The lifetime accumulation of low-abundance, somatic mtDNA re-arrangements (sublimons) has been proposed as a potential contributor to aging, and also to diseases such as cardiomyopathy or coronary heart disease. Tissue-specific sublimons, varying in abundance by three orders of magnitude between individuals, have recently been observed in myocardium of control subjects. To study the relationship between myocardial sublimon levels and various types of cardiac disease and aging, we applied a semi-quantitative fluorescent PCR assay on cellular DNA extracted from left ventricle in a series of 300 well characterized male victims of sudden death up to age 70 (Helsinki Sudden Death Study). The most prevalent classes of sublimons were present at <0.1 to 91 copies per cell, but their abundance did not correlate with any cardiac disease phenotype. In multiple regression analyses age (β = 0.43, P < 0.0001) and smoking (β = 0.25, P = 0.012) were the only independent factors found to correlate with sublimon levels. Thus, sublimons are inferred to accumulate with age in myocardium of a subset of individuals, but to levels where they do not appear to have any phenotypic effects during a typical life span. We propose that, instead of being a causal factor in cardiac aging, sublimons co-exist with wild-type mtDNA in an equilibrium which is regulated by as yet unknown mechanisms.
Received October 11, 2001; Revised and Accepted November 27, 2001.
INTRODUCTION
Rearrangements of mitochondrial DNA (mtDNA), in the form of deletions and partial duplications, are found in a variety of pathological states affecting skeletal muscle, heart, the nervous system and other organs (1). These pathological mtDNA re-arrangements always co-exist with wild-type mtDNA in variable proportions, a situation denoted as heteroplasmy. Single, clonally expanded mtDNA rearrangements are found in the affected tissues of patients with sporadic mitochondrial (encephalo)myopathy, who may show varying degrees of clinical severity ranging from progressive external ophthalmoplegia (PEO) to Kearns–Sayre syndrome (KSS) and also Pearson’s bone-marrow syndrome. Heart conduction abnormalities are often seen in KSS, including various branch blocks, atrioventricular or complete heart block (2,3).
Multiple mtDNA deletions can be found in cases of mitochondrial disorders showing autosomal inheritance, such as autosomal dominant (ad) PEO (4), or the recessively inherited mitochondrial neurogastrointestinal encephalomyopathy (MNGIE) syndrome (5). PEO with autosomal recessive inheritance and multiple mtDNA deletions is manifested as severe cardiomyopathy of infancy (6). In all these disorders, overt pathology is typically associated with a considerable mutational load, 10–50% or more of the total mtDNA being represented by the rearranged molecules in the relevant tissue(s), including heart.
Based on results obtained by PCR, it has been suggested that deleted mtDNAs accumulate during the course of aging in significant amounts, thus contributing to cellular senescence (7,8). In addition, some forms of idiopathic cardiomyopathies have been proposed to be associated with mtDNA deletions present in myocardium at low levels, comparable with those detected during normal aging (9–14). An association between deleted mtDNAs and coronary artery disease (CAD) has also been suggested (15). These reports are all based on studies of relatively few individuals. In recent reports, we and others have used long PCR to demonstrate the presence of rearranged mtDNAs even in healthy individuals (16–18), calling into question the significance of such molecules in cases of diseases of the myocardium or coronary arteries. A meaningful disease association of these low abundance rearranged mtDNAs has also been questioned by the lack of evidence that they are functionally deleterious at the levels observed. Long PCR can easily detect deletions present at only one or a few copies per cell (17), i.e. <0.1% of mtDNA.
