffin-embedded sections on poly (L-lysine)coated (0.1%) slides. Deparaffinize the sections through xylene and descending ethanol series (100%, 95%, 80%, and 75%). Incubate the sections in methanol containing 5% H2O2 for 30 min at RT. Place the slides in PBS, and then incubate the samples with 5% normal serum (from the same species as the host of the second antibody) for 1 h at RT. Incubate the slides with the primary antibody (1:1000–1:2000) overnight at 4 °C. Rinse the slides with PBS, 3 5 min at RT. Incubate the slides with the biotinilated second antibody at the optimal conditions (1:100–1:300) for 1 h at RT. Rinse the slides with PBS, 3 5 min at RT. Incubate the slides with ABC complex (prepare 30 min–1 h prior to use). Rinse the slides with PBS, 3 5 min at RT. Incubate the slides with DAB-H2O2 solution (40 mg of 3,30 diaminobenzidine tetrahydrochloride dissolved in 100 ml of PBS or 0.05 M Tris–HCl buffer (pH 7.6) containing 0.005% H2O2) for 1–3 min at RT. Rinse the slides with distilled water (dH2O) several times. Counterstain the slides briefly with hematoxylin. Rinse the slides with dH2O, dehydrate through ascending ethanol series, and clear in xylene. Mount the slides with synthetic resin (Permount).
Acknowledgment This work was supported by grants from the National Institutes of Health (NS11766 and PO1HD32062).
279
References [1] B.A.C. Ackrell, Mol. Aspects Med. 23 (2002) 369–384. [2] S. Anderson, A.T. Bankier, B.G. Barrel, N.H.L. de Bruin, A.R. Coulson, J. Drouin, I.C. Operon, D.P. Nierlich, B.A. Roe, F. Sanger, et al., Nature 290 (1981) 457–465. [3] F. Andreetta, H.J. Tritschler, E.A. Schon, S. DiMauro, E. Bonilla, J. Neurol. Sci. 105 (1991) 88–92. [4] G. Attardi, G. Schatz, Annu. Rev. Cell Biol. 4 (1988) 289–333. [5] C.D. Bedetti, J. Histochem. Cytochem. 33 (1985) 446–452. [6] D. Bogenhagen, D.A. Clayton, J. Biol. Chem. 249 (1984) 7991–7995. [7] E. Bonilla, D.L. Schotland, S. DiMauro, B. Aldover, J. Ultrastruct. Res. 51 (1975) 404–408. [8] E. Bonilla, M. Sciacco, K. Tanji, M. Sparaco, V. Petruzzella, C.T. Moraes, Brain Pathol. 2 (1992) 113–119. [9] E. Bonilla, K. Tanji, M. Hirano, T.H. Vu, S. DiMauro, E.A. Schon, Biochem. Biophys. Acta 1410 (1999) 171–182. [10] R.A. Capaldi, Arch. Biochem. Biophys. 280 (1990) 252–262. [11] C.E. Cooper, P. Nicholls, J.A. Freedman, Biochem. Cell Biol. 69 (1991) 596–607. [12] S. DiMauro, E.A. Schon, N. Engl. J. Med. 348 (2003) 2656–2668. [13] S. DiMauro, E. Bonilla, in: A.G. Engel, C. Franzini-Armstrong (Eds.), Myology Third Edition, McGraw-Hill, New York, 2004, pp. 1623–1662. [14] W.K. Engel, G.G. Cunningham, Neurology 13 (1963) 919–926. [15] R.E. Giles, H. Blanc, H.M. Cann, D.C. Wallace, Proc. Natl. Acad. Sci. USA 77 (1980) 6715–6719. [16] Y.I. Goto, I. Nonaka, S. Horai, Nature 348 (1990) 651–653. [17] Y.I. Goto, I. Nonaka, S. Horai, Biochem. Biophys. Acta 1097 (1991) 238–240. [18] R. Haller et al., J. Clin. Invest. 88 (1991) 1197–1206. [19] N. Hattori, M. Tanaka, T. Ozawa, Y. Mizuno, Ann. Neurol. 30 (1990) 563–571. [20] J.-I. Hayashi, S. Ohta, A. Kikuchi, M. Takemitsu, Y.I. Goto, I. Nonaka, Proc. Natl. Acad. Sci. USA 88 (1991) 10614–10618. [21] S.M. Hsu, L. Raine, H. Fanger, J. Histochem. Cytochem. 29 (1981) 577–580. [22] M.A. Johnson, D.M. Turnbull, D.J. Dick, H.S. Sherratt, J. Neurol. Sci. 60 (1983) 31–53. [23] M.A. Jonhson, B. Kadenbach, M. Droste, S.L. Old, D.M. Turnbull, J. Neurol. Sci. 87 (1988) 75–90. [24] B. Kadenbach, L. Kuhn-Nentwig, U. Buge, Curr. Top. Bioenerg. 