- Email: [email protected]
Evidence that the mitochondrial genome is the thrifty genome
Diabetes Research and Clinical Practice 45 (1999) 127 – 135 www.elsevier.com/locate/diabres
Evidence that the mitochondrial genome is the thrifty genome Hong Kyu Lee * Department of Internal Medicine, Seoul National Uni6ersity College of Medicine and Department of Biomedical Sciences, National Institute of Health, 28 Yongon-Dong, Chongno-Gu, Seoul 110 744, South Korea
Abstract Although mitochondrial DNA (mtDNA) abnormalities are known to cause insulin deficiency, insulin resistance and diabetes mellitus, it’s quantitative aspect was not addressed well. In this review, mitochondrial genome hypothesis of thrifty phenomenon is proposed, based on the data and review of literatures. From a population based epidemiologic study, it was found that mtDNA quantity was decreased in the peripheral blood of diabetic subjects, and also in those subjects who will convert to diabetes mellitus within 2 years. In this population, low mtDNA subjects were found to have higher blood pressure and high waist hip ratio. These findings suggested mtDNA status might be quantitatively linked to the insulin resistance syndrome. As quantitative relationships between peripheral blood mtDNA levels and insulin requirement, and energy utilization pattern (fat and carbohydrate oxidation during hyperinsulinemic clamp studies) were observed in a group of male students; and maternal mtDNA content (peripheral blood) correlated with birth weight and peripheral blood mtDNA content of the offspring in another study, possibility of thrifty phenotype phenomena might be due to the low mitochondrial status arose. As thrifty phenotype phenomenon shows the quantitatively continuous relationship between involved parameters and characteristics of ‘imprinting’, a possible mechanism is suggested. © 1999 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Mitochondrial genome; Thrifty genome; Mitochondrial DNA
1. Introduction: clustering of cardiovascular risk factors In 1988 Reaven [1] proposed that diabetes be studied as part of a syndrome, consisting of resistance to insulin-stimulated glucose uptake, glucose intolerance, hyperinsulinemia, increased very-low-density lipoprotein decreased high density lipoprotein, and hypertension. He named it * Tel.: +82-2-760-2228/2238; fax: +82-2-762-9662.
syndrome X, as the nature or the cause of this syndrome was unknown. In the following years the more appropriately named insulin resistance syndrome or Reaven’s syndrome became accepted; as this syndrome was difficult to define, however, it remains at best a loosely defined clinical entity [2]. Insulin resistance has been recognized as the fundamental underlying metabolic defect of this syndrome and various mechanisms such as dyslipidemia, non-insulin-dependent diabetes mellitus (NIDDM), hypertension, and coag-
0168-8227/99/$ - see front matter © 1999 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0 1 6 8 - 8 2 2 7 ( 9 9 ) 0 0 0 4 2 - X
128
H.K. Lee / Diabetes Research and Clinical Practice 45 (1999) 127–135
ulation abnormalities by which insulin resistance causes atherosclerosis are beginning to be understood [3]. However, the underlying cause(s) of insulin resistance itself remain largely unknown, and there is no easy quantitative measure for insulin resistance, as applied to the hypertension or the hypercholesterolemia. As insulin resistance phenotype is understood to be largely inherited [4], the search for the major cause, such as an insulin resistance gene is continuing. In this paper, it is proposed that the genetic defects underlying insulin resistance might be quantitative and possibly qualitative defects of mitochondria. The evidence is based on the results of the Yonchon study, glucose clamp studies involving a group of young Korean medical students, observations of the mitochondrial DNA (mtDNA) status of the new born and a review of the literature.
2. Mitochondrial abnormality can cause insulin deficiency, insulin resistance and diabetes mellitus Mitochondria are cytoplasmic lipoprotein particles and are the chief agents of intracellular respiration and energy metabolism. As a consequence of its endosymbiotic origin, the mitochondrion has an independent replication system and its own DNA. MtDNA codes for 13 subunits of respiratory chain complexes and its own structural rRNAs and tRNAs for its own expression. MtDNA is a 16 569 bp long double-stranded circular DNA and shows high polymorphism among different ethnic groups. Several mitochondrial deletions, mutations and other abnormalities have been implicated in the pathogenesis of diabetes and many other disease states [5] and their relation to diabetes was extensively reviewed by Gerbitz et al. [6,7]. The fact that a mitochondrial DNA mutation/ deletion can cause a unique type of diabetes is now well established. Recently Massen and Kadowaki [8] proposed that one mtDNA mutation, G to A substitution at position 3243 of mtDNA encoded tRNA, causing maternally inherited diabetes and deafness (MIDD), be incorporated into the WHO classification. The precise cellular defect
resulting from this mutation is under investigation, though there is evidence to suggest that decreased insulin secretion from pancreatic beta cells is the main mechanism; insulin resistance was observed in those subjects who manifested overt diabetes [9]. The possible modes of involvement of mitochondrial metabolism in glucose induced insulin secretion in beta cells have been illustrated by Gerbitz et al. [7]. Glucose enters the cell through a specific transporter (Glut 2) and stimulates binding of glucokinase (GK) to the mitochondrial pore protein, porin. This is followed by phosphorylation of glucose, activation of glycolysis, and stimulation of mitochondrial oxidative phosphorylation (OXPHOS), resulting in increased intracellular ATP. This leads to the closure of the ATP sensitive K + channel, the opening of the Ca2 + channel, and increased intracellular Ca2 + , which eventually triggers insulin secretion. It is thus reasonable to expect that pancreatic beta cells with abnormal mitochondria would show a poor insulin secretory response to glucose stimulation. Cell lines subjected to extended ethidium bromide treatment select out mtDNA depleted cells and these cells can be repopulated with exogenous mitochondria [10]. Clonal cell lines constructed by fusion of an mtDNA depleted osteosarcoma cell line with enucleated fibroblasts from diabetic patients carrying the tRNA 3243 mutation, show abnormal mitochondrial morphology and function, including markedly reduced glucose uptake and phosphorylation capacity, in comparison to cells hybridized with wild type mtDNA [11]. Soejima et al. [12] proved that mtDNA was required for regulation of glucose-stimulated insulin secretion using a mouse pancreatic beta-cell line, MIN6. This finding suggests that in certain circumstances abnormal mtDNA could cause insulin resistance. To confirm this notion, recently Poulton et al. [13] reported that a common mtDNA variant was associated with insulin resistance in adult life. Rothman et al. [14] showed that, not only in patients with NIDDM, but also in apparently healthy, normoglycemic, lean offspring of parents with NIDDM, the reduced insulin-stimulated muscle glycogen synthesis was the major aspect of
H.K. Lee / Diabetes Research and Clinical Practice 45 (1999) 127–135
insulin resistance. In this study, reduced glucose6-phosphate formation appeared to be the limiting step and it was followed by impaired glycogen synthesis. If mitochondrial function is deranged, the ATP supply to generate glucose-6-phosphate should decrease. It should be remembered, however, that while some cells have perfectly normal mtDNA, others can have different mtDNAs or show heteroplasmy, that is a fraction of the mtDNA population in a cell or tissue can have a point mutation, deletion or duplication. As suggested by Gerbitz et al. [7], genetic defects at quite different sites of cellular energy metabolism can lead to diabetes. A common defect in the cytosolic-mitochondrial interplay of energy production can result in both impaired insulin secretion in the beta cell and also peripheral resistance to the hormone in muscle and adipose tissue.
3. Mitochondrial DNA copy number in the peripheral blood can predict the development of diabetes, and correlate with insulin resistance parameters In 1993 Shin [15] found that in peripheral blood leukocytes of patients with diabetes mellitus (mostly NIDDM), the mtDNA copy number had decreased. Antonetti et al. [16] reported that it was lower in muscle, both in insulin-dependent diabetes mellitus (IDDM) and NIDDM. These two studies employed Southern blot analysis to measure mtDNA. This method suffers the problem of poor reproducibility and requires a large amount of DNA. The quantitative polymerase chain reaction (PCR) method [17] is rapid, does not employ a radioisotope and can be used when the DNA sample is small. As mtDNA accounts for 1% of total DNA, small changes in the amount applied could result in huge differences in the final calculation of mtDNA in both methods. However, Wang et al. [18] managed to apply a competitive PCR method using internal standard to mtDNA quantitation, and it was also found that this method shows quite high reproducibility, when measured in triplicate, employing a single batch of internal standard.