Recently, we have characterized rearranged mtDNAs in cells and tissues of healthy persons, and detected a specific set of these molecules in control adult heart (18). These molecules were termed sublimons, by analogy with plant mtDNA molecules having similar properties. The term was originally applied to low-abundance mtDNA molecules present in healthy plants, that are structurally indistinguishable from those that accumulate to high levels in plants affected by mitochondrial cytopathies, such as cytoplasmic male sterility (19). In humans, the term refers to low-abundance, rearranged mtDNAs, that are structurally similar to the multiple, rearranged mtDNAs found in patients with mitochondrial disorders such as adPEO (4) or MNGIE (5). The rearrangement break points of human myocardial sublimons were found to be scattered around the genome, but several hotspot regions were detected, and marked microscale heterogeneity was also observed in specific areas. One particularly prevalent class of myocardial sublimons had heterogeneous break points, respectively, in the region near the terminus of the D-loop (np 16067–16085) and in the gene for tRNA-leu(UUR) (np 3255–3272), generating a 3.75 kb partial duplication interconvertible by homologous recombination with a 12.8 kb deletion. Using semi-quantitative fluorescent PCR, the abundance of this prevalent class of sublimons in heart was found to vary over three orders of magnitude between different healthy individuals (18), but no simple relationship was found with age. Another prevalent myocardial sublimon, previously reported as a 7.4 kb deletion (9–14), had break-points at the D-loop terminus as above, and at a 12 bp repeat of the same sequence, located within the gene for ATP synthase subunit 6.
To assess the relationship between these molecules and cardiac pathology and aging we have now studied a series of 300 autopsied hearts of well characterized, consecutive male victims of sudden death (Helsinki Sudden Death Study). Based on a robust statistical analysis of the findings, we are able to conclude that despite a tendency to accumulate with age, there is no significant correlation between sublimon abundance and any common, cardiac disease phenotype, such as CAD, myocardial infarction or pathological changes in the myocardium (hypertrophy or dilatation) related to the development of cardiomyopathies.
RESULTS
The large set of autopsied hearts in this study represents a unique resource, because it offers the possibility of combining detailed phenotypic and pathological data with genetic analysis.
Besides a careful autopsy scrutiny, the areas of atherosclerotic lesions as well as cardiac dimensions were measured by planimetry and coronary narrowings via a silicone cast of the arteries; a structured interview of the spouse was conducted to obtain data on previous illnesses and cardiovascular risk factors (for details see Materials and Methods and references therein).
Prevalent-class myocardial sublimons accumulate with age
Myocardial DNA samples from all 300 individuals in the series were analysed for the presence and abundance of the most prevalent class of sublimons (3.75 kb partial duplication), alongside a single-copy nuclear DNA standard for quantitation and validation of the assay (18). The prevalent class sublimons were above the detection limit (approximately 0.1 copies per cell) in 266 of the 294 samples successfully analysed, varying in abundance up to 91 copies per cell (mean 6.1, SD 9.9). Repeat analyses for any given sample gave almost identical values (invariably within 10%).
Despite considerable scatter in the data, a tendency for sublimon abundance to increase with age is evident even without statistical analysis, although the scatterplot with age (Fig. 1) shows no clear threshold value of age for sublimon accumulation. Furthermore, high copy numbers were seen among subjects of all ages, but some of the oldest subjects also had sublimon levels at or near the detection limit. Mean ages for the three groups (N = 98 in each), defined by the 33.33rd and 66.67th percentiles of sublimon abundance, were 48.5, 51.6 and 56.6 years.
Abundance of prevalent-class myocardial sublimons is unrelated to cardiac pathology
We performed multiple regression analysis (RA) (Table 1) on those samples for which comprehensive phenotypic data were available (N = 101). When all selected factors (age, smoking, alcohol consumption, body surface area (BSA), post mortem time, severity of CAD and life-time diabetes) were entered into the model, the number of prevalent-class sublimon copies per cell was independently determined in a statistically significant manner only by age (β = 0.43, P < 0.0001) and by the number of cigarettes smoked (β = 0.25, P = 0.012). Forward and backward stepwise analysis methods gave the same significant factors.
The mean sublimon copy number did increase with the severity of CAD, expressed as the number of vessels (0, 1, 2–3) having >50% luminal obstruction (Table 2). However, in analysis of covariance (ANCOVA), after adjusting for age, the association of CAD severity with sublimon levels was not significant (P = 0.29, N = 242; Table 2). The total area of atherosclerotic lesions in each of the three main coronary arteries also did not correlate with cellular copy number of the 3.75 kb class of sublimons (P = 0.38–0.88, N = 265–283; Table 3).