15 (1987) 113– 161. [25] B. Kloeckener-Gruissem, J.E. McEwen, R.O. Poyton, Curr. Genet. 12 (1987) 311– 322. [26] C.T. Moraes, S. DiMauro, M. Zeviani, A. Lombes, S. Shanske, A.F. Miranda, H. Nakase, E. Bonilla, L.C. Werneck, S. Servidei, et al., N. Engl. J. Med. 320 (1989) 1293–1299. [27] H. Nakase, C.T. Moraes, R. Rizzuto, A. Lombes, S. DiMauoro, E.A. Schon, Am. J. Hum. Genet. 46 (1990) 418–427. [28] S.A. Neargarder, M.P. Murtagh, B. Wong, I.K. Hill, Cogn. Behav. Neurol. 2 (2007) 83–92. [29] W. Neupert, Ann. Rev. Biochem. 66 (1997) 863–917. [30] L.C. Papadopoulou, C.M. Sue, M. Davidson, K. Tanji, I. Nishino, J. Sadlock, J. Selby, D. Moira Glerum, R. Van Coster, G. Lyon, et al., Nat. Genet. 23 (1999) 333–337. [31] H.J. Pel, A. Tzagoloff, L.A. Grivell, Curr. Genet. 21 (1992) 139–146. [32] D. Pette, Histochem. J. 13 (1981) 319–327. [33] M. Satoh, T. Kuroiwa, Exp. Cell Res. 196 (1991) 137–140. [34] A.M. Seligman, R.M. Rutenburg, Science 113 (1951) 317–320. [35] A.M. Seligman, M.J. Karnovsky, H.L. Wasserkrug, J.S. Hanker, J. Cell Biol. 38 (1968) 1–15. [36] J.M. Shoffner, M.T. Lott, A.M.S. Lezza, P. Seibel, S.W. Ballinger, D.C. Wallace, Cell 61 (1990) 931–937. [37] E.A. Shoubridge, in: S. DiMauro, D.C. Wallace (Eds.), Mitochondrial DNA in Human Pathology, Raven Press, New York, 1993, pp. 109–123. [38] E.A. Shoubridge, Hum. Mol. Genet. 10 (2001) 2277–2284. [39] G. Silvestri, C.T. Moraes, S. Shanske, S.J. Oh, S. DiMauro, Am. J. Hum. Genet. 51 (1992) 1213–1217. [40] J.A. Smeitink, M. Zeviani, D.M. Turnbull, H.T. Jacobs, Cell Metb. 3 (2006) 9–13. [41] M. Sparaco, E.A. Schon, S. DiMauro, E. Bonilla, Brain Pathol. 5 (1995) 125–133. [42] M.J. Szabolics, R. Seigle, S. Shanske, E. Bonilla, S. DiMauro, V. D’Agati, Kidney Int. 45 (1994) 1388–1396. [43] J.W. Taanman, M.D. Burton, M.F. Marusich, N.G. Kennaway, R.A. Capaldi, Biochem. Biophys. Acta 1315 (1996) 199–207. [44] J.W. Taanman, J. Bioenerg. Biomembr. 29 (1997) 151–163. [45] K. Tanji, T.H. Vu, E.A. Schon, S. DiMauro, E. Bonilla, Ann. Neurol. 45 (1999) 377– 383. [46] K. Tanji, J. Gamez, C. Cervera, F. Mearin, A. Ortega, J. Torre, J. Montoya, A.L. Andreu, S. DiMauro, E. Bonilla, Acta Neuropathol. 105 (2003) 69–75. [47] R.W. Taylor, M.A. Birch-Machin, J. Schaefer, L. Taylor, R. Shakir, B.A.C. Ackrell, B. Cochran, L.A. Bindoff, M.J. Jackson, P. Griffiths, et al., Ann. Neurol. 39 (1996) 224–232. [48] V. Tiranti, K. Hoertnagel, R. Carrozo, C. Galimberti, M. Munaro, M. Granatiero, L. Zelante, P. Gasparini, R. Marzella, M. Rocchi, et al., Am. J. Hum. Genet. 63 (1998) 1609–1621.
280
K. Tanji, E. Bonilla / Methods 46 (2008) 274–280
[49] H.J. Tritschler, F. Andreetta, C.T. Moraes, E. Bonilla, E. Arnaudo, M.J. Danon, S. Glass, B.M. Zalaya, E. Vamos, S. Shanske, et al., Neurology 42 (1992) 209–217. [50] T.H. Vu, K. Tanji, H. Valsamis, S. DiMauro, E. Bonilla, Neurology 51 (1998) 1190–1193. [51] D.C. Wallace, Annu. Rev. Biochem. 61 (1992) 1175–1212.
[52] M.T.T. Wong-Riley, Trends Neurosci. 12 (1989) 94–101. [53] M. Zeviani, C.T. Moraes, S. DiMauro, H. Nakase, E. Bonilla, E.A. Schon, L.P. Rowland, Neurology 38 (1988) 1339–1346. [54] M. Zeviani, S. Di Donato, Brain 127 (2004) 2153–2172. [55] Z. Zhu, J. Yao, T. Johns, K. Fu, I. De Bie, C. Macmillan, A.P. Cuthber, R.F. Newbold, J. Wang, M. Chevrette, et al., Nat. Genet. 20 (1998) 337–343.