Segregation Patterns and Heteroplasmy Prevalence in Leber’s Hereditary Optic Neuropathy

Abstract

purpose. To investigate the segregation pattern of the mitochondrial DNA mutation at nucleotide position 3460 responsible for Leber’s hereditary optic neuropathy (LHON) and to determine the prevalence of heteroplasmy for the three primary LHON mutations at positions 11778, 3460, and 14484.

methods. Segregation analysis was performed in a cross-sectional study by determining the level of heteroplasmy in blood leukocytes of 23 LHON patients and unaffected carriers from four unrelated families. One family comprising two affected and three unaffected carriers was followed over 5.5 years for a longitudinal segregation analysis of heteroplasmy. The percentage of mutant mtDNA was determined using a novel procedure of fluorescence-based primer extension and restriction fragment length polymorphism analysis. The prevalence of heteroplasmy was assessed by determining the number of genealogically unrelated LHON pedigrees with heteroplasmic maternal family members from the LHON patient records of the Department of Ophthalmology, University of Tübingen, Germany.

results. The authors observed a marked variability in the degree of heteroplasmy levels within each pedigree and a tendency toward a higher mutant allele frequency in offspring generations. Disease expression was correlated with higher levels of mutant mtDNA molecules. Longitudinal analysis revealed no statistically significant decrease in the heteroplasmy level in the family studied but a reduction of 11% and 12% in one affected and one unaffected individual, respectively. In 167 genealogically unrelated LHON families the prevalence of heteroplasmy was 5.6%, 40%, and 36.4% for the 11778, 3460, and 14484 LHON mutations, respectively.

conclusions. Cross-sectional studies of heteroplasmy for the 3460 LHON mutation suggest that the genotype shifts toward a higher mutational load in offspring generations. Long-term decrease in the blood mutant load in single cases indicates negative selection of the mutant allele in the hematopoietic cell system. The prevalence of heteroplasmy varies significantly between the different primary LHON mutations, suggesting genotypical differences in disease expression.

Leber’s hereditary optic neuropathy (LHON) is a degenerative disease of the optic nerve commonly associated with one of three mitochondrial DNA (mtDNA) mutations. It typically affects young men, leading to a persistent loss of central vision, but may occasionally be encountered in conjunction with a spontaneous improvement of vision. Genetically, LHON is characterized by a maternal inheritance pattern, a predominantly homoplasmic genotype of the white blood cell fraction, where all mtDNA molecules carry the mutant allele, and a low penetrance. Approximately 7% of LHON patients positive for the LHON mutation at nucleotide position 11778 harbor a mixture of mutant and wild-type mtDNA, ¹ a condition known as heteroplasmy. Heteroplasmy has been reported in 14% of LHON pedigrees carrying the 11778 mutation. ^{2 3} Despite the small proportion of heteroplasmy, the study of heteroplasmy in LHON families is important, because it may help to elucidate the unusual transmission and segregation pattern of mtDNA mutations. The mitochondrial genotype is dynamic in that the proportion of mutant mtDNAs can change rapidly within a few generations or in a specific tissue over time. This shift of the mitochondrial genotype is mainly accounted for by the process of replicative segregation, which implies random replication and partitioning of mitochondrial genomes into daughter cells. Although the model of replicative segregation is consistent with many observations of mitochondrial inheritance, including LHON mutations, ¹ selective mechanisms may also influence the random shift of the mitochondrial genotype. ^{4 5} A clinical example is the late presentation and progression of neuromuscular mtDNA disease that has been explained by an increase in the mutant allele frequency with advancing age, that is, a positive selection of the mutant mtDNA. ⁶ With regard to LHON mutations, the prevailing opinion is that the wild-type to mutant allele ratio is stable in blood leukocytes of LHON patients or carriers, implying an absence of selection. ^{7 8}  

We here report the results of segregation analysis of the heteroplasmic LHON mutation at nucleotide 3460 in 23 LHON patients and unaffected carriers from four families. In one family, comprising three healthy and two affected individuals carrying the mutation, a longitudinal analysis of the blood heteroplasmy levels was carried out over a period of 5.5 years. The degree of heteroplasmy, that is, the percentage of mutant mtDNA, was determined by means of a novel procedure of fluorescence-based primer extension and restriction fragment length polymorphism (RFLP) analysis (Jacobi FK et al., unpublished results, 2001). 