129
Using this quantitative PCR, the mtDNA status of diabetic subjects was studied and it was confirmed that, in the peripheral blood leukocytes, mtDNA content was reduced. It was not certain then, however, whether decreased mtDNA was secondary to the diabetic state or was of primary importance. To clarify this point, mtDNA content in nested cohorts involved in the Yonchon study was therefore studied. In 1993 about 2500 residents of Yonchon County over age 30, participated in a community based health and nutrition survey. About 53% of these subjects were followed-up in 1995 [19]. During the intervening 2 years, 67 had developed diabetes mellitus, according to the WHO criteria, and in 23 of these cases it was possible to obtain stored blood samples in which, after clearing the impurities, sufficient DNA remained. When the mtDNA content of this blood was compared with that of 22 health age- and sex-matched subjects, the average mtDNA content in peripheral blood of subjects who had developed diabetes was found to be about 30% less than in controls [20]. A statistically significant correlation between the amount of mtDNA and diastolic blood pressure of all cohorts was also found (in both 1993 and 1995) and waist–hip ratio of all cohorts (in both 1993 and 1995). These findings suggested that the mtDNA content of peripheral blood could serve as a marker for the insulin resistance syndrome, and raised the possibility that mtDNA is a diabetogenic genome, or more appropriately, a genome for insulin resistance syndrome. The relationship between the mtDNA content of peripheral blood and whole body mtDNA status is still not understood, though research aiming to resolve this issue is in progress. It is believed, however, that peripheral blood mtDNA content reflects the status of the body in general; its level shows strong statistical correlation with many parameters of body energy, as described below. While different tissues have different amounts of mtDNA and could show heteroplasmy, mutated mtDNA is seldom contained in peripheral blood [21]. This is because cells containing mutated mtDNA cannot usually survive rapid cellular replication.
130
H.K. Lee / Diabetes Research and Clinical Practice 45 (1999) 127–135
The relation between peripheral blood mtDNA status and other tissue or whole body mtDNA status is not known. When a hybrid cell, containing both wild-type and mutated mtDNA, is produced and stimulated to replicate, cells containing either wild-type and wild-type and mutated mtDNA are produced (replicative segregation). King and Attardi [10] reported these normally restored cells contain a normal amount of mtDNA, suggesting that there is a kind of set point determined by nuclear gene(s). It is thus not unlikely that peripheral blood mtDNA status reflects the subjects genetic make-up, at least in relative terms. To determine whether the mtDNA content of peripheral blood correlated with insulin resistance parameters, we used certain data already obtained during a study involving a group of young Korean medical students at Ulsan University Medical College in 1996 [22]. Twenty-six subjects were studied using either a hyperglycemic (22 cases) or a euglycemic hyperinsulinemic clamp (14 cases); in 11 of the latter cases indirect calorimetry was also applied. When the peripheral blood mtDNA content of these subjects was analyzed, strong negative correlations with insulin secretion parameters of the subjects emerged; it correlated with basal insulin, basal c-peptide levels, and insulin levels stimulated with glucose or glucagon. The lower the mtDNA, the higher the insulin level and its secretion; peripheral mtDNA content was thus a parameter of insulin resistance. Indirect calorimetry showed that mtDNA content correlated positively with fat oxidation during the hyperinsulinemic clamp, and negatively with insulin-induced net fat oxidation decrease (difference between basal fat oxidation and fat oxidation during hyperinsulinemia). The higher the mtDNA level, the greater the degree of fat oxidation. This phenomenon was mirrored by less hyperinsulinemia-induced carbohydrate oxidation. Although this study involved only a small number of subjects, it is believed that the findings again suggest that peripheral blood mtDNA reflects whole body status. It is also believe that mtDNA content correlates not only with insulin secretion parameters, but with the subjects’ energy
utilization pattern. This is particularly significant, since it suggests a biologic mechanism by which decreased mtDNA could cause hyperinsulinemia.
4. Programmed effect of fetal or infant malnutrition (thrifty phenotype) causes insulin resistance syndrome In 1992 Hales and Barker [23] proposed a novel hypothesis of thrifty phenotype in the pathogenesis of type 2 diabetes mellitus; they suggested that poor fetal and early post-natal nutrition imposed mechanisms of nutritional thrift upon a growing individual. The first evidence for the thrifty phenotype hypothesis came from the study by Barker and Osmond [24] of the differences in death rates from cardiovascular disease in different areas of UK and Wales. The death rate was found to be closely related to differences in neonatal mortality 70 or more years ago. Since most neonatal deaths were associated with low birth weight, this finding was thought to indicate that cardiovascular disease is linked to impaired fetal growth or malnutrition in early life. This link was further demonstrated for ischemic heart disease [25], raised blood pressure [26], elevated plasma levels of fibrinogen [27], glucose intolerance [28] and raised serum cholesterol [29], thus encompassing all the features of insulin resistance syndrome. Nutritional thrift programmed early in life is detrimental at times of adequate or excessive nutrition, and can explain why Ethiopian Jews transported to Israel [30], and Nauruan islanders who suffered severe malnutrition during World War II show a high prevalence of diabetes mellitus. The same mechanism could explain also why it has declined recently among Nauruans [31]. Nutrient supply to the fetus is the major regulator of fetal growth, and in experimental animals, maternal undernutrition during pregnancy commonly leads to the birth of the low-weight offspring. However, birth weight is a poor measure of fetal growth, and using a simple chronically implanted measuring device, it was possible to monitor fetal length or girth on a daily basis [32]. The rate of fetal growth usually slows down within 2–3 days of the start of maternal undernu-
H.K. Lee / Diabetes Research and Clinical Practice 45 (1999) 127–135
trition, but resumes promptly after refeeding [33]. However, growth of fetuses that were already growing only slowly before diet restriction, does not slow down further due to undernutrition. Harding and Johnston [34] hypothesized that this slow growth rate resulted from earlier programming, affected by malnutrition. The relation between fetal malnutrition and the insulin resistance syndrome has also been demonstrated in animal models. Young rats fed low protein diet from 3 to 6 weeks of age exhibit impaired insulin secretion and decreased pancreatic beta cell mass, which do not recover for at least 12 weeks [35,36]. Fetal sheep growing slowly during late gestation were found to have a relatively high insulin:glucose ratio, suggesting the presence of insulin resistance in these animals [37]. Since slow growth during late gestation is programmed by nutritional status during early pregnancy, it can be stated that both insulin secretion capacity and insulin sensitivity are programmed in early life. In other animals, such as guinea pigs and rats, those in which growth is retarded inutero show elevated blood pressure at 3 – 4 months of age [38] and those exposed to a protein deficient diet in-utero are hypertensive when they become adult [39].
5. Inheritance pattern of thrifty phenotype With regard to the effect of nutrition on health or thrifty phenotype, the most important piece of information gathered from these animal studies, and some clinical observations, might be the fact that the programming effects of malnutrition in early pregnancy persist for more than one generation. This implies a kind of Lamarckian inheritance, not usual Mendellian inheritance. Zamenhof et al. [40] were the first to observe that rats exposed to undernutrition as fetuses gave birth to offspring which themselves had reduced brain weight and cell numbers. Furthermore, in a rat colony marginally undernourished for nine generations, nutritional improvement during the last week of pregnancy was found to lead to the birth of offspring which were larger than those of well-fed controls. The full correction of abnormal
131
adult body size and behavioral abnormalities required three generations of nutritional rehabilitation [41]. Similar observations were made in humans. Lumey [42] found that women exposed to famine as fetuses in the first half of pregnancy, despite being of normal birth weight themselves, gave birth to offspring with an increased prevalence of growth retardation. Ravelli et al. [43] reported that young men exposed to famine in-utero during late gestation were relatively protected from later obesity, while those exposed early in pregnancy showed an increased prevalence of obesity in adults. In Jamaica, Godfrey et al. [44] found a strong association between thin maternal triceps skinfold thickness at 15 weeks of gestation and raised blood pressure in their offspring, and lower weight gain between 15 and 35 weeks of gestation has been independently associated with raised blood pressure in children. These results suggest that women who became pregnant while in a poor nutritional state and who are programmed to show poor weight gain during pregnancy, will give birth to offspring who tends to have higher blood pressure. This kind of inheritance cannot be explained in the usual Mendellian fashion, but suggests an epigenetic mechanism, such as genetic imprinting. Imprinting of genes by methylation of cytosine residue of DNA occurs in many places, such as in those genes coding Xist, H19, IGF2R and insulin [45,46]. It is suspected that similar imprinting mechanism(s) might operate in this case, possibly involving the gene coding mitochondrial transcription factor A (Tfam), since cellular mtDNA depletion is known to be associated with a low level of Tfam [47]. If birth weight is correlated with insulin resistance syndrome, and mtDNA is a marker of this syndrome, there should be a certain relation between mtDNA content and birth weight. To find if there is any relationship, 27 singleton baby– mother pairs born during early 1997 in Seoul National University Hospital, Seoul, Korea were studied [48]. Maternal mtDNA content in peripheral blood correlated strongly with infant birth weight, and the prepregnancy body weight of mother, cord blood IGF1 level, cord blood
132
H.K. Lee / Diabetes Research and Clinical Practice 45 (1999) 127–135
IGFBP-3 levels also showed statistically significant correlation. In the multiple regression analysis, however, only maternal mtDNA remained significant. Cord blood mtDNA content did not correlate linearly with birth weight, although it correlated positively with maternal mtDNA content. Further studies with a larger number of subjects are needed before reaching a firm conclusion between birth weight and mtDNA of the baby. However, it is worthwhile to note that there is a quantitative relation in mitochondrial inheritance, and this finding fits well with the above mentioned inheritance pattern of thrifty phenotype.