The presence of diffuse interstitial microscopic myocardial fibrosis (P = 0.53, N = 294) or an autopsy diagnosis of dilatation, hypertrophy or other myocardial pathologic finding (cardiomyopathy, fatty infiltration of myocardium) (P = 0.62, N = 294), also showed no relationship with sublimon copy number.
In addition to these structural diagnoses, actual ventricular dimensions and weights measured at autopsy were analysed. In ANCOVA none of the cardiac dimensions, adjusted for BSA and age, was related to sublimon copy number (divided into three groups, see Materials and Methods; data not shown). Whether the subject was a smoker or a non-smoker (P = 0.24, N = 139), or had a history of diabetes (P = 0.54, N = 141), was not related to sublimon abundance. When all analyses were repeated with age as an independent factor (divided into three groups, see Materials and Methods), statistical associations remained essentially the same.
Among men with a recorded cardiovascular cause of death, the actual causes of death are provided in Table 4. Ischemic heart disease was the cause of death in the majority (76.1%) of these cases, whereas specific myocardial diseases were rare in the series. Adjusting for age, there were no statistically significant differences in abundance of the 3.75 kb sublimon class between cases with or without these diagnoses (Table 4). Eleven of the patients (9.4%) had alcoholic cardiomyopathy, and eight had either cardiomegaly or unknown cardiomyopathy. There were no cases with myocarditis, idiopathic dilated cardiomyopathy (DCM), hypertrophic obstructive cardiomyopathy or arrhythmogenic right ventricular dysplasia.
Abundance of different sublimons in a given sample shows strong correlation
The above findings raise the question of generality: i.e. whether the abundance of the prevalent class sublimons is indicative of sublimon levels in general, or whether this is a unique category whose abundance is independent of that of other mtDNA rearrangements. To test this, we compared the abundance of the next most prevalent class of sublimons (the 7.4 kb deletion) with that of the 3.75 kb class, in a subset of the individuals surveyed. The 7.4 kb deletion, where present, was found always as a single molecular species, without the break-point heterogeneity characteristic of the 3.75 kb sublimon class (Fig. 2). The abundance of the two sublimon classes was closely correlated in the 26 individuals compared (Fig. 3; Pearson’s correlation coefficient r = 0.81 on the square-root-transformed data as plotted, mean ratio of abundances 3.47, SD 4.16). This is consistent with the proposition that the abundance of the most prevalent sublimon class is a good measure of the overall abundance of sublimons in any given specimen. A rare sublimon was also studied by fluorescent PCR (data not shown), but was undetectable in all but a handful of cases, where it was found close to the detection limit of 0.1 copies per cell.
DISCUSSION
Previous studies have suggested the accumulation of rearranged mtDNAs with age as well as in association with various forms of cardiac disease. However, such reports have relied generally on small sample sets and are of low statistical reliability. The data reported here provide the first statistically robust approach to both of these key questions in human mitochondrial genetics.
Sublimon accumulation and heart disease
This analysis covers by far the largest, unbiased set of individuals yet investigated for the relationship between the abundance of rearranged mtDNAs in heart muscle and heart pathology. It demonstrates that mtDNA sublimon copy number in heart does not associate with any common cardiac disease or phenotypic trait studied. This conclusion is clearly at variance with the common view of the published literature concerning both ischemic heart disease and cardiomyopathy. The most likely explanation for this discrepancy lies in the small number of samples, and especially controls, studied by previous authors.
For example, Corral-Debrinski et al. (15) inferred an association between mtDNA rearrangements and ischemic heart disease, based on analysis of DNAs from just five patients. The 7.4 kb deletion was analysed quantitatively in just two patients and no controls, although it was detected qualitatively in each of four controls who were within the same age range as the patients. Pre-selection of a small group of patients with chronic, diagnosed CAD, may also have biased the earlier findings. The common (5 kb) deletion was reported at elevated levels in these patients, yet in control hearts or even those with severe DCM it is present at much lower levels than the 7.4 kb deletion (13), suggesting that tissue changes consequent upon chronic disease or its management could account for the findings in this limited group of patients. In contrast, our unbiased, population-based survey included near 100 cases in whom CAD, with or without myocardial infarction, was the diagnosed cause of sudden death. This allows for a much more rigorous statistical analysis which, after taking account of the confounding factor of age, did not support any association between CAD and mtDNA rearrangements.