Furthermore, a review of our repository of LHON family records revealed the prevalence of heteroplasmy in genealogically unrelated LHON families positive for either one of the three common LHON mutations. 

Methods

Patients and Subjects

The study included 10 LHON patients and 13 unaffected carriers from four unrelated German families (Families 24, 124, 179, 192). LHON diagnosis had been confirmed, and the carrier status for the guanine to adenine transition at np3460 was determined in all family members by a loss of the naturally occurring restriction site for BsaHI in the conventional polymerase chain reaction (PCR) and RFLP and by DNA sequencing. 

In one LHON family comprising a mother, her three sons, and a daughter, a longitudinal study was carried out over 5.5 years. Two sons had been diagnosed with LHON by clinical examination in the mid-1980s at the Department of Ophthalmology, University of Freiburg, Germany, whereas the other family members were visually asymptomatic. One son gradually recovered full central vision over a 5-year period (124-II:2), whereas his brother remained legally blind (124-II:1). Blood was first collected for LHON analysis in April 1994 from the offspring and in April 1996 from the mother. At that time, a second blood sample was obtained from individuals 124-II:2 and 124-II:3. Additional blood samples from all family members were collected in December 1999. 

For the assessment of heteroplasmy prevalence, records of patients and unaffected maternal family members from the LHON repository at the Department of Ophthalmology, University of Tübingen were reviewed, and the degree of heteroplasmy was estimated by agarose gel electrophoresis after conventional methods of PCR/RFLP analysis. 

All investigations complied with the regulations set out in the Declaration of Helsinki, and informed consent and fall institutional review board approval were obtained. 

DNA Isolation and PCR

Total DNA was extracted by a standard technique from whole blood samples. ⁹ An mtDNA fragment encompassing np3393 to np4203 was amplified in a 50 μl PCR containing 50 ng of blood leukocyte DNA, 10 pmol of L-Strand and H-Strand primers, 200 μM of each dNTP, 10 mM Tris-HCl (pH 8.6), 50 mM KCl, 0.001% gelatin, and 0.25 U AmpliTaq DNA polymerase (PE Biosystems, Weiterstadt, Germany). The sequence of the L-strand primer (nucleotides 3393–3414) was 5′-CTATATACAACTACGCAAAGGC-3′ and that of the H-strand (nucleotides 4203–4186) 5′-TGCTAGGGTGAGTGGTAG-3′. The cycling protocol was performed an a GeneAmp PCR System 9600 Thermal Cycler (PE Biosystems) using the following reaction conditions: an initial 5 minutes denaturation at 96°C, followed by 30 cycles of 94°C for 1 minute, 53°C for 1 minute, and 72°C for 1 minute, and a final 5-minute extension step at 72°C. PCR products were purified by agarose gel electrophoresis, and DNA was recovered using the QIAquick gel extraction kit (Qiagen, Hilden, Germany). The purified PCR product was eluted from the column in 50 μl of 10 mM Tris-HCl, pH 8.5. DNA from each sample was amplified in at least five independent reactions. 

Fluorescence-Based Primer Extension and RFLP Analysis

The quantitation procedure for the mtDNA mutation at np3460 is described in detail elsewhere (Jacobi FK et al., unpublished results, 2001). This is an improvement on the traditional PCR/RFLP method used in analysis of point mutations, the main modifications being (i) the inclusion of a fluorescence-based primer extension step using Vent (exo-) DNA polymerase to avoid quantitation errors due to heteroduplex formation in PCR, (ii) the quantitation of DNA fragments by polyacrylamide gel electrophoresis and automated spectrofluorometry an a DNA sequencer equipped with GENESCAN analysis software, and (iii) estimation of the percentage of mutant mtDNA using a simple mathematical model based on an in-tube digestion reaction of a homologous control DNA fragment. In contrast to the traditional PCR/RFLP assay, the present approach does not depend an complete endonuclease restriction, Instead, the proportion of uncleaved allele that comigrates with the “truly uncleavable” allele in incomplete digestion is estimated and subtracted from the latter, based on restriction efficiency determined by the homologous control. 

