- Top of page
- Patients and methods
Summary. Mitochondria (mt) play an important role in both apoptosis and haem synthesis. The present study was conducted to determine DNA mutations in mitochondrial encoded cytochrome c-oxidase I and II genes. Bone marrow (BM) biopsy and aspirate, peripheral blood (PB) and buccal smear samples were collected from 20 myelodysplastic syndrome (MDS) patients and 10 age-matched controls. Cytochrome c-oxidase I (CO I) and II (CO II) genes were amplified using polymerase chain reaction and sequenced. CO I mutations were found in 13/20 MDS patients and the CO II gene in 2/10 normal and 12/20 MDS samples, irrespective of MDS subtype. Mutations were substitutional, deletional and insertional. CO I mutations were most common at nucleotide positions 7264 (25%) and 7289 (15%), and CO II mutations were most common at nucleotide positions 7595 (40%) and 7594 (30%), suggesting the presence of potential ‘hot-spots’. Mutations were not found in buccal smears of MDS patients and were significantly higher in MDS samples compared with age-matched controls in all cell fractions (P < 0·05), with bone marrow high-density fraction (BMHDF) showing a higher mutation rate than other fractions (P < 0·05). MDS marrows showed higher levels of apoptosis than normal controls (P < 0·05), and apoptosis in BMHDF was directly related to cytochrome c-oxidase I gene mutations (P < 0·05). Electron microscopy revealed apoptosis affecting all haematopoietic lineages with highly abnormal, iron-laden mitochondria. These results suggest a role for mt-DNA mutations in the excessive apoptosis and resulting cytopenias of MDS patients.
Mitochondria (mt) were once independently living bacteria that have become a permanent part of the eukaryotic cell through an ancient symbiotic contract (Margulis & Schwartz, 1999). They continue to exert some level of autonomy, however, as they maintain 5–10 copies of their own 16 500 base pair circular DNA (mt-DNA), replicate independently of nuclear DNA, and code for 13 subunits of the respiratory chain multienzyme complex, 22 transfer RNAs and 2 ribosomal RNAs (Anderson et al, 1981). As every cell has several mitochondria, and every mitochondrion has multiple copies of mt-DNA, this amounts to thousands of copies of mt-DNA per cell. In addition to coding for 13/80 subunits of the respiratory chain multienzyme complex essential for the electron transport system involved in haem synthesis, a role for mitochondria has also been identified in the initiation of cellular apoptosis (Petit et al, 1995; Green & Reed, 1998; Hunault-Berger et al, 1999). Given the appreciation of their significant role in maintaining cellular homeostasis, it is no surprise that mitochondrial mutations are associated with a number of human diseases. These include both congenital diseases such as Pearson's syndrome (Rotig et al, 1989) characterized by a refractory sideroblastic anaemia and bone marrow dysplasia, as well as acquired ones such as various cancers (colorectal, head and neck, lung and primary bladder cancers) (Habano et al, 1999; Fliss et al, 2000) and acquired idiopathic sideroblastic anaemia (AISA) (Gatterman et al, 1992, 1993, 1996, 1997; Gatterman, 1999; Matthes et al, 2000). Because of the contribution of mitochondria in both haem synthesis and apoptosis, there has been an increasing interest in investigating the role of mt-DNA mutations in myelodysplastic syndromes (MDS) in which iron-laden, swollen and bizarre mitochondria are associated with an acquired anaemia, variable cytopenia and excessive apoptosis of haematopoietic cells (Jacobs, 1986; Gatterman, 1999).