Similar considerations may apply to the report by Kim et al. (13), where the prevalence of the 7.4 kb deletion was studied in 24 patients with severe DCM, compared with just nine controls. The median age of the controls was also 33 years, whereas that of the patients was 43. Only three controls were above age 40, and all three had detectable levels of the 7.4 kb deletion. From the scatter-plots shown in Figures 1 and 3, it is clear that a minority of individuals, especially in the older age groups, have sublimon levels far outside of the range seen in the majority. Inclusion of just one or two such individuals in the patient group studied by Kim et al. (13) would have been sufficient to shift the mean of the distribution by a considerable amount, whilst having little obvious effect on the SEs. In any case, this extreme form of cardiac disease, affecting only a very small minority of individuals, is a clinical diagnosis based on hospital examinations. At autopsy, DCM is only manifested by changes in the dimensions of the cardiac chambers. In our series, the mean sublimon copy number was not elevated among patients with either alcoholic cardiomyopathy (one type of DCM), cardiomegaly or unknown cardiomyopathy as an underlying cause of death. Moreover, we were able to study sublimon levels in relation to changes in cardiac dimensions, indicative of the subclinical stage of myocardial diseases. No association was found.
Thus, although our series did not include a sufficient number of cases to prove or disprove previous hypotheses concerning specific, rare forms of cardiomyopathy, we found no evidence to support the notion that sublimon accumulation accompanies the pathological (sub- or preclinical) changes underlying this clinical disease entity. Earlier it has been estimated that cases of idiopathic DCM associated with significant levels of deleted mtDNAs, such as the inherited form described by Suomalainen et al. (20), may be very rare. Overall, our findings exclude mtDNA deletions from being of causal significance in common heart diseases underlying sudden cardiac death.
Another serious drawback of several previously published analyses is the type of quantitation employed (serial-dilution PCR, using primers for a rarely deleted region of mtDNA as a control for ‘total’ mtDNA). Because mtDNA copy number varies considerably between cell types, and is also influenced by tissue remodelling in response to external stimuli (21,22), the use of this standard is also potentially subject to the effects of tissue changes in individuals with chronic disease. Another problem frequently encountered with mtDNA amplimers is cross-reaction with mitochondrial pseudogenes in nuclear DNA (23). Tissue type or quality plus details of sample preparation seem to be major variables influencing the degree to which a given primer pair amplifies bona fide mtDNA as opposed to nuclear pseudogenes (24–26). Our use of a primer pair for a single-copy nuclear gene as an internal standard for quantitating mtDNA rearrangements avoids these potential artefacts. Our sublimon-specific primers also do not amplify mtDNA-related sequences in nuclear DNA (18).
Phenotypic consequences of age-associated sublimon accumulation
In multi-variate analyses we found myocardial sublimon abundance to correlate with age and the intensity of lifetime smoking, independently of other cardiac risk factors. However, although some of the older men had elevated levels of myocardial sublimons, most had only a few copies per cell or less, in common with the younger subjects. Therefore, sublimon accumulation is not a necessary concomitant of aging. However, the data for two different rearrangements that we studied are highly concordant, implying that sublimon accumulation, where it occurs, is general, rather than applicable only to specific rearrangements.
The finding of an overall age association is in line with previous studies carried out on many fewer individuals. These earlier studies, using semi-quantitative PCR to detect specific, deleted mtDNA species in heart (7,15,27,28), also found considerable variation in deletion levels amongst subjects of any given age. The amount of variation indicates that factors other than age must be involved in determining cardiac sublimon levels. These may be environmental and/or genetic. Cigarette smoking appears to be one such factor, but alcohol consumption is not. Nevertheless, the lack of phenotypic correlations in our study strongly suggests that the age-associated accumulation of somatic mtDNA re-arrangements seen in some individuals is of no pathological significance in regard to cardiac aging, at least where sublimon levels remain within the limits revealed in this study (<100 copies per cell of the prevalent class).