The assay uses a polyacrylamide gel-purified 5′-JOE–labeled nested primer with the sequence 5′-GATCAGAGGATTGAGTAAACGC-3′ (nucleotides 3662–3640 of the H-strand) and the Vent (exo-) DNA polymerase in the primer extension reaction. The primer extension products, a 269-bp and a 281-bp fragment for the sample and the control, respectively, are restricted with the endonuclease BsaHI, which cleaves the wild-type mtDNA into 204- and 65-bp fragments and the control DNA into 216- and 65-bp fragments, respectively. The mutant mtDNA remains uncleaved. All enzymes used in this assay were purchased from New England Biolabs (Schwalbach, Germany). 

The level of heteroplasmy was measured in at least five independent reactions and averaged for each DNA sample using PCR-amplified DNA as a template. 

Multiple Cloning and RFLP Analysis of PCR Products

Gel-extracted PCR products from reactions using the above-mentioned primers and conditions were cloned into a pCR 2.1 vector using the TA Cloning Kit (Invitrogen, Groningen, The Netherlands). Recombinant DNA was isolated from bacterial colonies by a standard alkaline lysis minipreparation, and 50 ng of plasmid DNA were digested overnight with BsaHI. Wild-type and mutant-type clones were differentiated by a four- and three-band pattern, respectively, an agarose gel electrophoresis. Clones showing a mixed digestion pattern were found in some instances and excluded from analysis. A total of 80 colonies were processed for each sample, and the percentage of mutant-type clones was determined. 

Statistical Analysis

The statistical significance of changes in the heteroplasmy ratio over time was determined using the Wilcoxon matched-pairs signed rank test and a representative value from repeated measurements of each dated DNA sample. The difference in heteroplasmy rate between the three primary LHON mutations was tested for statistical significance usingχ ² statistics for 2 × 3 cross-tabulations (both SPSS 8.0 for Windows; SPSS GmbH Software, Munich, Germany). 

Results

The families investigated in this study were selected on the basis of heteroplasmy for the 3460 mutation in either a LHON patient or an unaffected carrier and are therefore representative with regard to the observed transmission and segregation patterns. The significance of the findings is, however, limited by bias introduced through incomplete ascertainment, families with LHON mutations having been identified solely by the presence of at least one affected individual. This, by virtue of the high expression threshold, automatically implies higher heteroplasmy levels. 

Quantitative assessment of heteroplasmy revealed a marked variability of degrees both among siblings and between different generations of a family (Fig. 1) . The analysis of the transmission pattern in the four families studied demonstrated an average increase in the heteroplasmy ratio. The affection status correlated roughly with the level of heteroplasmy, higher levels being found in LHON patients than in their unaffected relatives. However, a blood heteroplasmy level of only 26% was found in one patient of LHON family 192. 

The longitudinal analysis of LHON family 124 revealed no statistically significant (P = 0.1441) decrease in the heteroplasmy level in five members (Fig. 1) . However, in one affected and one unaffected individual we measured a reduction of 11% and 12% mutant mtDNA, respectively, whereas in the remaining family members heteroplasmy was maximally decreased by 3%. The LHON patient (124-II:2) displaying a decrease in the mutant mtDNA level had experienced a severe loss of vision due to an episode of LHON, followed by a gradual recovery of vision over 4 years (Fig. 2) . In a second affected individual in this pedigree (124-II:1), the stability of the high mutant mtDNA level was consistent with the persistent blindness. To rule out some unusual procedural artifacts as a cause of the observed decrease in the level of heteroplasmy, we independently analyzed heteroplasmy by multiple cloning and RFLP analysis, using a common but more labor-intensive method for quantifying heteroplasmy. ¹⁰ The results were consistent with our previous findings (Fig. 1) . 

