Abstract
Mitochondrial disorders can lead to a confusing array of symptoms, which frequently makes a diagnosis difficult. Traditional approaches to such diagnoses are based on enzyme activity assays, with further characterization provided by genetic analysis. However, these methods require relatively large sample sizes, are time-consuming, labor-intensive, and show variability between laboratories. Here, we report an immunocytochemical test that makes use of monoclonal antibodies to subunits from each of the oxidative phosphorylation complexes and pyruvate dehydrogenase to aid in the detection of mitochondrial disorders. It can be completed and data analyzed in less than 4 hr. We have used this test to study fibroblast cultures from patients with mitochondrial disorders arising from both mitochondrial DNA and nuclear DNA defects. We have also examined cases of Leigh syndrome arising from different genetic causes. We show that patients can be categorized on the basis of which complexes are affected and whether or not the defect being studied shows a mosaic distribution, an indicator of whether the causal mutation(s) is/are in the mitochondrial or nuclear genome. Immunocytochemical analysis as described here should be considered as an initial screen for mitochondrial disorders by which to direct (and limit) the subsequent enzymatic and genetic tests required to make an unambiguous diagnosis.
Keywords
M
Diagnosis of mitochondrial disorders can be complicated because of a confusing array of symptoms, the complex genetics, and the large number of proteins associated with these disorders. There are an estimated 87 different structural proteins that comprise the OX PHOS complexes alone and many other assembly and import factors required for the proper functioning of the OX PHOS system. Mitochondria also have their own DNA, which encodes 13 subunits of the OX PHOS complexes, 22 tRNAs, and two rRNAs. Therefore, disorders of mitochondrial function can stem from nuclear and/or mitochondrial DNA defects. Disorders stemming from mitochondrial DNA defects have added complexity in that the defect can be heteroplasmic. Although there are just two copies of each nuclear gene, every cell contains hundreds to thousands of copies of mitochondrial DNA and hence mitochondrion-encoded genes (Pon and Schon 2001). Owing to the variable segregation of such mutations, often only a proportion of the mitochondrial genomes within a cell carry a mutation (Leonard and Schapira 2000). Heteroplasmy can vary among individuals, from cell to cell, or among different tissues, and it is the level of mutated genomes within a cell and the particular energy requirements of that cell that affect the severity of the resulting disorder. Further complicating matters are X-linked nuclear genes. The E1α-subunit of the pyruvate dehydrogenase (PDH) complex is encoded on the X-chromosome, and mutations in this gene are the most common causes of PDH deficiency (Lissens et al. 2000). However, because the gene is X-linked, male patients often show more severe symptoms than female patients since they receive only a single copy of the gene (Lissens et al. 2000). Furthermore, females that have inherited one normal copy of the gene and one defective copy of the gene can show mosaicism as a result of × inactivation (Lissens et al. 2000). All of these factors lead to symptoms that can be quite diverse, ranging from relatively mild symptoms, such as exercise intolerance, to neurodegeneration, leading to death. It is this complexity that led Munnich and colleagues (1992) to conclude that a disorder of OX PHOS can give rise to any symptoms in any organ or tissue with any mode of inheritance (Munnich et al. 1992).
Traditional approaches to the diagnosis of mitochondrial disorders have been based on enzymatic activity assays, with further characterization coming from genetic analysis. However, these methods require relatively large sample sizes and are time-consuming, labor-intensive, and are subject to variability from one laboratory to another. We therefore set out to devise a test that could be performed quickly on very small samples and that would aid in the diagnosis of mitochondrial disorders by identifying defective complexes for subsequent analysis by enzymatic and genetic methods. Here we report an immunocytochemical test that can be used to identify defects in which assembly of one or more of the five OX PHOS complexes or PDH is altered, which comprise the vast majority of cases. We demonstrate the utility of this test by studies of fibroblast cultures from patients with a range of mitochondrial disorders including nuclear and mitochondrial DNA defects. Four of the patients had Leigh syndrome, a rapidly progressing neurometabolic disorder that leads to degeneration of the central nervous system. The disease usually strikes young children between the ages of 3 months and 2 years, with death typically occurring within a few years. The genetic causes of Leigh syndrome are quite varied, with defects in complex I, complex II, complex IV, complex V, and PDH all having been reported to give rise to the disease (Dahl 1998; Carrozzo et al. 2001). With our analysis, we were able to distinguish among the various genetic causes of Leigh syndrome. In addition, we studied four patients in whom the mutation occurred in the mtDNA and was heteroplasmic. We show that such patients present as mosaic in the immunocytochemical analysis, with a mixture of normal cells as well as partly reduced or almost fully reduced levels of one or more of the OX PHOS complexes.