Could sublimons, at the maximum levels observed here in aged subjects, be expected to account for any cardiac disease phenotype? Although there is at present no reliable method to quantitate total sublimon load in any DNA sample, the prevalent sublimon class studied here probably represents a significant fraction of all sublimons present, based on previous analyses (17,18). Therefore, the total sublimon load is unlikely to exceed a few percent of heart muscle mtDNA in any case studied in the present series.
In cultured cell models, deleted molecules must comprise at least 60% of the total mtDNA to produce a marked impairment of respiration (29). In human mtDNA disorders, a mutant gene dosage of at least 10–50%, usually more, has been observed in clinically affected, post-mitotic tissues. In one post mortem study (30), deleted mtDNAs were found to represent ∼15 and ∼35% of total heart mtDNA in two adPEO patients, yet were not associated systematically with any specific heart pathology. Cardiac symptoms are generally absent from adPEO patients (M.Zeviani, personal communication), although cardiomyopathy is present in some cases of arPEO (6), where amounts of deleted mtDNA may be even higher. To be of physiological significance, rearranged mtDNAs would have to segregate to high levels in individual cells, but in this case, only a small minority of cells would be affected, even in those individuals in whom we found the highest levels of sublimons.
Reports of an age-associated decline in respiratory chain function (31–34) have been used to support the functional significance of small amounts of specific mtDNA deletions. However, other authors have found no such decline in respiratory chain function with age (35), which is the prevailing conclusion from more recent studies (36,37), including, in particular, studies of the senescent heart (38). Histological studies (33) have shown the presence of an increasing, though still small number of COX-negative muscle fibres in aged individuals, but their functional impact at the whole tissue level is doubtful.
One remaining, although questionable possibility is that tiny amounts of rearranged mtDNA that accumulate in individual cells might somehow trigger apoptosis, leading to tissue destruction combined with loss of the mutant mtDNA molecules responsible. Recent studies in human cells (39) and in mice (40) have indeed provided evidence that loss of respiratory chain function can sometimes provoke apoptosis. However, mice with significant amounts of deleted mtDNA in all tissues, including heart, show normal respiratory chain activity until the deleted mtDNA molecules predominate (41), suggesting that a ‘hit and run’ mechanism whereby low levels of rearranged mtDNA molecules trigger apoptosis and lead to gross tissue damage is extremely unlikely in human heart.
In earlier studies, the wide data scatter and low number of subjects studied made it impossible to judge the degree to which sublimon accumulation was a linear function of age, or even to estimate the extent of this variation in a statistically meaningful manner. Our study, based on 300 individuals, now makes it possible to provide at least tentative answers to both questions. After square-root transformation (Fig. 4), we found sublimon levels within any given 5 year age class to be distributed around a mean value with an approximately constant SD of slightly <100% of the mean. Furthermore, we found that the mean square-root of sublimon abundance increases with age in an approximately linear fashion. Extrapolating from this analysis, the prevalent sublimon class would reach approximately 100 copies per cell (some 2% of all mtDNA) in a typical individual at around age 210 years. Even within the tiny subpopulation of individuals who accumulate sublimons at the upper end of the rate distribution, the prevalent sublimons are unlikely to reach a pathologically significant level within any foreseeable human life span. Thus, sublimon accumulation is best regarded as an epi-phenomenon of aging.
MATERIALS AND METHODS
The autopsy series
A series of 300 consecutive victims of sudden death, all males (aged 33–69 years, mean 51.2 years, SD 9.6 years), was collected at the Department of Forensic Medicine, University of Helsinki in 1991–1992 (Helsinki Sudden Death Study). The cause of death was cardiovascular disease in 39%, other disease in 21% and accident or suicide in 40%. The Ethics Committee of the Department of Forensic Medicine, University of Helsinki, approved the study.