The review of our records of LHON families positive for either one of the three common LHON mutations revealed a statistically significant difference (P < 0.0001) in the prevalence of heteroplasmy between the 11778 mutation and the 3460 and 14484 mutation, respectively. Of 167 LHON families 7 (5.6%) of 125 had at least one affected or unaffected family member heteroplasmic for the 11778 mutation, whereas the ratios were 8 (40%) of 20 and 8 (36.4%) of 22 for the 3460 and the 14484 mutation, respectively. 

Discussion

We hereby report the results of heteroplasmy analysis of a large number of LHON patients and unaffected individuals carrying the 3460 mtDNA mutation. We used a novel method for quantitating heteroplasmy that has advantages over a number of conventional techniques in terms of accuracy, reliability, and handling (Jacobi FK et al., unpublished results, 2001). As shown by reconstruction experiments, our testing methodology is highly accurate for the assessment of high percentages of the mutant allele in a mixed DNA population because it is based on cleavage of the wild-type allele. Repeated measurements of a constructed DNA mixture of 95% mutant mtDNA yielded an average of 93.25% (range, 90.45–95.57%, n = 5). Subjection of 100% mutant mtDNA sample to the quantitating procedure has consistently yielded a 100% value (data not shown); a deviating result would indicate star activity (i.e., relaxation of restriction endonuclease specificity) and cast doubt on the usefulness of this method. We therefore deem measurements of 98% mutant allele, as in two patients in the present study (in 124-II:1: range, 97.81–98.38%, n = 5), to be reliable and clearly distinct from a homoplasmic mutant genotype. 

The results of the cross-sectional study demonstrate a marked genotypic variability and a poor correlation between the level of heteroplasmy in white blood cells and disease expression, shown most strikingly in one male patient of family 192 with a blood mutation load of only 26%. This is extremely low, in view of the previously estimated threshold for expression of 75% to 80% ¹ and probably reflects the variation of the mutant allele frequency in different tissues arising from random segregation of wild-type and mutant mtDNA during embryogenesis. In a postmortem study of a female LHON patient carrying the 11778 mutation, Howell et al. ¹¹ reported a heteroplasmy level of 33% in the white blood cell fraction compared with 95% and 100% in the optic nerve and retina, respectively. In a study by Black et al. ¹² the mutant mtDNA fraction was only 15% in white blood cells of one patient with the 3460 mutation. 

The increase in the heteroplasmy level in subsequent generations observed in our study has been attributed in the past to an ascertainment bias, according to which only families with a high average heteroplasmy level are studied, which through random segregation creates a genetic shift toward homoplasmy of the mutant allele. ⁶ Interpreting our observed decrease in the level of heteroplasmy over a limited time span in two individuals in terms of a lifelong reduction of mutant allele frequency, the suspected intergenerational shift of heteroplasmy toward higher mutant allele frequency might be explained in part by the fact that in pedigree analyses members of the parental generation are examined at a greater age, when literally “blood stem cells have had more time to sort out the mutant allele.” 

The data from longitudinal segregation analysis of five individuals heteroplasmic for the 3460 mutation fail to show a statistically significant decrease in blood heteroplasmy. However, the low P value obtained from statistical analysis suggests the possibility of a systematic decrease in blood heteroplasmy over time in a larger study population. The marked reduction in mutant load in one affected and one unaffected individual is a remarkable finding, contrasting with the apparently stable heteroplasmy levels in the other family members. It is important in that it suggests a negative selection of the mutant allele in the hematopoietic stem cells or some committed progenitor cells of white blood cells in selected individuals carrying the heteroplasmic mtDNA mutation. It has long been assumed that blood heteroplasmy of LHON mutations segregates stably over an extended period, ^{7 8} but only recently Howell et al. ¹³ reported a reduction in heteroplasmy levels of LHON patients. Earlier studies may have failed to detect negative selection because it is not an ubiquitous event in LHON mutation carriers or because the quantitating procedures used were not sufficiently accurate or the follow-up period was too short. In subject II:3 of family 124, we noted a slight change between April 1994 (30%) and April 1996 (29%), but an 11% decrease in December 1999. This suggests that a decrease in mutant allele frequency may occur at a variable rather than a constant rate, which would be in line with the subtle change observed in other family members. The results presented in the longitudinal study by Howell et al. ¹³ do not rule out the possibility of the 3460 mutant allele’s decrease in heteroplasmy being discontinuous rather than continuous, and they show such a time course in at least 2 of 8 individuals. It may therefore be inappropriate to express mutant load decreases generally as mean changes per year as has previously been done. 