The MDS are a group of clonal, heterogeneous haematopoietic disorders presenting with anaemia, a variable cytopenia and bone marrow dysplasia (Jacobs & Bowen, 1992). Excessive intramedullary apoptosis of haematopoietic cells is postulated to account for the variable cytopenias of MDS (Clark & Lampert, 1990; Mundle et al, 1995; Raza et al, 1995; Shetty et al, 1996). However, the precise mechanism by which haem synthesis is impaired leading to anaemia and accumulation of iron in the mitochondria of erythroid precursors is not clear. Normally, iron is imported into the mitochondria of erythropoietic cells, combines with protoporphyrin IX to form haem and subsequently leaves the mitochondria as haem iron (Jacobs, 1986; Gatterman, 1999). In MDS, utilization of iron is obviously disturbed, so that iron accumulates in the mitochondrial matrix and gives the appearance of ringed sideroblasts (Jacobs, 1986; Jacobs & Bowen, 1992). A possible explanation for this defect is that iron may not be in the right chemical form (Gatterman et al, 1992, 1993, 1997; Matthes et al, 2000). Iron deposits in the sideroblastic mitochondria are in the ferric (Fe3+) state, while only ferrous iron can be used for incorporation into protoporphyrin IX by ferrochelatase (Porra & Jones, 1963). As ferrous iron is not stable under aerobic conditions, it is necessary for erythropoietic cells to have an enzyme system that can maintain a supply of Fe2+ as substrate for ferrochelatase. It was shown that the electrons for the conversion of ferric iron into ferrous iron are provided by the mitochondrial respiratory chains (Flatmark & Romslo, 1975), in particular complexes III (cytochrome b) and IV (cytochrome c-oxidase) (Gatterman, 1999). It has been hypothesized that malfunction of one or of several of the enzyme complexes of the mitochondrial encoded respiratory chain genes may contribute to the inefficient reduction of iron in MDS and that this malfunction could be caused by mutations of nuclear DNA or of mitochondrial DNA, both of which contribute to the assembly of respiratory chain complexes (Jacobs, 1986; Jacobs & Bowen, 1992; Gatterman, 1999). Previous studies have revealed the presence of mt-DNA mutations affecting cytochrome c-oxidase I gene in two patients with AISA (Gatterman et al, 1997), but the other varieties of MDS have not been investigated. The current study was conducted to not only detect mutations in mt-DNA encoded cytochrome c-oxidase I and II genes, but also to explore the relevance of these mutations in the excessive apoptosis of haematopoietic cells in MDS. Our results indicate an unexpectedly high rate of mutations among 16/20 MDS patients examined.
An additional unique feature of this study is related to detailed morphological examination of mitochondria in bone marrow (BM) biopsy tissue of MDS patients. In the past, electron microscopic (EM) studies could not be performed on BM biopsies because the decalcification process destroys ultrastructural detail (Cohen et al, 1997). A novel decortication technique was used to mechanically tease apart the pieces of bone from the BM biopsy samples prior to embedding the tissue for EM studies (Weiss, 1976). By eliminating the chemical decalcification step, it was possible to examine the morphological details of iron-laden, swollen and bizarrely shaped mitochondria present in both apoptotic and non-apoptotic dysplastic cells of MDS patients. In addition, both stromal and parenchymal MDS cells could be investigated precisely as they exist in vivo with well-preserved geographical relationships. These studies have resulted in the recognition of novel, previously unrecognized ultrastructural details of mitochondria as well as apoptotic stromal and parenchymal cells in MDS bone marrows. To our knowledge, the present study is the first to investigate simultaneously the morphological ultrastructural details of mitochondria in MDS BM biopsies, and to examine mitochondrial DNA for the presence of mutations in cytochrome c-oxidase genes. Both aspects of the study shed new light on the role of mitochondria in the pathology and perhaps aetiology of cytopenias in patients with myelodysplastic syndromes.
- Top of page
- Patients and methods
We report an exceptionally high rate of DNA mutations in the mitochondrial encoded cytochrome c-oxidase I and II genes in 16/20 MDS patients studied. Gatterman et al (1997) described mitochondrial DNA mutations in cytochrome c-oxidase I in acquired idiopathic sideroblastic anaemia, but to our knowledge this is the first report of mt-DNA mutations affecting cytochrome c-oxidase gene II in all subcategories of MDS patients. Several features of this study were unique and are summarized below.
• mt-DNA mutations were only found in the haematopoietic cells of MDS patients and were not found in matched buccal smear samples.
• Mutations were insertional, deletional, substitutional or a combination.
• Cells obtained from the bone marrow high-density fraction were the most likely to contain mt-DNA mutations. These cells represent the more mature fraction of the sample, and have already been shown to contain large numbers of apoptotic cells in MDS patients. Cells containing mt-DNA mutations, albeit with a lesser frequency, were also found in low-density BM fraction and both low- and high-density fractions of peripheral blood compartments.