Autopsy methods
A post mortem coronary angiography was performed and all local coronary narrowings were measured from a silicone cast model of the coronary tree (42). The presence or absence of coronary thrombus was recorded at autopsy by opening the coronary arteries longitudinally. The presence of myocardial infarction was based on macroscopic and microscopic examination and by the nitroblue-tetrazolium enzyme test (43). The heart and separated ventricles were weighed and cardiac dimensions at the equatorial region of the ventricles drawn into transparent sheets and later analysed by computer-assisted planimetry (44). The areas of different types of atherosclerotic lesions (fatty streaks, fibrotic lesions, complicated lesions, calcifications) were separately analysed from whole-mount preparations of coronary arteries by histochemical stainings followed by computer-assisted planimetric analysis (45). A measure of the total atherosclerotic burden, the total area of atherosclerotic lesions, was calculated in each coronary artery as a sum of the areas of fatty streaks and fibrotic lesions (complicated lesions and calcifications are formed on these lesions, thus not adding to the total lesion area) (45). Complete data on previous illnesses and cardiovascular risk factors was obtained by a personal interview of the spouse or close acquaintance in 147 cases (44).
DNA extraction and fluorescent PCR
Heart muscle samples from left ventricular myocardium were frozen at –70°C and DNA was extracted by standard methods (17). The break-point regions of the two most prevalent classes of sublimons were amplified by PCR, as previously described by Kajander et al. (18), using one fluorescently labelled oligonucleotide primer (np 16153–16133 of the human mtDNA sequence) and one unlabelled primer (np 3150–3168 for the 3.75 kb partial duplication and np 8531–8550 for the 7.4 kb deletion). For semi-quantitative analysis, multiplex reactions included additional primers to amplify a single-copy gene as an internal standard, as previously described by Kajander et al. (18). Fluorescent products were analysed by capillary electrophoresis using GeneScan software on an Applied Biosystems 310 Genetic Analyzer, which resolves at the nucleotide level. Sublimon abundances were computed relative to the single-copy standard, under conditions where saturation effects were excluded (18).
Statistical analysis
The factors included in the multiple RA were age, BSA, smoking (number of cigarettes/day), alcohol consumption (g/day), maximal coronary stenosis, the presence of hypertension and diabetes (both dummy-coded variables) and post mortem time. The multiple RA results are given as standardized regression coefficients (β) and P-values. For ANCOVA, sublimon abundance (copies/cell) was divided into three equally sized groups (N = 98 in each) by the 33.33rd and 66.67th percentiles, corresponding to values 1.19 and 4.86 copies per cell, respectively. Age was similarly divided into three equally sized groups of <46, 46–58 and >58 years (N = 103, 100 and 97, respectively). Where used as a categorical variable, the severity of CAD was classified as the number of arteries having >50% luminal stenosis at autopsy (corresponding to 0, 1 or 2–3 vessel disease). Variables with grossly asymmetric distribution were square-root- or log-transformed before statistical analyses. P-values <0.05 were considered as statistically significant. The analyses were made using Statistica/Win (Version 5.0, StatSoft) and SPSS/Win (Version 9.0).
ACKNOWLEDGEMENTS
We are extremely grateful to Professor Antti Penttilä for initiating the collection of the autopsy series used in this study, to Kaisa Lalu for participating in its collection, to Seppo Tyynelä and Synnöve Staff for assistance with measuring cardiac dimensions and to Jarkko Pajarinen, Vesa Savolainen and Markus Perola for DNA extraction. This study was supported by grants from the Academy of Finland, Tampere University Hospital Medical Research Fund, European Union, Yrjö Jahnsson Foundation, Finnish Foundation of Alcohol Research and the Pirkanmaa Region Fund of the Finnish Cultural Foundation.
To whom correspondence should be addressed at: Institute of Medical Technology, University of Tampere, 33014 Finland. Tel: +358 3 215 7731; Fax: +358 3 215 7710; Email: howy.jacobs@uta.fi
Figure 1. Relationship between sublimons and age, N = 294. (A) Sublimon abundance (3.75 kb class) plotted against age. (B) Bar chart showing distribution of the men according to age and sublimon copy number, both divided into three equal sized groups by 33.3rd and 66.7th percentiles of the continuous variables. Cases in each age group sum up to 100%.