One proposed mechanism for the long-term reduction in mutant load is that individual hematopoietic stem cells homoplasmic for wild-type mtDNA will divide faster than cells with a significant proportion of mutant mtDNA and thus shift the stem cell population toward wild-type over time. ¹³ The advantage of this hypothesis is that it explains the decrease in mutant load without preferential replication, segregation, or partitioning of the wild-type necessarily being involved as others have proposed. ¹⁴ Another possible mechanism could be that preferential mtDNA degradation rather than synthesis causes a decrease in the mutant load over time. In view of the poor repair mechanisms of mtDNA, it has been suggested that preferential degradation has the function of removing defective mtDNA. ¹⁵ Mitochondrial degradation or autophagy is part of the physiological life cycle of mitochondria and is probably mediated by the process of so-called mitochondrial permeability transition, a critical step in the events leading to cell apoptosis, and is induced, among other things, by reactive oxygen species (ROS). Given the increase in ROS formation caused by primary LHON mutations, this pathway could possibly account for long-term reduction in the mutant load of white blood cells. ¹⁶  

One question raised by the present study is what implication the possibility of negative selection in dividing cells could have for the mutant load in retinal ganglion cells in LHON. This issue is particularly intriguing in the light of the spontaneous visual recovery observed in a small number of LHON cases, which occasionally occurs up to several years after visual loss. This indicates that the impairment of oxidative phosphorylation associated with LHON mutations may be sufficient to cause retinal ganglion cell dysfunction but not irreversible cell damage under certain circumstances. A recent study suggests that with more refined testing methods, partial improvement of visual function may be found to be more widespread in patients with the 11778 mutation in LHON than was previously thought. ¹⁷ Though clearly other regulating factors, such as hormonal, environmental, and tissue-specific factors are likely to be responsible for visual recovery, it is not erroneous to assume that a selective mechanism—either replicative, degradative, or both—could be operating in retinal ganglion cells of heteroplasmic LHON cases. Knowledge of the factors regulating mtDNA maintenance at the cellular and mitochondrial level in a cell- and tissue-specific manner is still incomplete. ¹⁸ If there is an inherent randomness in mitochondrial kinetics, genotypic shifts should theoretically evolve at a faster rate in mitotically active tissue than in postmitotic tissue, because mitotic segregation and proliferation of mtDNA segregating units (whether as a single mtDNA molecule or “mitochondrial nucleoid”) at the intercellular level would intensify replicative segregation at the intracellular level. 

Mitochondrial genotypic stability, together with increases and decreases in the mutant allele frequency in tRNA and protein encoding genes, has been reported in several longitudinal studies. ^{6 19 20 21} Kawakami et al. ¹⁹ described a patient with mitochondrial myopathy, encephalopathy, lactic acidosis, and stroke-like episodes (MELAS), in whom muscle weakness improved gradually with age and was paralleled by a decrease in the population of mutant genomes. However, there is a general trend emerging that suggests that pathogenic mtDNA mutations undergo segregational loss in dividing cells, such as those of the hematopoietic cell system, ^{13 20} but not in nondividing cells such as neurons. ^{6 22 23} Thus, despite the many discrepancies observed, the issue of potential selective mechanisms in target tissue of LHON merits further investigation on account of its potential clinical importance. 

