Association of the mitochondrial 8344 MERRF mutation with maternally inherited spinocerebellar degeneration and Leigh disease

MERRF, myoclonic epilepsy with ragged-red fibers, is one of the mitochondrial encephalopathies, a clinically diverse group of disorders that also includes Kearns-Sayre syndrome (KSS) and MELAS (mitochondrial encephalopathy, lactic acidosis, and stroke-like episodes). ^[1,2] In many cases, these disorders involve a mutation within the mitochondrial genome as the primary etiologic factor. MERRF is usually causatively associated with a mutation at nucleotide 8344 in the mitochondrial tRNA^lys gene, although there is one report of a matrilineal MERRF pedigree that carried a mutation in the same gene at nucleotide 8356. ^[3]

We report here the clinical and molecular genetic analysis of a family who carry the 8344 MERRF mutation, but who exhibit an array of neurologic disorders previously unassociated with this mitochondrial mutation. It was initially thought that the inheritance was autosomal dominant, ^[4] but the clinical abnormalities have been maternally transmitted through three generations, thereby alerting us to the possibility of mitochondrial inheritance.

Clinical findings.

The pedigree of this family displays a multigenerational neurologic disorder that is striking in terms of both phenotypic variability and age of onset (Figure 1).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1. Pedigree of a family with suspected maternal transmission of diverse neurologic abnormalities including Leigh disease and spinocerebellar degeneration.

Methods.

Molecular genetic analysis of the mitochondrial genome from family members was carried out with the strategy used routinely in our studies. ^[5] All procedures were conducted under a protocol approved by an institutional review board and with informed consent of the patients. Briefly, total DNA was isolated from either primary fibroblast cultures or platelets, and this DNA was then used for PCR amplification of short fragments (approximately 300 bp) of the mitochondrial genome with pairs of primers that contained Sau3A restriction endonuclease sites. After purification and Sau3A digestion of the amplified DNA fragments, they were cloned into BamH1-cleaved M13 sequencing vectors and transformed into an Escherichia coli host strain. Recombinant plasmids were used for dideoxy chain termination nucleotide sequencing. A cumulative span of about 4,000 bp of the mitochondrial genome coding region has been sequenced for members of this maternal lineage. The regions analyzed spanned the sites of previously identified pathogenetic mutations; these included the entire length of the tRNA^leu, tRNA^lys, cytochrome b, and ATPase6 genes. Finally, for each mtDNA region, the sequence from at least two maternal relatives was determined.

Mitochondria were prepared as previously described from platelets obtained by plateletpheresis. ^[6] The concentration of cytochrome b was determined from dithionite-reduced minus oxidized spectra in a Shimadzu UV-3000 spectrophotometer at 25 degrees C after solubilization with n-dodecyl-beta-D-maltoside (Boehringer-Mannheim).

Results.

Because the maternal transmission of the neurologic abnormalities in this family (Figure 1) suggested mitochondrial inheritance, rather than nuclear, we investigated the possibility of a pathogenetic mutation in the mitochondrial DNA. The DNA sequencing analysis showed that the three members of this matrilineal pedigree who were tested carry the 8344 MERRF mutation (Table 1). The mutation was heteroplasmic in all three family members and in DNA from both platelets and fibroblasts. As anticipated, the proportion of the 8344 mutant allele was highest in the most severely affected family member (III-3). Previous studies have indicated a correlation between the proportion of the 8344 mutation and the severity of clinical symptoms. ^[7,8] However, this trend is not absolute, the proportion of the mutant allele in platelet mtDNA was higher in the minimally affected brother (III-2) than in his more severely affected mother (II-2). The proportion of the 8344 mutation should have been very high in the tissues of the two sons, IV-3 and IV-4, because they were the most severely affected. Unfortunately, no tissue was available to test this prediction.

Other than the MERRF mutation at nucleotide 8344, none of the other changes from the Cambridge standard sequence were likely to be pathogenetic because they were phenotypically silent or were polymorphisms that had been detected previously in normal controls (data not shown). For example, the nucleotide sequence of the entire cytochrome b gene was determined for three family members. The only sequence difference from the Cambridge standard was an AT[arrow right]GC transition at nucleotide 15326, a polymorphism observed in more than 95% of the subjects whom we have analyzed.

Prior to the mitochondrial genetic analysis, biochemical assays were carried out with platelet mitochondria to ascertain whether the clinical symptoms in this matrilineal lineage were associated with respiratory chain dysfunction. In an attempt to minimize spurious results arising from the relatively harsh conditions required for isolation of platelet mitochondria, and to normalize activity to an intrinsic mitochondrial membrane redox protein, alpha-glycerophosphate:cytochrome c oxidoreductase flux was titrated as a function of the concentration of antimycin or myxothiazol ^[6] respiratory chain inhibitors that bind to the protonmotive cytochrome b subunit of Complex III. ^[9] This assay measures the ubiquinone-mediated transfer of reducing equivalents from the flavin-linked mitochondrial alpha-glycerophosphate dehydrogenase to Complex III (ubiquinol:cytochrome c oxidoreductase). The flux values were then normalized to the concentration of cytochrome b. The results of these experiments (Figure 2) show that there is a slight increase in both antimycin and myxothiazol resistance in mitochondria isolated from minimally affected (III-2) and severely affected (III-3) family members, but not in those from the clinically normal daughter (IV-2) of III-2.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2. Titration curves of alpha-glycerophosphate-cytochrome c oxidoreductase activity as a function of antimycin (open symbols) and myxothiazol (filled symbols). (A) Results for mitochondria from a single control subject (circles) and for the pooled results from assays of a group (n = 5) of normal subjects (squares). (B) MERRF family member III-2. (C) MERRF family member III-3. (D) MERRF family member IV-2. It should be noted that family member IV-2 (D) is the clinically normal daughter of male family member III-2 (C) and that she does not carry the 8344 MERRF mutation (data not shown).