Figure 1. Relationship between sublimons and age, N = 294. (A) Sublimon abundance (3.75 kb class) plotted against age. (B) Bar chart showing distribution of the men according to age and sublimon copy number, both divided into three equal sized groups by 33.3rd and 66.7th percentiles of the continuous variables. Cases in each age group sum up to 100%.
Figure 2. Semi-quantitative, multiplex fluorescent PCR analysis of the representation of the 7.4 kb deletion sublimon. Panels shown are outputs from GeneScan analysis carried out using the Applied Biosystems 310 capillary electrophoresis instrument, which resolves to the 1 nt level. The three panels shown represent examples of individuals with high (BT-180), intermediate (BT-166) and low (BT-26) abundance of the 7.4 kb deletion, compared with the single-copy internal standard.
Figure 2. Semi-quantitative, multiplex fluorescent PCR analysis of the representation of the 7.4 kb deletion sublimon. Panels shown are outputs from GeneScan analysis carried out using the Applied Biosystems 310 capillary electrophoresis instrument, which resolves to the 1 nt level. The three panels shown represent examples of individuals with high (BT-180), intermediate (BT-166) and low (BT-26) abundance of the 7.4 kb deletion, compared with the single-copy internal standard.
Figure 3. Correlation between the abundances of the two most prevalent myocardial sublimon classes. The abundance of the 7.4 kb deletion was measured in DNA from 26 individuals, and plotted against that of the most prevalent (3.75 kb) sublimon class in the same specimens.
Figure 3. Correlation between the abundances of the two most prevalent myocardial sublimon classes. The abundance of the 7.4 kb deletion was measured in DNA from 26 individuals, and plotted against that of the most prevalent (3.75 kb) sublimon class in the same specimens.
Figure 4. Profile of sublimon accumulation with age. Data from Figure 1 square-root transformed, grouped into 5 year age intervals and plotted as a moving average (mean square-root of sublimon abundance, expressed as copies per cell, for all individuals within each 5 year age interval). For clarity, SDs are plotted only for the intervals ending at ages 40, 50, 60 and 70, but are similar for all intermediate intervals.
Figure 4. Profile of sublimon accumulation with age. Data from Figure 1 square-root transformed, grouped into 5 year age intervals and plotted as a moving average (mean square-root of sublimon abundance, expressed as copies per cell, for all individuals within each 5 year age interval). For clarity, SDs are plotted only for the intervals ending at ages 40, 50, 60 and 70, but are similar for all intermediate intervals.
Multiple RA results of potential factors affecting sublimon abundance, given as standardized regression coefficients (β) and P-values (N = 101)
| Model | β | P-value |
| Age | 0.43 | <0.0001 |
| Smoking (cigarettes/day) | 0.25 | 0.012 |
| Alcohol consumption (g/day) | –0.14 | 0.14 |
| BSA | 0.12 | 0.21 |
| Post mortem time | 0.09 | 0.33 |
| Coronary artery diseasea | –0.02 | 0.85 |
| Diabetesb | –0.001 | 0.95 |
| Model | β | P-value |
| Age | 0.43 | <0.0001 |
| Smoking (cigarettes/day) | 0.25 | 0.012 |
| Alcohol consumption (g/day) | –0.14 | 0.14 |
| BSA | 0.12 | 0.21 |
| Post mortem time | 0.09 | 0.33 |
| Coronary artery diseasea | –0.02 | 0.85 |
| Diabetesb | –0.001 | 0.95 |
aMaximal luminal stenosis of any coronary artery.
bAs recorded by the interview.
Severity of CAD and sublimon abundance in heart (N = 242)
| Severity of CADa | N | Mean sublimon copy number/cell ± SEMb | Mean age |
| 0 vessels | 177 | 4.94 ± 0.59 | 49.5 |
| 1 vessel | 43 | 6.35 ± 1.21 | 55.1 |
| 2–3 vessels | 22 | 9.89 ± 2.60 | 57.5 |
| Severity of CADa | N | Mean sublimon copy number/cell ± SEMb | Mean age |
| 0 vessels | 177 | 4.94 ± 0.59 | 49.5 |
| 1 vessel | 43 | 6.35 ± 1.21 | 55.1 |
| 2–3 vessels | 22 | 9.89 ± 2.60 | 57.5 |
aNumber of coronary arteries with >50% luminal diameter stenosis.
bAdjusting for age, overall P = 0.29 from ANCOVA for severity of CAD.