• mt-DNA mutations present in the bone marrow were not necessarily the same as, or even present, in the peripheral blood and vice versa.
• Both cytochrome c-oxidase gene I (13/20 patients) and cytochrome c-oxidase gene II (12/20 patients) were affected equally, and the majority of mutations were substitutional.
• Certain frequently involved ‘hot spots’ were identified for cytochrome c-oxidase gene I mutations affecting nucleotide positions 7264 (25%) and 7289 (15%), and cytochrome c-oxidase gene II mutations affecting nucleotide positions 7595 (40%) and 7594 (30%). The mutation at np7595 resulted in a substitution of G to C changing the amino acid from alanine to proline, and at np7594, between T to G changing the amino acid from histamine to glutamine. Seven MDS patients had the substitutional mutation at np7595 and six of these same patients also had the substitutional mutation at np7594. Although these mutations were within the primer binding sites, they could not be artefactual because we sequenced both strands to detect these mutations.
• Patients belonging to all FAB categories of MDS were found to have mt-DNA mutations, but they were most consistently present in RARS patients (4/4). No mt-DNA mutation was uniquely related to a specific FAB category.
• There was a significant association between mitochondrial encoded cytochrome c-oxidase gene I mutations in the high-density fraction of mononuclear cells and the level of apoptosis in the BM biopsies (P < 0·05).
• None of the mitochondrial mutations described here are polymorphic as they have not been recorded in the human mitochondrial genome database (MITOMAP, 2001). This does not exclude the presence of unknown polymorphisms.
• Electron microscopic studies conducted on decorticated bone marrow biopsy tissue enabled the first investigation of ultrastructural detail of mitochondria in both apoptotic and non-apoptotic cells belonging to haematopoietic stromal and parenchymal lineages. Normal as well as bizarre, swollen and iron-laden mitochondria were clearly identified.
The first important point is that mt-DNA mutations identified in this study did not represent previously recorded polymorphisms. Rather, they appeared to be acquired mutations which were unique to the haematopoietic cells of MDS patients as they were absent in buccal smears obtained from the same patients. Mutations were more common in cytochrome c-oxidase gene I than in gene II. Of the 18 instances of mt-DNA mutations identified in gene I, there were nine substitutions, the most common being three instances of substitution between C–A. In cytochrome c-oxidase gene II, the most common type of mutations were substitutions present in 25 instances, followed by five insertions. Certain nucleotide positions were frequently affected in gene II and two of these at np7594 and np7595 can definitely be considered as ‘hot spots’ for mt-DNA mutations. These resulted in an amino acid change from alanine to proline when the G–C substitution occurred at np7595, and from histamine to glutamine when the T–G substitution occurred at np7594.
Some of these mutations were only found in cells obtained from either blood or the marrow, but not in both samples of the same patient. Even within a haematopoietic compartment, mutations were frequently present only in a fraction of low- versus high-density cells. At first glance, this appears counter intuitive. However, there are several plausible explanations for this observation. First, this cannot represent a sampling error as no mutations were found in normal cells of the same patients (buccal smears). Furthermore, many of these were recurrent rather than random mutations both across patients and across various samples within the same patients, making it less likely to be a methodological issue. Second, it must be remembered that the polymerase chain reaction (PCR) method used for gene amplification followed by sequencing is likely to detect the most predominant cell population within a given sample. The marrow contains more immature cells than the blood, and if the mt-DNA predominated in the mature cells, then it would only be detectable in the blood and vice versa. The same applies to various fractions of mononuclear cells separated on a density gradient, where the more mature as well as apoptotic cells being heavier sink to the bottom and are recovered in the high-density fraction. Again, if the majority of mutated mitochondria are present in the more mature or dying cells, then only cells obtained from the high-density fraction of blood or marrow will show the presence of the mutation.
Haematopoiesis in patients with myelodysplastic syndromes has repeatedly been shown to be monoclonal (Anan et al, 1995). An important question relates to the biological versus aetiological significance of mt-DNA mutations if they are only present in a fraction of cells, be they mature or immature, in an otherwise monoclonal population. Mutations could not possibly be the cause of the disease if they are absent from a subpopulation of cells which have otherwise descended from the transformed MDS parent cell as it would mean that the transformed MDS parent cell is capable of generating both cells with mutated and wild-type mt-DNA. In other words, the presence of mt-DNA mutations in a subset of cells would represent derivative populations and evolving clones, and thereby reflect the biology and not the aetiology. Given the unique multiplicity of mt-DNA and its peculiar independent replication, however, it is still actually possible for such mutations to be the initiating event despite being manifested in a subset of cells in a monoclonal population. Further studies are needed to understand the role of mitochondrial DNA mutations in clonal selection by using the phenotypically well defined early haematopoietic stem cells.