6. Mitochondrial biogenesis in ovum The inheritance of mitochondrial and nuclear genes is fundamentally different, in that mitochondrial gene copy number per cell usually exceeds 1000, replication occurs throughout the cell cycles, and its origin is strict maternal. Even though a small amount of sperm mitochondria are transmitted during fertilization, paternal mtDNA is eliminated early in development by active degradation [49]. The details of mitochondrial biogenesis in oocyte development and gametogenesis remain largely unknown; most existing information was obtained by Hauswirth et al. in the 1980s [50,51]. They showed that in bovine oocytes, mtDNA copy number increased 50 – 200-fold from that of resting germ cell line, which seems to have a similar copy number to somatic cells. Between a primary germ cell line and the largest non-zone pellucida encased fetal oocyte, the increase in mtDNA copy number is at least 3-fold. The same amount of mtDNA is found in primordial follicles in the adult ovary, suggesting that these cells are dormant. From this point and the development of an intact zona pellucida (class II oocytes), further amplification occurs (at least 15-fold). In the final stages of oocyte development, when the amount of cytoplasm increases by at least 6-fold, less than 2-fold increase in mtDNA occurs. MtDNA amplification is essentially complete by the time the zona pellucida is added to a developing oocyte. Mature
mammalian oocytes contain 4–7 pg of mtDNA or 200 000 mtDNA molecules. Along with the increase in mtDNA copy numbers, oocyte cytoplasmic volume increases 240fold, and the number of organelles and the total volume of mitochondria increase proportionally. During rapid expansion of cytoplasmic volume and mitochondrial number, the ratio of mitochondrial to cytoplasmic volume therefore remains constant. These findings suggest that mitochondrial biogenesis is directly linked to developmental increases in cell volume, and imply that to be viable, an oocyte requires a fixed amount of functional mitochondrial surface per unit volume of cytoplasm. A somatic cell contains about ten copies of mtDNA per mitochondrion, as does the primordial oocyte. When an oocyte becomes the smallest class of zone-encased oocytes, there is a rapid reduction in the copy number per mitochondrion, termed mitochondrial meiosis, to about one or 1.5 genomes per daughter mitochondrion. The phenomenon is comparable to that of bacterial replication in fission. This low copy number does not recover significantly during the later stages of oocyte development, rising only to two copies per organelle. No additional mtDNA synthesis occurs until after the 64-cell blastocyst stage after fertilization, and repopulation of each organelle does not occur until after early cell divisions of the embryo. Although few studies have been performed it is likely that some changes occurring during oocyte development persist during embryogenesis and possibly into fetal and adult life, although this has not been well studied. If decreased mtDNA status of the body represents, de facto, the thrifty phenotype, this phenotype could result from this very complex process of mitochondrial biogenesis. MtDNA exists mostly in circular and monomeric forms, but it is known that it can be dimeric or multimeric. Michaels et al. [51] observed that 30–40% of bovine oocyte mtDNA exists in multimeric form, much more than in mouse oocytes, as reported by Piko and Matsumoto [52]. It is not known if an egg with many multimeric forms of mtDNA is fertile, nor what would happen to this multimeric form of mtDNA
H.K. Lee / Diabetes Research and Clinical Practice 45 (1999) 127–135
during mitochondrial meiosis. Because of highly competitive selection of wild-type mtDNA during oogenesis and embryogenesis (only 1 in 1000 is chosen), it is not likely that much of the multimeric form survives.
7. Conclusion and suggestion Further studies are needed to confirm the role of the quantitative mitochondrial paradigm of insulin resistance syndrome pathogenesis. However, it matches well with the concept that there is common soil for this syndrome [53], and the suggestion that the underlying abnormality is a genetic program to fail, for the beta cell to compensate for insulin resistance [54]. Mitochondria are energy-supplying cytoplasmic soil and mtDNA is quite vulnerable to free radicals and fails easily. It is believed that the quantitative mitochondrial paradigm proposed here (Fig. 1) complements the already established qualitative mitochondrial paradigm of diabetogenesis [6]. This paradigm does not contradict either the thrifty phenotype hypothesis [23], supported by considerable epidemiological and experimental evidence, or the thrifty genotype hypothesis, proposed by Neel [55]. Furthermore, it suggests a novel biologic mechanism, that a bodily set-point of mtDNA content exists, that is changes according to the nutritional status of fetal or early development, and that it is inherited. As mammalian mtDNA undergoes enormous amplification and selection during oogenesis and
Fig. 1. Relations between the concepts of low mitochondrial status, insulin resistance, thrifty phenotype and genotype. Insulin resistance state is considered as an adaptive state of low mitochondrial functional capacity induced by malnutrition, which require more insulin or hyperglycemia to overcome mitochondrial energy supply, Tfam, mitochondrial transcription factor A; mtSSBP, mitochondrial single stranded DNA binding protein; POL-r, DNA polymerase gamma.
133
subsequent embryogenesis, it is also possible that abnormal biogenesis of mtDNA could result in an unfavorable environment. Cellular mitochondrial status is determined by nuclear factors, such as Tfam, mitochondrial single stranded DNA binding protein and RNA polymerase gamma [56,57]. Since inheritance of the thrifty phenotype shows characteristics of genetic imprinting, we suspect that the gene coding these genes might be differently methylated during malnutrition and a lower cellular mitochondrial set point. Much is known about mtDNA replication and its control, but to understand the detailed mechanisms of mitochondrial biogenesis and the implications of this poor development, further studies are needed. References [1] G.M. Reaven, Banting lecture 1988. Role of insulin resistance in human disease, Diabetes 37 (1988) 1595 – 1607. [2] I.F. Godsland, J.C. Stevenson, Insulin resistance: syndrome or tendency?, Lancet 346 (1995) 100 – 103. [3] A. Garg, S.M. Haffner, Insulin resistance and atherosclerosis. An overview, Diab. Care 19 (1996) 274. [4] D.E. Moller, C. Bjobek, A. Vidal-Puig, Candidate genes for insulin resistance, Diab. Care 19 (1996) 396 – 400. [5] D.C. Wallace, Diseases of the mitochondrial DNA, Annu. Rev. Biochem. 61 (1992) 1175 – 1212. [6] K.D. Gerbitz, K. Gempel, D. Brdiczka, Mitochondria and diabetes. Genetic, biochemical, and clinical implications of the cellular energy circuit, Diabetes 45 (1996) 113 – 126. [7] K.D. Gerbitz, J.M.W. van den Ouweland, J.A. Massen, M. Jaksch, Mitochondrial diabetes mellitus: a review, Biochim. Biophys. Acta 1271 (1995) 253 – 260. [8] J.A. Maassen, T. Kadowaki, Maternally inherited diabetes and deafness: a new diabetes subtype, Diabetologia 39 (1996) 375 – 382. [9] G. Velho, M.M. Byrne, K. Clement, J. Sturis, M.E. Pueyo, H. Blanche, N. Vionet, J. Fiet, P. Passa, J.J. Robert, K.S. Polonsky, P. Froguel, Clinical phenotype, insulin secretion, and insulin sensitivity in kindreds with maternally inherited diabetes and deafness due to mitochondrial tRNA Leu(UUR) gene mutation, Diabetes 45 (1996) 478 – 487. [10] M.P. King, G. Attardi, Human cells lacking mtDNA: repopulation with exogenous mitochondria by complementation, Science 246 (1989) 500 – 503. [11] J.M.W. van den Ouweland. A New Subypte of Non-Insulin Dependent Diabetes Mellitus Is Associated With a Mitochondrial Gene Mutation. PhD Thesis. Netherlands, University of Leiden, 1994 (cited from [6]).