Sublimon abundance in heart and coronary atherosclerosis in each of the three main arteries (N = 265–283)
| Sublimon copy number/cell | P b | ||||
| Quartiles of total atherosclerosis area in each vessela | 1st qrt (N = 66–71) | 2nd qrt (N = 66–72) | 3rd qrt (N = 66–70) | 4th qrt (N = 67–70) | |
| LAD | 4.60 | 6.08 | 5.46 | 8.11 | 0.45 |
| LCX | 4.86 | 6.03 | 6.85 | 6.46 | 0.88 |
| RCA | 1.73 | 1.66 | 2.19 | 2.00 | 0.38 |
| Sublimon copy number/cell | P b | ||||
| Quartiles of total atherosclerosis area in each vessela | 1st qrt (N = 66–71) | 2nd qrt (N = 66–72) | 3rd qrt (N = 66–70) | 4th qrt (N = 67–70) | |
| LAD | 4.60 | 6.08 | 5.46 | 8.11 | 0.45 |
| LCX | 4.86 | 6.03 | 6.85 | 6.46 | 0.88 |
| RCA | 1.73 | 1.66 | 2.19 | 2.00 | 0.38 |
aFatty streak and fibrotic lesions combined, divided into four equal-sized groups by the lower and upper quartiles and the median.
bAge-adjusted P-value for mtDNA sublimon (3.75 kb class) abundance from ANCOVA for square-root transformed values.
qrt, quartile; LAD, left anterior descending; LCX, left circumflex; RCA, right coronary artery.
Relationship between mtDNA sublimon abundance (3.75 kb class) in heart and specific causes of cardiovascular death (N = 114)
| Causes of deatha | N (%) | Mean sublimon copy number/cell (mean ± SEM) | P b | |
| Those with this diagnosis | Those without this diagnosis | |||
| Ischemic heart disease | 89 (76.1) | 8.73 ± 1.42 | 4.33 ± 1.19 | 0.25 |
| Alcoholic cardiomyopathy | 11 (9.4) | 4.00 ± 1.42 | 8.08 ± 1.24 | 0.97 |
| Unspecific myodegeneration | 10 (8.5) | 5.08 ± 2.45 | 7.94 ± 1.22 | 0.28 |
| Cardiomegalia NUD | 7 (6.0) | 5.32 ± 3.49 | 7.85 ± 1.19 | 0.41 |
| Unknown cardiomyopathy | 1 (0.9) | 0.98 | 7.75 ± 1.14 | 0.71 |
| Lethal arrhythmia | 1 (0.9) | 7.04 | 7.70 ± 1.14 | 0.84 |
| Causes of deatha | N (%) | Mean sublimon copy number/cell (mean ± SEM) | P b | |
| Those with this diagnosis | Those without this diagnosis | |||
| Ischemic heart disease | 89 (76.1) | 8.73 ± 1.42 | 4.33 ± 1.19 | 0.25 |
| Alcoholic cardiomyopathy | 11 (9.4) | 4.00 ± 1.42 | 8.08 ± 1.24 | 0.97 |
| Unspecific myodegeneration | 10 (8.5) | 5.08 ± 2.45 | 7.94 ± 1.22 | 0.28 |
| Cardiomegalia NUD | 7 (6.0) | 5.32 ± 3.49 | 7.85 ± 1.19 | 0.41 |
| Unknown cardiomyopathy | 1 (0.9) | 0.98 | 7.75 ± 1.14 | 0.71 |
| Lethal arrhythmia | 1 (0.9) | 7.04 | 7.70 ± 1.14 | 0.84 |
aNote that five cases had two different cardiovascular diagnoses (there are altogether 119 diagnoses).
bAge-adjusted P-value from ANCOVA for square-root transformed data.