The second issue dealt with in the present study is the prevalence of heteroplasmy in genealogically unrelated LHON pedigrees. In reviewing our records of LHON families, we found a significant, sevenfold higher prevalence of heteroplasmy in families carrying the 3460 (40%) or 14484 (36.4%) LHON mutation as opposed to the 11778 LHON mutation (5.6%), which is estimated to account for 50% to 70% of the LHON cases. Although an ascertainment bias among the primary LHON mutations cannot be totally excluded, owing to an element of bias in the selection of the families studied, it is unlikely to have had a profound impact on these estimates because of the randomness of the ascertainment process. The heteroplasmy prevalence for the less common 3460 and 14484 LHON mutations, each accounting for approximately 15% of cases, ²⁴ is not explicitly evident from literature data, whereas the 11778 mutation has repeatedly been reported as being heteroplasmic in approximately 14% of LHON families. ^{2 3} The 3460 and the 14484 mutations have been found to be homoplasmic in some cases and heteroplasmic in others, with the 14484 mutation being more commonly associated with homoplasmy. ^{24 25} Harding et al. ²⁶ reported heteroplasmy in 5 (7.6%) of 66 families carrying the 11778 mutation, in 2 (25%) of 8 families with the 3460 mutation and in 2 (18%) of 11 families with the 14484 mutation (one of which was additionally homoplasmic for the 11778 mutation). The authors also found heteroplasmy in LHON cases with the 3460 and 14484 mutation more frequently than with the 11778 mutation, however, at a low percentage (4%). Howell et al. ²⁷ studied 18 LHON patients and unaffected matrilineal relatives from six families carrying the 3460 mutation and found heteroplasmy in 2 (33%) of 6 families. Two (22%) of 9 LHON patients studied were heteroplasmic for the mutation. In contrast, Huoponen et al. ²⁸ excluded heteroplasmy in 10 LHON patients and healthy matrilineal relatives from three families carrying the 3460 mutation. Oostra et al. ²⁹ reported heteroplasmy in 2 (13%) of 15 compared with 2 (100%) of 2 families carrying the 11778 and 3460 mutations, respectively. The reason for the discrepancy in the reported heteroplasmy rate for the 3460 mutation, and possibly for the less reported 14484 mutation, is probably the small number of families analyzed in most studies, some pedigrees being represented only by single cases. ²⁷ In our review of LHON pedigrees, which contains the largest collective so far described, we corroborate the finding of a higher level of heteroplasmy for the 3460 and 14484 mutations than for 11778. 

This variation in levels of heteroplasmy may relate to both genetic and biochemical factors associated with the primary LHON mutations, such as their transmission and segregation kinetics, their tendency to coexist with certain mtDNA lineages, or the severity of the oxidative phosphorylation defect underlying disease expression. First, little is known about the differences in transmission and segregation kinetics of the various primary LHON mutations. Our study results confirm previous findings associated with the 11778 and 3460 mutations, suggesting that there are substantial intrafamilial variations in mutant allele frequency ^{30 31} as well as a genetic drift toward a higher mutant allele frequency in offspring generations ^{1 6} and a possible segregation of mutant mtDNA under negative selection. ¹³ A notable point of distinction among the primary LHON mutations is the high percentage of singleton cases (58%) with the 11778 mutation, ² which together with the higher rate of homoplasmy could be brought about by a rapid replicative drift toward the mutant mtDNA genome. ³²  