134
H.K. Lee / Diabetes Research and Clinical Practice 45 (1999) 127–135
[12] A. Soejima, K. Inoue, D. Taka, M. Kaneko, H. Ishihara, Y. Oka, J.-I. Hayashi, Mitochondrial DNA is required for the regulations of glucose-stimulated insulin secretion in a mouse pancreatic beta cell line, MIN6, J. Biol. Chem. 42 (1996) 26194 – 26199. [13] J. Poulton, M.S. Brown, A. Cooper, D.R. Marchington, D.I.W. Phillips, A common mitochondrial DNA variant is associated insulin resistance in adult life, Diabetologia 41 (1998) 54 – 58. [14] D.L. Rothman, I. Magnusson, G. Cline, D. Gerard, C.R. Kahn, R.G. Shulman, G.I. Shulman, Decreased muscle glucose transport/phosphorylation is an early defect in the pathogenesis of non-insulin-dependent diabetes mellitus, Proc. Natl. Acad. Sci. USA 92 (1995) 983–987. [15] C.S. Shin, Quantitative abnormalities of mitochondrial DNA in noninsulin dependent diabetes mellitus, Kor. J. Diab. Ass. 18 (1994) 344–350. [16] D.A. Antonetti, C. Reynet, C.R. Kahn, Increased expression of mitochondrial-encoded genes in skeletal muscle of humans with diabetes mellitus, J. Clin. Invest. 95 (1995) 1383 – 1388. [17] F. Cottrez, C. Auriault, A. Capron, H. Groux, Quantitative PCR: validation of the use of a multispecific internal control, Nucleic Acids Res. 22 (1994) 2712–2713. [18] H. Wang, L. Fliegel, C.E. Cass, A.M.W. Penn, P.M. Michalak, J.H. Weiner, B.D. Lemire, Quantitation of mitochondrial DNA in heteroplasmic fibroblasts with competitive PCR, BioTechniques 17 (1994) 76–78. [19] C.S. Shin, H.K. Lee, C.-S. Koh, Y.I. Kim, Y.S. Shin, K.Y. Yoo, H.Y. Paik, Y.S. Park, B.K. Yang, Risk factors for the development of NIDDM in Yonchon county, Korea, Diab. Care 20 (1997) 1842–1846. [20] H.K. Lee, J.H. Song, C.S. Shin, D.J. Park, K.S. Park, K.U. Lee, C.S. Koh, Peripheral blood mitochondrial DNA content can predict the development of diabetes (Abstract), Diabetes 46 (Suppl 1) (1997) 175A. [21] B. Obermaier-Kusser, J. Muller-Hocker, I. Nelson, P. Lestienne, C. Enter, T. Riedele, K.D. Gerbitz, Different copy numbers of apparently identically deleted mitochondrial DNA in tissues from a patient with Kearnes–Sayer syndrome detected by PCR, Biochem. Biophys. Res. Commun. 169 (1995) 1007–1015. [22] H.K. Lee, K.-U. Lee, J.H. Song, J.Y. Park, C.S. Shin, K.S. Park, D.J. Park. Peripheral mitochondrial DNA content correlated with insulin secretion in young Koreans (Abstract). International Diabetes Epidemiology Symposium, Savonlinna, Finland, 1997. [23] C.N. Hales, D.J. Barker, Type 2 (non-insulin-dependent) diabetes mellitus: the thrifty phenotype hypothesis, Diabetologia 35 (1992) 595–601. [24] D.J.P. Barker, C. Osmond, Infant mortality, childhood nutrition and ischemic heart disease in England and Wales, Lancet 1 (1986) 1077–1081. [25] D.J. Barker, P.D. Winter, C. Osmond, B. Margetts, S.J. Simmonds, Weight in infancy and death from ischaemic heart disease, Lancet 2 (1989) 577–580.
[26] D.J. Barker, A.R. Bull, C. Osmond, S.J. Simmonds, Fetal and placental size and risk of hypertension in adult life, Br. Med. J. 301 (1990) 259 – 262. [27] D.J. Barker, T.V. Meade, C.H.D. Fall, A. Lee, C. Osmond, K. Phipps, Y. Stirling, The relation of fetal and infant growth to plasma fibrinogen and factor VII levels in adult life, Br. Med. J. 304 (1992) 148 – 152. [28] C.N. Hales, D.J. Barker, P.M.S. Clark, L.J. Cox, C. Fall, C. Osmond, P.D. Winter, Fetal and infant growth and impaired glucose tolerance at age 64 years, Br. Med. J. 303 (1991) 1019 – 1022. [29] D.J. Barker, C.N. Martyn, C. Osmond, C.H. Fall, Growth in utero and serum cholesterol concentrations in adult life, Br. Med. J. 307 (1993) 1524 – 1527. [30] M.P. Cohen, E. Stern, Y. Rusecki, A. Zeidler, High prevalence of diabetes in young adult Ethiopian immigrants to Israel, Diabetes 37 (1988) 824 – 828. [31] G.K. Dowse, P.Z. Zimmet, C.F. Finch, V.R. Collins, Decline in incidence of epidemic glucose intolerance in Nauruans: implications for the thrifty genotype, Am. J. Epidemiol. 133 (1991) 1093 – 1104. [32] D.J. Mellor, I.C. Matheson, Chronic catheterisation of the aorta and umbilical vessels of fetal sheep, Res. Vet. Sci. 18 (1975) 221 – 223. [33] D.J. Mellor, I.C. Matheson, Daily changes in the curved crown-rump length of individual sheep fetuses during the last 60 days of pregnancy and effects of different levels of maternal nutrition, Q. J. Exp. Physiol. Cog. Med. Sci. 64 (1979) 119 – 131. [34] J.E. Harding, B.M. Johnston, Nutrition and fetal growth, Reprod. Fertil. Dev. 7 (1995) 539 – 547. [35] I. Swenne, C.J. Crace, R.D.G. Milner, Persistent impairment of insulin secretory response to glucose in adult rats after limited period of protein caloriemalnutrition early in life, Diabetes 36 (1987) 454 – 458. [36] I. Swenne, L.A.H. Borg, C.J. Crace, A.S. Landstrom, Persistent reduction of pancreatic beta-cell mass after a limited period of protein-energy malnutrition in the young rat, Diabetologia 35 (1992) 939 – 945. [37] P.D. Gluckman, J.E. Harding, The physiology and pathophysiology of intrauterine growth retardation, Horm. Res. 48 (Suppl 1) (1997) 11 – 16. [38] E. Persson, T. Jansson, Low birth weight is associated with elevated adult blood pressure in the chronically catheterized guinea pig, Acta Physiol. Scand. 145 (1992) 195 – 196. [39] S.M. Woodall, B.M. Johnston, B.H. Breier, P.D. Gluckman, Chronic maternal undernutrition in the rat leads to delayed postnatal growth and elevated blood pressure of offspring, Pediatr. Res. 40 (1996) 438 – 443. [40] S. Zamenhof, E. van Marthens, L. Grauel, DNA (cell number) in neonatal brain: second generation (F2) alteration by maternal (Fo) dietary protein restriction, Science 172 (1971) 850 – 851. [41] R.J. Stewart, H. Sheppard, R. Preece, J.C. Waterlow, The effect of rehabilitation at different stages of development of rats marginally malnourished for ten to twelve generations, Br. J. Nutr. 43 (1980) 403 – 412.
H.K. Lee / Diabetes Research and Clinical Practice 45 (1999) 127–135 [42] L.H. Lumey, Decreased birth weights in infants after maternal in-utero exposure to the Dutch famine of 1944– 1945, Paediatr. Perinat. Epidemiol. 6 (1992) 240. [43] G.P. Ravelli, Z.A. Stein, M.W. Susser, Obesity in young men after famine exposure in utero and early infancy, N. Engl. J. Med. 295 (1976) 349–353. [44] K.M. Godfrey, T. Forrester, D.J. Barker, A.A. Jackson, J.P. Landman, J.S.E. Hall, V. Cox, C. Osmond, Maternal nutritional status in pregnancy and blood pressure in childhood, Br. J. Obstet. Gynaecol. 101 (1994) 398–403. [45] M.S. Bartolomei, S.M. Tilghman, Genomic imprinting in mammals, Annu. Rev. Genet. 31 (1997) 493–525. [46] M.A. Surani, Imprinting and the initiation of gene silencing in the germ line, Cell 93 (1988) 309–312. [47] N.G. Larsson, A. Oldfors, E. Holme, D.A. Clayton, Low levels of mitochondrial transcription factor A in mitochondrial depletion, Biochem. Biophys. Res. Commun. 200 (1994) 1374 – 1381. [48] K.S. Park, Y.Y. Lee, J.H. Song, J.J. Koh, C.S. Shin, D.J. Park, B.H. Yoon, J.K. Jun, H.K. Lee, The relationship of birth weight to insulin-like growth factors (IGF) and the amount of mitochondrial DNA in umbilical cord blood (Abstract), Diabetes 47 (Suppl 1) (1998) A321. [49] H. Kaneda, J. Hayashi, S. Takahama, C. Taya, K.F. Lindahl, H. Yonekawa, Elimination of paternal mitochondrial DNA in intraspecific crosses during early mouse embryogenesis, Proc. Natl. Acad. Sci. USA 92 (1995) 4542 – 4546.
.