The increased level of apoptosis observed in MDS subjects in the present study is in agreement with earlier reports (Mundle et al, 1995; Raza et al, 1995; Shetty et al, 1996). Recently, Shetty et al (2000) reported that the highest percentage of apoptotic cells are recovered from the high-density fraction of MDS BM aspirates. The current observation that the highest percentage of mitochondrial encoded cytochrome c-oxidase I gene mutations are present in the high-density fraction of MDS BM aspirates suggests that there may be a relationship between the mt-DNA mutations and the level of apoptosis. However, the hypothesis that the two are related must remain speculative at best. Three main mechanisms have been proposed to implicate the role of mitochondria in apoptosis. The earliest mechanism involved the disruption of the electron transport system (ETS). For instance, gamma irradiation induces apoptosis in thymocytes by disrupting the electron transport chain, probably at the cytochrome b-c1/cytochrome c step (Scalfe, 1966). The consequence of the loss of ETS is a drop in ATP production. Such a drop in ATP production has been noted in apoptosis (Bossy-Wetzel et al, 1998). The second mechanism proposed was the activation of caspases by the cytochrome ‘c’ released from the mitochondria. Cytosolic cytochrome ‘c’ forms an essential part of the vertebrate ‘apoptosome’ which is composed of cytochrome c, Apaf 1 and procaspase-9 (Li et al, 1997). The result is activation of caspase-9, which then processes and activates other caspases to orchestrate the biochemical execution of cells. In addition to cytochrome ‘c’, mitochondria are known to release other pro-apoptotic mediators such as apoptosis-inducing factor (AIF) (Susan et al, 1999). The third mechanism proposed is the increased production of reactive oxygen species by mitochondria during apoptosis (Thress et al, 1999). Recently, Matthes et al (2000) showed an increased apoptosis owing to decreased mitochondrial membrane potential in acquired sideroblastic anaemia. Although it is not known which of these mechanisms predominates in MDS, it is clear that mitochondria constitute one of the key components in the apoptotic process and the defect in complex III and complex IV of respiratory chain might contribute towards an excessive apoptosis.
Another hallmark of the disease is that the bone marrow of MDS patients contain erythroblasts with a perinuclear ring of iron-laden mitochondria as illustrated in Figs 4A and C. Similar observations were made earlier (Cohen et al, 1997). The intramitochondrial iron deposits contain iron in the ferric (Fe3+) state and this ferric iron could be responsible for an increased production of reactive oxygen species via the Fenton reaction (Halliwell & Gutteridge, 1984). However, ferrochelatase, the enzyme that catalyses haem synthesis, is known to use iron in the ferrous form. The electrons needed for the reduction of Fe3+ to Fe2+ are physiologically generated through the electron transport system (Flatmark & Romslo, 1975). Mutations in cytochrome c-oxidase I and II genes might contribute towards the dysregulation of ETS such that the iron can no longer be converted into the right chemical form for haem synthesis. The result is a deposition of the under-utilized iron in the mitochondrial matrix of erythroid progenitors, giving rise to the pathognomonic ringed sideroblasts.
The underlying cause for increased mutations in mitochondrial encoded cytochrome c-oxidase I and II genes remains unknown. However, it has been reported that exogenous factors such as toxins and viral infections can cause mutations in mitochondria. It has also been reported that one of the known human retroviruses, the human immunodeficiency virus (HIV), causes mitochondrial dysfunction by selectively concentrating in mitochondria of infected cells (Somasundaran et al, 1994). Both toxins and viruses (Raza & Preisler, 2000; Reddy et al, 2000) have been proposed as being aetiological agents in the development of myelodysplastic syndromes, and may exert their pathological effects through a disruption of mitochondrial function. Further studies are needed to explore the possible role of mitochondria in the aetiology of MDS.