Materials and Methods
Cell Lines
MRC5 fibroblasts were obtained from the American Type Culture Collection (Manassas, VA). Primary fibroblast cultures were obtained with consent from needle biopsies performed during diagnostic evaluations. The data presented in Table 1 on genetic cause and symptoms were obtained in the clinics at which the patients first presented. The various cell lines were provided as follows. Patient fibroblast cultures were obtained from the following sources. Patient 1 was from the Nijmegen Center for Mitochondrial Disorders (Nijmegen, The Netherlands). This patient has been described in Triepels et al. (2001). Patient 2 was from Oregon Health Sciences University (Portland, OR). Patients 3 and 4 were from the Hospital for Sick Children (Toronto, Ont, Canada). These two patients have been described previously by Hanson et al. (2001). Patients 5 and 7 were from Columbia University College of Physicians and Surgeons (New York, NY). Patients 6 and 8 were from Oregon Health Sciences University (Portland, OR). All cells were grown in high-glucose Dulbecco's modified Eagle's medium supplemented with 10% bovine calf serum, 50 μg/ml uridine, 110 μg/ml pyruvate, and 10 mM HEPES buffer to maximize growth rates as previously described (Marusich et al. 1997).
Antibodies
All antibodies were from Molecular Probes, Inc. (Eugene, OR). For specifics about each antibody see Table 2.
Immunocytochemistry
Fibroblasts were grown to ∼50% confluency on glass coverslips. The coverslips were removed from medium and washed in PBS before fixation in 4% paraformaldehyde for 20 min at room temperature (RT). Antigen retrieval was performed by submerging the coverslips in 100 mM Tris pH 9.5, 5% urea for 10 min at 95C. The coverslips were rinsed in PBS and placed in 0.1% Triton X-100 for 10 min at RT. Nonspecific antibody binding was blocked by placing the coverslips in 10% normal goat serum for 30 min at RT. Coverslips from each cell line were then exposed for 1 hr at RT to primary antibody cocktails consisting of anti-porin and one of each of the following antibodies: anti-complex I, 30-kD FeS protein; anti-complex II, 30-kD protein; anti-complex III, core 2; anti-complex IV-I, anti-complex IV-IV, anti-complex V, oligomycin sensitivity conferral protein (OSCP) or anti-pyruvate dehydrogenase E1α. Specific details of the isotype and concentration of each antibody used for immunocytochemistry are listed in Table 2. Coverslips were washed three times for 2 min each in PBS. Coverslips were then exposed for 45 min at RT to secondary antibody cocktails consisting of Alexa Fluor 594 goat anti-mouse IgG2b specific antibody (Molecular Probes) and either Alexa Fluor 488 goat anti-mouse IgG1 specific antibody (Molecular Probes, Inc.) (for the complex I, complex III, complex V, and PDH slides) or Alexa Fluor 488 goat anti-mouse IgG2a specific antibody (Molecular Probes) (for the complex II and complex IV slides). Coverslips were washed three times for 2 min in PBS before mounting in Vectashield Mounting Medium with DAPI (Vector Laboratories; Burlingame, CA). Results were imaged using a Nikon Eclipse E800 microscope (Nikon Instrument Group; Melville, NY) with a MicroMAX digital cooled CCD camera system (Princeton Instruments; Trenton, NJ).
Characteristics of the investigated fibroblast cell lines
Results
Patient Cell Lines Include Nuclear and mtDNA Mutations
In this study we compared the immunocytochemical characteristics of fibroblast cultures from eight patients in whom enzymatic deficiencies in either OX PHOS complexes or PDH had been confirmed. In six of these patients the pathogenic mutation had been identified genetically, but in two of the patients the genetic defect had not been identified. Four of the patients had heteroplasmic mitochondrial tRNA mutations giving rise either to mitochondrial encephalomyopathy with lactic acidosis and stroke-like episodes (MELAS) or to myoclonic epilepsy with ragged red fibers (MERRF). The other four patients had all been diagnosed with Leigh syndrome. Two of the patients showed complex IV deficiency, one of whom carried a mutation in an intron of the SURF-1 gene, which codes for a protein involved in complex IV assembly. Although the mutation occurred in an intron, we have previously shown that this cell line has undetectable levels of the Surf-1 protein by immunoblot (Hanson et al. 2001). The other complex IV-deficient patient has an as yet undefined defect that was not in the SURF-1 gene. One patient had an isolated complex I deficiency of unknown genetic origin, and the final patient had PDH deficiency caused by a mutation in the E1α-subunit. Specifics for each patient are provided in Table 1.