135
[50] W.W. Hauswirth, P.J. Laipis, Transmission genetics of mammalian mitochondria: a molecular model and experimental evidence, in: E. Quagliariello, et al. (Eds.), Achievements and Perspectives of Mitochondrial Research: Biogenesis, vol. 2, Elsevier, Amsterdam, 1985. [51] G.S. Michaels, W.W. Hauswirth, P.J. Laipis, Mitochondrial copy number in bovine oocytes and somatic cells, Dev. Biol. 94 (1982) 246 – 251. [52] L. Piko, L. Matsumoto, Number of mitochondria and some properties of mitochondrial DNA in the mouse egg, Dev. Biol. 49 (1976) 1 – 10. [53] M.P. Stern, Diabetes and cardiovascular disease: the common soil hypothesis, Diabetes 44 (1995) 369 – 374. [54] J.L. Leahy, Impaired b-cell function with chronic hyperglycemia: overworked b-cell hypothesis, Diab. Rev. 4 (1996) 298 – 319. [55] J.V. Neel, The thrifty genotype revisited, in: J. Kobberling, R. Tattersal (Eds.), The Genetics of Diabetes Mellitus, Academic Press, New York, 1982. [56] R.N. Lightowlers, P.F. Chinnery, D.M. Turnball, N. Howell, Mammalian mitochondrial genetics: heredity, heteroplasmy and disease, Trends Genet. 13 (1997) 450 – 455. [57] R.A. Schultz, S.J. Swoap, L.D. McDaniel, B. Zhang, E.C. Koon, D.J. Garry, K. Li, R.S. Williams, Differential expression of mitochondrial DNA replication factors in mammalian tissues, J. Biol. Chem. 273 (1998) 3447 – 3451.
Evidence that the mitochondrial genome is the thrifty genome Hong Kyu Lee * Department of Internal Medicine, Seoul National Uni6ersity College of Medicine and Department of Biomedical Sciences, National Institute of Health, 28 Yongon-Dong, Chongno-Gu, Seoul 110 744, South Korea
Abstract Although mitochondrial DNA (mtDNA) abnormalities are known to cause insulin deficiency, insulin resistance and diabetes mellitus, it’s quantitative aspect was not addressed well. In this review, mitochondrial genome hypothesis of thrifty phenomenon is proposed, based on the data and review of literatures. From a population based epidemiologic study, it was found that mtDNA quantity was decreased in the peripheral blood of diabetic subjects, and also in those subjects who will convert to diabetes mellitus within 2 years. In this population, low mtDNA subjects were found to have higher blood pressure and high waist hip ratio. These findings suggested mtDNA status might be quantitatively linked to the insulin resistance syndrome. As quantitative relationships between peripheral blood mtDNA levels and insulin requirement, and energy utilization pattern (fat and carbohydrate oxidation during hyperinsulinemic clamp studies) were observed in a group of male students; and maternal mtDNA content (peripheral blood) correlated with birth weight and peripheral blood mtDNA content of the offspring in another study, possibility of thrifty phenotype phenomena might be due to the low mitochondrial status arose. As thrifty phenotype phenomenon shows the quantitatively continuous relationship between involved parameters and characteristics of ‘imprinting’, a possible mechanism is suggested. © 1999 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Mitochondrial genome; Thrifty genome; Mitochondrial DNA
1. Introduction: clustering of cardiovascular risk factors In 1988 Reaven [1] proposed that diabetes be studied as part of a syndrome, consisting of resistance to insulin-stimulated glucose uptake, glucose intolerance, hyperinsulinemia, increased very-low-density lipoprotein decreased high density lipoprotein, and hypertension. He named it * Tel.: +82-2-760-2228/2238; fax: +82-2-762-9662.
syndrome X, as the nature or the cause of this syndrome was unknown. In the following years the more appropriately named insulin resistance syndrome or Reaven’s syndrome became accepted; as this syndrome was difficult to define, however, it remains at best a loosely defined clinical entity [2]. Insulin resistance has been recognized as the fundamental underlying metabolic defect of this syndrome and various mechanisms such as dyslipidemia, non-insulin-dependent diabetes mellitus (NIDDM), hypertension, and coag-
0168-8227/99/$ - see front matter © 1999 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0 1 6 8 - 8 2 2 7 ( 9 9 ) 0 0 0 4 2 - X
128
H.K. Lee / Diabetes Research and Clinical Practice 45 (1999) 127–135
ulation abnormalities by which insulin resistance causes atherosclerosis are beginning to be understood [3]. However, the underlying cause(s) of insulin resistance itself remain largely unknown, and there is no easy quantitative measure for insulin resistance, as applied to the hypertension or the hypercholesterolemia. As insulin resistance phenotype is understood to be largely inherited [4], the search for the major cause, such as an insulin resistance gene is continuing. In this paper, it is proposed that the genetic defects underlying insulin resistance might be quantitative and possibly qualitative defects of mitochondria. The evidence is based on the results of the Yonchon study, glucose clamp studies involving a group of young Korean medical students, observations of the mitochondrial DNA (mtDNA) status of the new born and a review of the literature.
2. Mitochondrial abnormality can cause insulin deficiency, insulin resistance and diabetes mellitus Mitochondria are cytoplasmic lipoprotein particles and are the chief agents of intracellular respiration and energy metabolism. As a consequence of its endosymbiotic origin, the mitochondrion has an independent replication system and its own DNA. MtDNA codes for 13 subunits of respiratory chain complexes and its own structural rRNAs and tRNAs for its own expression. MtDNA is a 16 569 bp long double-stranded circular DNA and shows high polymorphism among different ethnic groups. Several mitochondrial deletions, mutations and other abnormalities have been implicated in the pathogenesis of diabetes and many other disease states [5] and their relation to diabetes was extensively reviewed by Gerbitz et al. [6,7]. The fact that a mitochondrial DNA mutation/ deletion can cause a unique type of diabetes is now well established. Recently Massen and Kadowaki [8] proposed that one mtDNA mutation, G to A substitution at position 3243 of mtDNA encoded tRNA, causing maternally inherited diabetes and deafness (MIDD), be incorporated into the WHO classification. The precise cellular defect
resulting from this mutation is under investigation, though there is evidence to suggest that decreased insulin secretion from pancreatic beta cells is the main mechanism; insulin resistance was observed in those subjects who manifested overt diabetes [9]. The possible modes of involvement of mitochondrial metabolism in glucose induced insulin secretion in beta cells have been illustrated by Gerbitz et al. [7]. Glucose enters the cell through a specific transporter (Glut 2) and stimulates binding of glucokinase (GK) to the mitochondrial pore protein, porin. This is followed by phosphorylation of glucose, activation of glycolysis, and stimulation of mitochondrial oxidative phosphorylation (OXPHOS), resulting in increased intracellular ATP. This leads to the closure of the ATP sensitive K + channel, the opening of the Ca2 + channel, and increased intracellular Ca2 + , which eventually triggers insulin secretion. It is thus reasonable to expect that pancreatic beta cells with abnormal mitochondria would show a poor insulin secretory response to glucose stimulation. Cell lines subjected to extended ethidium bromide treatment select out mtDNA depleted cells and these cells can be repopulated with exogenous mitochondria [10]. Clonal cell lines constructed by fusion of an mtDNA depleted osteosarcoma cell line with enucleated fibroblasts from diabetic patients carrying the tRNA 3243 mutation, show abnormal mitochondrial morphology and function, including markedly reduced glucose uptake and phosphorylation capacity, in comparison to cells hybridized with wild type mtDNA [11]. Soejima et al. [12] proved that mtDNA was required for regulation of glucose-stimulated insulin secretion using a mouse pancreatic beta-cell line, MIN6. This finding suggests that in certain circumstances abnormal mtDNA could cause insulin resistance. To confirm this notion, recently Poulton et al. [13] reported that a common mtDNA variant was associated with insulin resistance in adult life. Rothman et al. [14] showed that, not only in patients with NIDDM, but also in apparently healthy, normoglycemic, lean offspring of parents with NIDDM, the reduced insulin-stimulated muscle glycogen synthesis was the major aspect of
H.K. Lee / Diabetes Research and Clinical Practice 45 (1999) 127–135
insulin resistance. In this study, reduced glucose6-phosphate formation appeared to be the limiting step and it was followed by impaired glycogen synthesis. If mitochondrial function is deranged, the ATP supply to generate glucose-6-phosphate should decrease. It should be remembered, however, that while some cells have perfectly normal mtDNA, others can have different mtDNAs or show heteroplasmy, that is a fraction of the mtDNA population in a cell or tissue can have a point mutation, deletion or duplication. As suggested by Gerbitz et al. [7], genetic defects at quite different sites of cellular energy metabolism can lead to diabetes. A common defect in the cytosolic-mitochondrial interplay of energy production can result in both impaired insulin secretion in the beta cell and also peripheral resistance to the hormone in muscle and adipose tissue.