Discussion.

This family illustrates two important results that often confound the clinical, biochemical, and molecular genetic analysis of the mitochondrial encephalomyopathies. First, maternal transmission of the disorder was problematic because of the small size of the pedigree and the marked variability in the clinical symptoms. Second, although this family has some of the abnormalities that typify MERRF patients, including lipomas, other features were atypical. For example, Leigh syndrome is an infrequent occurrence in patients with the 8344 MERRF mutation. ^[10] Furthermore, to our knowledge, this is the first instance of an association of spinocerebellar degeneration, in the absence of pronounced myoclonus and seizures, with this pathogenetic mitochondrial DNA mutation. Greenwood et al ^[11] reported a small family with both Leigh disease and spinocerebellar degeneration among the members. It would be of interest to screen this family for mitochondrial mutations, particularly those that cause MERRF. The neurologic deficits in that family ^[11] were characterized as autosomal dominant, but the transmission is also compatible with maternal inheritance, and there were elevated CSF lactate/pyruvate ratios. There is also a recent report of a patient with Ekbom's syndrome. ^[12] Both ataxia and neuropathy were present in association with the 8344 mutation, but there was no spontaneous or evoked myoclonus in this patient.

No biochemical or mtDNA sequence data are available for patient II-2, but we hypothesize that his late-onset, atypical Charcot-Marie-Tooth disease may represent yet another phenotype associated with the 8344 MERRF mutation. Patient III-2, who carries the 8344 mutation at an unexpectedly high level in blood and fibroblasts and who apparently has a respiratory chain defect (Figure 2), is mildly affected at the present time. Whether or not he will develop a similar neurodegenerative condition later in life is unknown, but the present studies support the possibility of long latencies of mitochondrial neurodegenerative diseases.

The biochemical assays of respiratory chain function were illustrative of the difficulties encountered in the analysis of the mitochondrial encephalopathies, and they underscore the value of integrated genetic, biochemical, and clinical studies. At first consideration, the increase in myxothiazol and antimycin resistance (Figure 2) suggested that, in the members of this maternal lineage, there was a reduction in the binding of these inhibitors to the mitochondrially encoded cytochrome b subunit of Complex III. However, the results of nucleotide sequencing revealed unequivocally that there were no mutations in the mitochondrial cytochrome b gene, thus obviating a direct explanation for the increased inhibitor resistance. On the basis of previous studies with yeast mutants that showed increased resistance to the Complex III inhibitor diuron in assays of NADH-cytochrome c oxidoreductase, ^[13] an explanation can be put forward that resolves this apparent paradox. In brief, the alphaglycerophosphate:cytochrome c oxidoreductase assay involves two redox enzyme complexes "connected" by a ubiquinone pool. Furthermore, the overall rate of electron transfer is limited by the rate at which the dehydrogenase complex reduces the quinone intermediates (V_red), not by the rate at which Complex III reoxidizes these intermediates (V_ox). As one consequence of the complex kinetics of this assay, a decrease in the flux rate through Complex III will increase the measured antimycin and myxothiazol resistance (figure 4 of ^[13]). The results in Figure 2, therefore, most likely reflect Complex III dysfunction in platelet mitochondria, although more detailed studies will be required to confirm this conclusion. Interestingly, other studies have also obtained evidence for mitochondrial Complex III deficiencies in MERRF patients. ^[14]

The prevailing explanation for the marked clinical variability among the mitochondrial encephalomyopathies (MERRF, MELAS, KSS) is differential, local neurodegeneration within the CNS due to mitochondrial respiratory chain dysfunction. ^[1,2] The pathophysiology is a complex function of several determinant factors. On the one hand, there will be genetic factors that include the phenotype of the mitochondrial mutation and its genetic "load" (eg, the proportion of the mutant allele when heteroplasmic) in different regions of the CNS. In addition, however, there are also important physiologic and developmental factors that result in a "selective regional vulnerability within the nervous system." ^[15] In the example of MERRF, the pathology is relatively diffuse, but there is neuronal loss in the cerebellar dentate nucleus and brainstem. Although cerebral cortical pathologic lesions are absent, Thompson et al ^[16] reported that, irrespective of the severity of myoclonus in 8344 MERRF patients, there is a common pattern of abnormal somatosensory evoked potentials consistent with cortical reflex myoclonus. Future studies of the mutant allele frequency in local areas of the CNS, particularly of the cortex, in these patients may help to clarify the neuropathology and, from this knowledge, to develop more effective therapeutic strategies. The occurrence of atypical 8344 MERRF families, as exemplified by the present report, is a further indication that mitochondrial encephalomyopathies have probably been underdiagnosed when purely clinical criteria were used. Therefore, molecular genetic screening of a broad range of patients is amply justified. In particular, the 8344 mutation should be considered when evaluating patients with Leigh syndrome, unusual spinocerebellar degenerations, or unusual neuropathies.

Acknowledgments

The technical assistance of Ms. Toyya Kinsey in the nucleotide sequencing is gratefully acknowledged. We thank the patients who participated in this research.