A Rapid Immunocytochemical Approach to Identifying Defective OX PHOS Complexes
An important aim of the present study was to develop a screen for mitochondrial diseases that is rapid and that requires only small amounts of sample. For each of the eight patients, fibroblasts were cultured on six or seven glass coverslips. Analysis was performed before the cell cultures were confluent (∼4 × 105 cells per coverslip). Fibroblasts on each coverslip were immunostained with monoclonal antibody cocktails containing an antibody to porin as a control and an antibody to one each of the following proteins: complex I, 30-kD subunit; complex II, 30-kD FeS subunit; complex III, core 2 subunit; complex IV, subunit I; complex IV, subunit IV; complex V, OSCP. For patients 2–4, fibroblasts were also immunostained with an antibody cocktail containing antibodies to porin and to PDH-E1α-subunit. In cases where more than one immunocytochemically reactive monoclonal antibody (MAb) to a particular complex was available, the antibody showing the largest decrease in Western blots of patient cell lysates was chosen (data not shown). For complex IV, two different MAbs were used, because we expected that the mitochondrially encoded subunit I would show the largest decrease in staining for mitochondrial DNA defects but would be present, albeit in reduced amounts, in cell lines from patients with mutations in nuclear genes, particularly those with SURF-1 defects. A key aspect of the test is that the control antibody, porin, has a different isotype from all of the OX PHOS and PDH antibodies, so that each coverslip can be double stained by incubating the cells in a primary antibody cocktail followed by a incubation in an isotype-specific secondary antibody cocktail. Control experiments in which one of the primary antibodies was omitted showed no cross-label between the isotype-specific secondaries (data not shown). The double staining of cells serves as an internal control that helps to correct for intensity differences that may have been seen in single-labeled cells due to such things as variable thickness of cells or varying levels of mitochondrial mass between cells. Each step of the immunocytochemistry was analyzed to find the shortest time required while retaining good staining quality. The entire procedure can be completed and data obtained within 4 hr.
Antibodies used in this study
ICC, immunocytochemistry.
Patients 1–4 had all been diagnosed with Leigh syndrome, but with four different genetic causes of the disease represented among these patients (see Table 2). Immunocytochemical evaluation of these patient cell lines revealed which complex was defective in each case (Figure 1). Patient 1 had almost complete loss of staining for the complex I 30-kD subunit but normal staining for the other complexes. Patient 2 showed a uniform reduction in the levels of PDH-E1α staining, but normal staining of the other complexes. Patients 3 and 4 showed reduction in both complex IV-I and complex IV-IV staining but normal staining of the other complexes. Therefore, the immunocytochemical analysis was able to distinguish among different genetic causes of Leigh syndrome.
The Immunocytochemical Approach Reveals Mosaicism in mtDNA Defects
Patients 5–8 all had tRNA mutations resulting in a phenotype of either MELAS or MERRF. Immunocytochemical staining of these cell lines showed heterogeneous expression of complex I, III, and IV subunits. Some cells showed virtually no immunostaining, others had near normal levels, and still others had intermediate levels of staining. However, the two antibodies that revealed the highest proportion of affected cells were the complex I 30-kD antibody and the complex IV subunit I antibody. This is to be expected because complex I contains the largest number of mitochondrially encoded subunits and complex IV-subunit I is a mitochondrially encoded subunit. In addition, one of the MERRF cell lines (patient 8) showed heterogeneous expression of a complex II subunit. This was surprising because complex II is the only OX PHOS complex that has no mitochondrially encoded subunits. The number of cells showing reduced complex II staining was much smaller than the number of cells showing reduced complex I, III, or IV staining. Nevertheless, this suggests that complex II is downregulated in some cells with severely altered OX PHOS function as a consequence of mutations in other respiratory chain complexes.
Figure 2A shows the results for patient 8. The staining seen for complexes I, III, IV, and V for this patient is representative of the staining also seen in patients 5–7. The more normal pattern of staining of complex II is exemplified by patient 7 (Figure 2B). The heterogeneous staining patterns caused by the heteroplasmic expression of the tRNA mutations contrasts sharply with the uniform reduction of staining seen in cell lines from patients with nuclear DNA defects (Figure 1). In a situation in which the mutation is unknown, this difference is an important diagnostic clue. The mosaic expression of an OX PHOS subunit strongly suggests that a defect is mitochondrially encoded or involves a nucleus-encoded subunit that affects mtDNA copy number, as in mtDNA depletion patients (Marusich et al. 1997). Mosaicism can also result from mutations of an X-linked gene. However, only one polypeptide of the OX PHOS machinery (NDUFA1 of complex I) is on the X chromosome (Zhuchenko et al. 1996), and no patients with a mutation in this gene have thus far been reported. Therefore, with respect to OX PHOS defects, the vast majority of mosaics are a consequence of mitochondrial DNA heteroplasmy or depletion. In addition, when multiple complexes are affected, this points toward a defect in a mitochondrially encoded tRNA, a mitochondrially encoded rRNA, or a large-scale deletion of mitochondrial DNA rather than a point mutation or deletion of a structural gene on the mtDNA.