3. Mitochondrial DNA copy number in the peripheral blood can predict the development of diabetes, and correlate with insulin resistance parameters In 1993 Shin [15] found that in peripheral blood leukocytes of patients with diabetes mellitus (mostly NIDDM), the mtDNA copy number had decreased. Antonetti et al. [16] reported that it was lower in muscle, both in insulin-dependent diabetes mellitus (IDDM) and NIDDM. These two studies employed Southern blot analysis to measure mtDNA. This method suffers the problem of poor reproducibility and requires a large amount of DNA. The quantitative polymerase chain reaction (PCR) method [17] is rapid, does not employ a radioisotope and can be used when the DNA sample is small. As mtDNA accounts for 1% of total DNA, small changes in the amount applied could result in huge differences in the final calculation of mtDNA in both methods. However, Wang et al. [18] managed to apply a competitive PCR method using internal standard to mtDNA quantitation, and it was also found that this method shows quite high reproducibility, when measured in triplicate, employing a single batch of internal standard.
129
Using this quantitative PCR, the mtDNA status of diabetic subjects was studied and it was confirmed that, in the peripheral blood leukocytes, mtDNA content was reduced. It was not certain then, however, whether decreased mtDNA was secondary to the diabetic state or was of primary importance. To clarify this point, mtDNA content in nested cohorts involved in the Yonchon study was therefore studied. In 1993 about 2500 residents of Yonchon County over age 30, participated in a community based health and nutrition survey. About 53% of these subjects were followed-up in 1995 [19]. During the intervening 2 years, 67 had developed diabetes mellitus, according to the WHO criteria, and in 23 of these cases it was possible to obtain stored blood samples in which, after clearing the impurities, sufficient DNA remained. When the mtDNA content of this blood was compared with that of 22 health age- and sex-matched subjects, the average mtDNA content in peripheral blood of subjects who had developed diabetes was found to be about 30% less than in controls [20]. A statistically significant correlation between the amount of mtDNA and diastolic blood pressure of all cohorts was also found (in both 1993 and 1995) and waist–hip ratio of all cohorts (in both 1993 and 1995). These findings suggested that the mtDNA content of peripheral blood could serve as a marker for the insulin resistance syndrome, and raised the possibility that mtDNA is a diabetogenic genome, or more appropriately, a genome for insulin resistance syndrome. The relationship between the mtDNA content of peripheral blood and whole body mtDNA status is still not understood, though research aiming to resolve this issue is in progress. It is believed, however, that peripheral blood mtDNA content reflects the status of the body in general; its level shows strong statistical correlation with many parameters of body energy, as described below. While different tissues have different amounts of mtDNA and could show heteroplasmy, mutated mtDNA is seldom contained in peripheral blood [21]. This is because cells containing mutated mtDNA cannot usually survive rapid cellular replication.
130
H.K. Lee / Diabetes Research and Clinical Practice 45 (1999) 127–135
The relation between peripheral blood mtDNA status and other tissue or whole body mtDNA status is not known. When a hybrid cell, containing both wild-type and mutated mtDNA, is produced and stimulated to replicate, cells containing either wild-type and wild-type and mutated mtDNA are produced (replicative segregation). King and Attardi [10] reported these normally restored cells contain a normal amount of mtDNA, suggesting that there is a kind of set point determined by nuclear gene(s). It is thus not unlikely that peripheral blood mtDNA status reflects the subjects genetic make-up, at least in relative terms. To determine whether the mtDNA content of peripheral blood correlated with insulin resistance parameters, we used certain data already obtained during a study involving a group of young Korean medical students at Ulsan University Medical College in 1996 [22]. Twenty-six subjects were studied using either a hyperglycemic (22 cases) or a euglycemic hyperinsulinemic clamp (14 cases); in 11 of the latter cases indirect calorimetry was also applied. When the peripheral blood mtDNA content of these subjects was analyzed, strong negative correlations with insulin secretion parameters of the subjects emerged; it correlated with basal insulin, basal c-peptide levels, and insulin levels stimulated with glucose or glucagon. The lower the mtDNA, the higher the insulin level and its secretion; peripheral mtDNA content was thus a parameter of insulin resistance. Indirect calorimetry showed that mtDNA content correlated positively with fat oxidation during the hyperinsulinemic clamp, and negatively with insulin-induced net fat oxidation decrease (difference between basal fat oxidation and fat oxidation during hyperinsulinemia). The higher the mtDNA level, the greater the degree of fat oxidation. This phenomenon was mirrored by less hyperinsulinemia-induced carbohydrate oxidation. Although this study involved only a small number of subjects, it is believed that the findings again suggest that peripheral blood mtDNA reflects whole body status. It is also believe that mtDNA content correlates not only with insulin secretion parameters, but with the subjects’ energy
utilization pattern. This is particularly significant, since it suggests a biologic mechanism by which decreased mtDNA could cause hyperinsulinemia.
4. Programmed effect of fetal or infant malnutrition (thrifty phenotype) causes insulin resistance syndrome In 1992 Hales and Barker [23] proposed a novel hypothesis of thrifty phenotype in the pathogenesis of type 2 diabetes mellitus; they suggested that poor fetal and early post-natal nutrition imposed mechanisms of nutritional thrift upon a growing individual. The first evidence for the thrifty phenotype hypothesis came from the study by Barker and Osmond [24] of the differences in death rates from cardiovascular disease in different areas of UK and Wales. The death rate was found to be closely related to differences in neonatal mortality 70 or more years ago. Since most neonatal deaths were associated with low birth weight, this finding was thought to indicate that cardiovascular disease is linked to impaired fetal growth or malnutrition in early life. This link was further demonstrated for ischemic heart disease [25], raised blood pressure [26], elevated plasma levels of fibrinogen [27], glucose intolerance [28] and raised serum cholesterol [29], thus encompassing all the features of insulin resistance syndrome. Nutritional thrift programmed early in life is detrimental at times of adequate or excessive nutrition, and can explain why Ethiopian Jews transported to Israel [30], and Nauruan islanders who suffered severe malnutrition during World War II show a high prevalence of diabetes mellitus. The same mechanism could explain also why it has declined recently among Nauruans [31]. Nutrient supply to the fetus is the major regulator of fetal growth, and in experimental animals, maternal undernutrition during pregnancy commonly leads to the birth of the low-weight offspring. However, birth weight is a poor measure of fetal growth, and using a simple chronically implanted measuring device, it was possible to monitor fetal length or girth on a daily basis [32]. The rate of fetal growth usually slows down within 2–3 days of the start of maternal undernu-
H.K. Lee / Diabetes Research and Clinical Practice 45 (1999) 127–135
trition, but resumes promptly after refeeding [33]. However, growth of fetuses that were already growing only slowly before diet restriction, does not slow down further due to undernutrition. Harding and Johnston [34] hypothesized that this slow growth rate resulted from earlier programming, affected by malnutrition. The relation between fetal malnutrition and the insulin resistance syndrome has also been demonstrated in animal models. Young rats fed low protein diet from 3 to 6 weeks of age exhibit impaired insulin secretion and decreased pancreatic beta cell mass, which do not recover for at least 12 weeks [35,36]. Fetal sheep growing slowly during late gestation were found to have a relatively high insulin:glucose ratio, suggesting the presence of insulin resistance in these animals [37]. Since slow growth during late gestation is programmed by nutritional status during early pregnancy, it can be stated that both insulin secretion capacity and insulin sensitivity are programmed in early life. In other animals, such as guinea pigs and rats, those in which growth is retarded inutero show elevated blood pressure at 3 – 4 months of age [38] and those exposed to a protein deficient diet in-utero are hypertensive when they become adult [39].
5. Inheritance pattern of thrifty phenotype With regard to the effect of nutrition on health or thrifty phenotype, the most important piece of information gathered from these animal studies, and some clinical observations, might be the fact that the programming effects of malnutrition in early pregnancy persist for more than one generation. This implies a kind of Lamarckian inheritance, not usual Mendellian inheritance. Zamenhof et al. [40] were the first to observe that rats exposed to undernutrition as fetuses gave birth to offspring which themselves had reduced brain weight and cell numbers. Furthermore, in a rat colony marginally undernourished for nine generations, nutritional improvement during the last week of pregnancy was found to lead to the birth of offspring which were larger than those of well-fed controls. The full correction of abnormal
131
adult body size and behavioral abnormalities required three generations of nutritional rehabilitation [41]. Similar observations were made in humans. Lumey [42] found that women exposed to famine as fetuses in the first half of pregnancy, despite being of normal birth weight themselves, gave birth to offspring with an increased prevalence of growth retardation. Ravelli et al. [43] reported that young men exposed to famine in-utero during late gestation were relatively protected from later obesity, while those exposed early in pregnancy showed an increased prevalence of obesity in adults. In Jamaica, Godfrey et al. [44] found a strong association between thin maternal triceps skinfold thickness at 15 weeks of gestation and raised blood pressure in their offspring, and lower weight gain between 15 and 35 weeks of gestation has been independently associated with raised blood pressure in children. These results suggest that women who became pregnant while in a poor nutritional state and who are programmed to show poor weight gain during pregnancy, will give birth to offspring who tends to have higher blood pressure. This kind of inheritance cannot be explained in the usual Mendellian fashion, but suggests an epigenetic mechanism, such as genetic imprinting. Imprinting of genes by methylation of cytosine residue of DNA occurs in many places, such as in those genes coding Xist, H19, IGF2R and insulin [45,46]. It is suspected that similar imprinting mechanism(s) might operate in this case, possibly involving the gene coding mitochondrial transcription factor A (Tfam), since cellular mtDNA depletion is known to be associated with a low level of Tfam [47]. If birth weight is correlated with insulin resistance syndrome, and mtDNA is a marker of this syndrome, there should be a certain relation between mtDNA content and birth weight. To find if there is any relationship, 27 singleton baby– mother pairs born during early 1997 in Seoul National University Hospital, Seoul, Korea were studied [48]. Maternal mtDNA content in peripheral blood correlated strongly with infant birth weight, and the prepregnancy body weight of mother, cord blood IGF1 level, cord blood
132
H.K. Lee / Diabetes Research and Clinical Practice 45 (1999) 127–135
IGFBP-3 levels also showed statistically significant correlation. In the multiple regression analysis, however, only maternal mtDNA remained significant. Cord blood mtDNA content did not correlate linearly with birth weight, although it correlated positively with maternal mtDNA content. Further studies with a larger number of subjects are needed before reaching a firm conclusion between birth weight and mtDNA of the baby. However, it is worthwhile to note that there is a quantitative relation in mitochondrial inheritance, and this finding fits well with the above mentioned inheritance pattern of thrifty phenotype.