Second, the coexistence of primary LHON mutations and certain mtDNA lineages may indirectly account for the observed difference in the heteroplasmy rate in LHON pedigrees. Although the primary LHON mutations have appeared in a large number of different mtDNA haplogroups, suggesting de novo mutational events in most pedigrees, it is also evident from phylogenetic studies that the primary LHON mutations differ in their preferential association with haplogroup J, which occurs in approximately 9% of the general European population. ^{25 33 34} The latter is partially defined by variants or so-called secondary mutations, such as 13708 G-A and 4216 T-C or, to a lesser degree, the 15257 G-A mutation, which may account for this haplogroup’s predilection toward an increased disease expression of primary LHON mutations. The strongest preferential association with haplogroup J has been demonstrated for the 14484 mutation (80%), followed by the 11778 mutation (37%). ³⁵ The 3460 mutation represents an extreme in that it appears to be distributed randomly among all mtDNA haplogroups. ³⁶ This implies that the 3460 mutation does not require additional, secondary mutations for disease expression and should thus be the most susceptible to selective pressure. ³⁷ Under the assumption that the kinetics of mitochondrial genetics are predominantly determined by random genetic drift, heteroplasmy can be considered an intermediate condition in the transition from one homoplasmic sequence to another, although this has not been observed in all mutational events ^{38 39} or in silent polymorphisms. ⁴⁰ Hence, heteroplasmy also implies that mutations have occurred relatively recently. This is supported by phylogenetic analyses of LHON disease that have revealed high incidences of heteroplasmy (32.1% for the 11778 mutation and 85.7% for the 3460 mutation). ³⁶ Accordingly, the predominance of homoplasmy of the 3460 mutation in some studies ²⁷ may suggest a more distant origin on a generational time scale. 

Third, in addition to differences in the genetic basis of the various primary LHON mutations, there are marked differences in clinical presentation and biochemical defects. The mildness of the clinical course and the rate of spontaneous visual recovery correlate with the type of mutation, increasing from mutation 11778 through 3460 to 14484. A recent functional analysis of the primary LHON mutation has demonstrated that the biochemical defect is more severe in the 3460 than in 11778 and mildest in the 14484 mutation. ¹⁶ This fits well with the theory that heteroplasmy is a more important determinant in the 3460 than in the 11778 LHON mutation, although it does not explain the proportionally high rate of heteroplasmy observed in our LHON families carrying the 14484 mutation. The threshold for disease expression in heteroplasmic cases with the 14484 mutation may be lowered by a synergistic effect of haplotype J and the 14484 mutation on energy metabolism deficiency in LHON. ⁴¹ However, at least with respect to the 11778 mutation, the augmentation of disease expression by secondary LHON mutations has been the subject of much controversy. ^{3 42 43 44}  

In summary, the key findings of the present study, namely the possible operation of negative selective mechanisms on the 3460 mutant allele in dividing cells and the higher prevalence of blood heteroplasmy in LHON families carrying the 3460 or 14484 mutation compared with the 11778 mutation, provide new insights into the kinetics of disease-related mitochondrial genetics and suggest genotypical differences in disease expression. Drawing on current knowledge of mitochondrial genetics related to LHON, we discuss the possible reasons and implications of these observations. 

 Supported by the Fortüne Grant 744-0-0 of the Medical Faculty, University of Tübingen, Germany.

 Submitted for publication August 14, 2000; revised December 22, 2000; accepted January 24, 2001.

 Commercial relationships policy: N.

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked“ advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Matrilineal pedigrees of heteroplasmic 3460 LHON families. ▪ and •, affected individuals; □, male; ○, female;□ /, deceased male; , deceased female. The mean percentage of the mutant allele as determined by primer extension and RFLP analysis is shown below the symbols. Six members of LHON family 124 were assayed for multiple blood samples collected over 5.5 years and one (I:1) for 3.5 years. Age data of family members correspond to the time of first blood sampling. In two individuals with a marked decrease in the heteroplasmy level over time, the SD of cumulative results and the percentage of mutant allele (in parentheses) expressed in the number of mutant-type clones as determined by multiple cloning and RFLP analysis have been added.

Figure 2.

 

Densitometric tracings from heteroplasmy analysis of LHON patient 124-II:2 showing a decrease in the heteroplasmy level over time. Note the mode for calculating the percentage of mutant mtDNA, which uses peak area values corresponding to the fluorescence intensities of DNA bands. Designation of DNA bands is as follows: peak A, proportion of cleaved wild-type allele; peak B, proportion of cleaved control DNA; peak C, uncleaved mutant allele and proportion of uncleaved wild-type allele; and peak D, proportion of uncleaved control DNA. Σ, the sum of peak areas corresponding to a single signal, as determined by the default of analysis parameters.

The authors gratefully acknowledge the statistical assistance of Wolfgang Pabst from Institute for Statistics and Informatics, University of Giessen, Germany.