Discussion
The clinical spectrum of mitochondrial disorders is quite diverse. Although there is growing awareness of the scope of these disorders, patients are often difficult to diagnose because the nature of the defect in individual cases is frequently poorly understood. Not only do the phenotypes of mitochondrial disorders vary greatly but the same phenotype can be caused by many different genetic defects, as is the case for Leigh syndrome. In addition, mutations in somatic nuclear, X-linked nuclear, and mitochondrial genes can all cause mitochondrial disorders, leading to a wide range of inheritance patterns. Therefore, methods that help to quickly identify the source of a defect are clearly needed if better treatment options for mitochondrial disorders are to be developed and to aid in genetic counseling.
Fibroblasts from patients 1–4 were labeled with MAb cocktails containing antibodies to subunits from the indicated mitochondrial complexes (stained in green) and a control protein, porin (stained in red). Shown are the signals for complex I 30-kD subunit, complex II 30-kD FeS subunit, complex III core 2 subunit, complex IV-subunit I, complex V-OSCP, and pyruvate dehydrogenase E1α-subunit. Nuclei were labeled with DAPI. Shown are the merged red, green, and blue signals for each image. Cells that show reduced labeling of a particular mitochondrial complex appear red, and cells with normal levels of a particular complex appear yellow.
Here we show that a simple immunocytochemical test can help to rapidly identify defective OX PHOS and PDH complexes. This test can be completed in less than 4 hr and requires minimal levels of sample, allowing the analysis to be performed shortly after the establishment of the primary cell cultures. Each sample is stained by two MAbs, one a control for mitochondrial mass (porin) and the second to identify the presence or absence of a particular complex.
To demonstrate the utility of this immunocytochemical analysis in pinpointing the nature of mitochondrial defects, we studied four patients who had been diagnosed with Leigh syndrome stemming from four different genetic causes. Our test was able to clearly differentiate among Leigh syndrome arising from a PDH defect, a complex I defect, and complex IV defects. Although Leigh syndrome affecting complex IV arising from a SURF-1 defect and a non-SURF-1 related defect could not be distinguished in this instance, previous reports have indicated that in many cases this may be possible. Williams et al. (2001) reported that complex IV-I staining was unaltered in three different SURF-1 patients while complex IV-IV staining was reduced. In our study, both complex IV-I and complex IV-IV staining were reduced in the SURF-1 patient. However, full genetic analysis has not been carried out on our patient, so it is possible that this patient carries an additional mutation that leads to a more severe immunocytochemical phenotype than has been previously observed in other SURF-1 patients. If, as Williams et al. (2001) have suggested, the complex IV-I staining is unaltered in the majority of SURF-1 patients, then patients showing complex IV defects could be further categorized into those showing a reduction in both complex IV-I and complex IV-IV staining and those showing a reduction only in complex IV-IV staining, with the majority of SURF-1 patients falling into the latter category. Alternatively, it may be possible to incorporate a SURF-1 MAb into the analysis described here, as one becomes available. Because E1 is an X-linked gene whereas SURF-1 is a recessive nuclear gene, knowing the precise cause of Leigh syndrome is important for genetic counseling. Our patient with the PDH mutation was a male and therefore showed a uniform reduction of PDH E1-α staining in all cells. However, females with PDH deficiencies should be mosaics, and this mosaicism is a valuable identifier that an individual is a carrier of a PDH defect even when major symptoms of PDH deficiency are not manifested (Lib et al. 2002).
As we show, an immunocytochemical analysis helps to distinguish between mitochondrial DNA defects and nuclear DNA defects. Two patients diagnosed with MELAS and two patients diagnosed with MERRF each showed a mosaic expression of OX PHOS subunits, as would be expected for a mitochondrial DNA mutation that was heteroplasmic. This type of staining pattern is quite different from that observed in cell lines carrying a somatic nuclear mutation, which shows a uniform reduction of a subunit in all of the cells. A heterogeneous staining pattern, when observed for multiple OX PHOS complexes, is an important indicator of a mitochondrial DNA defect and points to a mutation in a mitochondrially encoded tRNA or rRNA or a large-scale mitochondrial DNA deletion as the likely cause, or to a mitochondrial DNA depletion syndrome, which is much rarer.
In summary, we provide a simple, rapid immunocytochemical test by which to detect defects in the OX PHOS complexes or PDH. These defects account for the majority of mitochondrial disorders. Because this test is quick, relatively easy to perform, and requires very small sample volumes, we believe that it complements the current mitochondrial diagnostic tests available by focusing attention on the proteins that are defective and thereby directing and greatly limiting the number of biochemical and genetic analyses that must be performed for an unequivocal diagnosis.
(