6. Mitochondrial biogenesis in ovum The inheritance of mitochondrial and nuclear genes is fundamentally different, in that mitochondrial gene copy number per cell usually exceeds 1000, replication occurs throughout the cell cycles, and its origin is strict maternal. Even though a small amount of sperm mitochondria are transmitted during fertilization, paternal mtDNA is eliminated early in development by active degradation [49]. The details of mitochondrial biogenesis in oocyte development and gametogenesis remain largely unknown; most existing information was obtained by Hauswirth et al. in the 1980s [50,51]. They showed that in bovine oocytes, mtDNA copy number increased 50 – 200-fold from that of resting germ cell line, which seems to have a similar copy number to somatic cells. Between a primary germ cell line and the largest non-zone pellucida encased fetal oocyte, the increase in mtDNA copy number is at least 3-fold. The same amount of mtDNA is found in primordial follicles in the adult ovary, suggesting that these cells are dormant. From this point and the development of an intact zona pellucida (class II oocytes), further amplification occurs (at least 15-fold). In the final stages of oocyte development, when the amount of cytoplasm increases by at least 6-fold, less than 2-fold increase in mtDNA occurs. MtDNA amplification is essentially complete by the time the zona pellucida is added to a developing oocyte. Mature
mammalian oocytes contain 4–7 pg of mtDNA or 200 000 mtDNA molecules. Along with the increase in mtDNA copy numbers, oocyte cytoplasmic volume increases 240fold, and the number of organelles and the total volume of mitochondria increase proportionally. During rapid expansion of cytoplasmic volume and mitochondrial number, the ratio of mitochondrial to cytoplasmic volume therefore remains constant. These findings suggest that mitochondrial biogenesis is directly linked to developmental increases in cell volume, and imply that to be viable, an oocyte requires a fixed amount of functional mitochondrial surface per unit volume of cytoplasm. A somatic cell contains about ten copies of mtDNA per mitochondrion, as does the primordial oocyte. When an oocyte becomes the smallest class of zone-encased oocytes, there is a rapid reduction in the copy number per mitochondrion, termed mitochondrial meiosis, to about one or 1.5 genomes per daughter mitochondrion. The phenomenon is comparable to that of bacterial replication in fission. This low copy number does not recover significantly during the later stages of oocyte development, rising only to two copies per organelle. No additional mtDNA synthesis occurs until after the 64-cell blastocyst stage after fertilization, and repopulation of each organelle does not occur until after early cell divisions of the embryo. Although few studies have been performed it is likely that some changes occurring during oocyte development persist during embryogenesis and possibly into fetal and adult life, although this has not been well studied. If decreased mtDNA status of the body represents, de facto, the thrifty phenotype, this phenotype could result from this very complex process of mitochondrial biogenesis. MtDNA exists mostly in circular and monomeric forms, but it is known that it can be dimeric or multimeric. Michaels et al. [51] observed that 30–40% of bovine oocyte mtDNA exists in multimeric form, much more than in mouse oocytes, as reported by Piko and Matsumoto [52]. It is not known if an egg with many multimeric forms of mtDNA is fertile, nor what would happen to this multimeric form of mtDNA
H.K. Lee / Diabetes Research and Clinical Practice 45 (1999) 127–135
during mitochondrial meiosis. Because of highly competitive selection of wild-type mtDNA during oogenesis and embryogenesis (only 1 in 1000 is chosen), it is not likely that much of the multimeric form survives.
7. Conclusion and suggestion Further studies are needed to confirm the role of the quantitative mitochondrial paradigm of insulin resistance syndrome pathogenesis. However, it matches well with the concept that there is common soil for this syndrome [53], and the suggestion that the underlying abnormality is a genetic program to fail, for the beta cell to compensate for insulin resistance [54]. Mitochondria are energy-supplying cytoplasmic soil and mtDNA is quite vulnerable to free radicals and fails easily. It is believed that the quantitative mitochondrial paradigm proposed here (Fig. 1) complements the already established qualitative mitochondrial paradigm of diabetogenesis [6]. This paradigm does not contradict either the thrifty phenotype hypothesis [23], supported by considerable epidemiological and experimental evidence, or the thrifty genotype hypothesis, proposed by Neel [55]. Furthermore, it suggests a novel biologic mechanism, that a bodily set-point of mtDNA content exists, that is changes according to the nutritional status of fetal or early development, and that it is inherited. As mammalian mtDNA undergoes enormous amplification and selection during oogenesis and
Fig. 1. Relations between the concepts of low mitochondrial status, insulin resistance, thrifty phenotype and genotype. Insulin resistance state is considered as an adaptive state of low mitochondrial functional capacity induced by malnutrition, which require more insulin or hyperglycemia to overcome mitochondrial energy supply, Tfam, mitochondrial transcription factor A; mtSSBP, mitochondrial single stranded DNA binding protein; POL-r, DNA polymerase gamma.
133
subsequent embryogenesis, it is also possible that abnormal biogenesis of mtDNA could result in an unfavorable environment. Cellular mitochondrial status is determined by nuclear factors, such as Tfam, mitochondrial single stranded DNA binding protein and RNA polymerase gamma [56,57]. Since inheritance of the thrifty phenotype shows characteristics of genetic imprinting, we suspect that the gene coding these genes might be differently methylated during malnutrition and a lower cellular mitochondrial set point. Much is known about mtDNA replication and its control, but to understand the detailed mechanisms of mitochondrial biogenesis and the implications of this poor development, further studies are needed. References [1] G.M. Reaven, Banting lecture 1988. Role of insulin resistance in human disease, Diabetes 37 (1988) 1595 – 1607. [2] I.F. Godsland, J.C. Stevenson, Insulin resistance: syndrome or tendency?, Lancet 346 (1995) 100 – 103. [3] A. Garg, S.M. Haffner, Insulin resistance and atherosclerosis. An overview, Diab. Care 19 (1996) 274. [4] D.E. Moller, C. Bjobek, A. Vidal-Puig, Candidate genes for insulin resistance, Diab. Care 19 (1996) 396 – 400. [5] D.C. Wallace, Diseases of the mitochondrial DNA, Annu. Rev. Biochem. 61 (1992) 1175 – 1212. [6] K.D. Gerbitz, K. Gempel, D. Brdiczka, Mitochondria and diabetes. Genetic, biochemical, and clinical implications of the cellular energy circuit, Diabetes 45 (1996) 113 – 126. [7] K.D. Gerbitz, J.M.W. van den Ouweland, J.A. Massen, M. Jaksch, Mitochondrial diabetes mellitus: a review, Biochim. Biophys. Acta 1271 (1995) 253 – 260. [8] J.A. Maassen, T. Kadowaki, Maternally inherited diabetes and deafness: a new diabetes subtype, Diabetologia 39 (1996) 375 – 382. [9] G. Velho, M.M. Byrne, K. Clement, J. Sturis, M.E. Pueyo, H. Blanche, N. Vionet, J. Fiet, P. Passa, J.J. Robert, K.S. Polonsky, P. Froguel, Clinical phenotype, insulin secretion, and insulin sensitivity in kindreds with maternally inherited diabetes and deafness due to mitochondrial tRNA Leu(UUR) gene mutation, Diabetes 45 (1996) 478 – 487. [10] M.P. King, G. Attardi, Human cells lacking mtDNA: repopulation with exogenous mitochondria by complementation, Science 246 (1989) 500 – 503. [11] J.M.W. van den Ouweland. A New Subypte of Non-Insulin Dependent Diabetes Mellitus Is Associated With a Mitochondrial Gene Mutation. PhD Thesis. Netherlands, University of Leiden, 1994 (cited from [6]).
134
H.K. Lee / Diabetes Research and Clinical Practice 45 (1999) 127–135
[12] A. Soejima, K. Inoue, D. Taka, M. Kaneko, H. Ishihara, Y. Oka, J.-I. Hayashi, Mitochondrial DNA is required for the regulations of glucose-stimulated insulin secretion in a mouse pancreatic beta cell line, MIN6, J. Biol. Chem. 42 (1996) 26194 – 26199. [13] J. Poulton, M.S. Brown, A. Cooper, D.R. Marchington, D.I.W. Phillips, A common mitochondrial DNA variant is associated insulin resistance in adult life, Diabetologia 41 (1998) 54 – 58. [14] D.L. Rothman, I. Magnusson, G. Cline, D. Gerard, C.R. Kahn, R.G. Shulman, G.I. Shulman, Decreased muscle glucose transport/phosphorylation is an early defect in the pathogenesis of non-insulin-dependent diabetes mellitus, Proc. Natl. Acad. Sci. USA 92 (1995) 983–987. [15] C.S. Shin, Quantitative abnormalities of mitochondrial DNA in noninsulin dependent diabetes mellitus, Kor. J. Diab. Ass. 18 (1994) 344–350. [16] D.A. Antonetti, C. Reynet, C.R. Kahn, Increased expression of mitochondrial-encoded genes in skeletal muscle of humans with diabetes mellitus, J. Clin. Invest. 95 (1995) 1383 – 1388. [17] F. Cottrez, C. Auriault, A. Capron, H. Groux, Quantitative PCR: validation of the use of a multispecific internal control, Nucleic Acids Res. 22 (1994) 2712–2713. [18] H. Wang, L. Fliegel, C.E. Cass, A.M.W. Penn, P.M. Michalak, J.H. Weiner, B.D. Lemire, Quantitation of mitochondrial DNA in heteroplasmic fibroblasts with competitive PCR, BioTechniques 17 (1994) 76–78. [19] C.S. Shin, H.K. Lee, C.-S. Koh, Y.I. Kim, Y.S. Shin, K.Y. Yoo, H.Y. Paik, Y.S. Park, B.K. Yang, Risk factors for the development of NIDDM in Yonchon county, Korea, Diab. Care 20 (1997) 1842–1846. [20] H.K. Lee, J.H. Song, C.S. Shin, D.J. Park, K.S. Park, K.U. Lee, C.S. Koh, Peripheral blood mitochondrial DNA content can predict the development of diabetes (Abstract), Diabetes 46 (Suppl 1) (1997) 175A. [21] B. Obermaier-Kusser, J. Muller-Hocker, I. Nelson, P. Lestienne, C. Enter, T. Riedele, K.D. Gerbitz, Different copy numbers of apparently identically deleted mitochondrial DNA in tissues from a patient with Kearnes–Sayer syndrome detected by PCR, Biochem. Biophys. Res. Commun. 169 (1995) 1007–1015. [22] H.K. Lee, K.-U. Lee, J.H. Song, J.Y. Park, C.S. Shin, K.S. Park, D.J. Park. Peripheral mitochondrial DNA content correlated with insulin secretion in young Koreans (Abstract). International Diabetes Epidemiology Symposium, Savonlinna, Finland, 1997. [23] C.N. Hales, D.J. Barker, Type 2 (non-insulin-dependent) diabetes mellitus: the thrifty phenotype hypothesis, Diabetologia 35 (1992) 595–601. [24] D.J.P. Barker, C. Osmond, Infant mortality, childhood nutrition and ischemic heart disease in England and Wales, Lancet 1 (1986) 1077–1081. [25] D.J. Barker, P.D. Winter, C. Osmond, B. Margetts, S.J. Simmonds, Weight in infancy and death from ischaemic heart disease, Lancet 2 (1989) 577–580.
[26] D.J. Barker, A.R. Bull, C. Osmond, S.J. Simmonds, Fetal and placental size and risk of hypertension in adult life, Br. Med. J. 301 (1990) 259 – 262. [27] D.J. Barker, T.V. Meade, C.H.D. Fall, A. Lee, C. Osmond, K. Phipps, Y. Stirling, The relation of fetal and infant growth to plasma fibrinogen and factor VII levels in adult life, Br. Med. J. 304 (1992) 148 – 152. [28] C.N. Hales, D.J. Barker, P.M.S. Clark, L.J. Cox, C. Fall, C. Osmond, P.D. Winter, Fetal and infant growth and impaired glucose tolerance at age 64 years, Br. Med. J. 303 (1991) 1019 – 1022. [29] D.J. Barker, C.N. Martyn, C. Osmond, C.H. Fall, Growth in utero and serum cholesterol concentrations in adult life, Br. Med. J. 307 (1993) 1524 – 1527. [30] M.P. Cohen, E. Stern, Y. Rusecki, A. Zeidler, High prevalence of diabetes in young adult Ethiopian immigrants to Israel, Diabetes 37 (1988) 824 – 828. [31] G.K. Dowse, P.Z. Zimmet, C.F. Finch, V.R. Collins, Decline in incidence of epidemic glucose intolerance in Nauruans: implications for the thrifty genotype, Am. J. Epidemiol. 133 (1991) 1093 – 1104. [32] D.J. Mellor, I.C. Matheson, Chronic catheterisation of the aorta and umbilical vessels of fetal sheep, Res. Vet. Sci. 18 (1975) 221 – 223. [33] D.J. Mellor, I.C. Matheson, Daily changes in the curved crown-rump length of individual sheep fetuses during the last 60 days of pregnancy and effects of different levels of maternal nutrition, Q. J. Exp. Physiol. Cog. Med. Sci. 64 (1979) 119 – 131. [34] J.E. Harding, B.M. Johnston, Nutrition and fetal growth, Reprod. Fertil. Dev. 7 (1995) 539 – 547. [35] I. Swenne, C.J. Crace, R.D.G. Milner, Persistent impairment of insulin secretory response to glucose in adult rats after limited period of protein caloriemalnutrition early in life, Diabetes 36 (1987) 454 – 458. [36] I. Swenne, L.A.H. Borg, C.J. Crace, A.S. Landstrom, Persistent reduction of pancreatic beta-cell mass after a limited period of protein-energy malnutrition in the young rat, Diabetologia 35 (1992) 939 – 945. [37] P.D. Gluckman, J.E. Harding, The physiology and pathophysiology of intrauterine growth retardation, Horm. Res. 48 (Suppl 1) (1997) 11 – 16. [38] E. Persson, T. Jansson, Low birth weight is associated with elevated adult blood pressure in the chronically catheterized guinea pig, Acta Physiol. Scand. 145 (1992) 195 – 196. [39] S.M. Woodall, B.M. Johnston, B.H. Breier, P.D. Gluckman, Chronic maternal undernutrition in the rat leads to delayed postnatal growth and elevated blood pressure of offspring, Pediatr. Res. 40 (1996) 438 – 443. [40] S. Zamenhof, E. van Marthens, L. Grauel, DNA (cell number) in neonatal brain: second generation (F2) alteration by maternal (Fo) dietary protein restriction, Science 172 (1971) 850 – 851. [41] R.J. Stewart, H. Sheppard, R. Preece, J.C. Waterlow, The effect of rehabilitation at different stages of development of rats marginally malnourished for ten to twelve generations, Br. J. Nutr. 43 (1980) 403 – 412.
H.K. Lee / Diabetes Research and Clinical Practice 45 (1999) 127–135 [42] L.H. Lumey, Decreased birth weights in infants after maternal in-utero exposure to the Dutch famine of 1944– 1945, Paediatr. Perinat. Epidemiol. 6 (1992) 240. [43] G.P. Ravelli, Z.A. Stein, M.W. Susser, Obesity in young men after famine exposure in utero and early infancy, N. Engl. J. Med. 295 (1976) 349–353. [44] K.M. Godfrey, T. Forrester, D.J. Barker, A.A. Jackson, J.P. Landman, J.S.E. Hall, V. Cox, C. Osmond, Maternal nutritional status in pregnancy and blood pressure in childhood, Br. J. Obstet. Gynaecol. 101 (1994) 398–403. [45] M.S. Bartolomei, S.M. Tilghman, Genomic imprinting in mammals, Annu. Rev. Genet. 31 (1997) 493–525. [46] M.A. Surani, Imprinting and the initiation of gene silencing in the germ line, Cell 93 (1988) 309–312. [47] N.G. Larsson, A. Oldfors, E. Holme, D.A. Clayton, Low levels of mitochondrial transcription factor A in mitochondrial depletion, Biochem. Biophys. Res. Commun. 200 (1994) 1374 – 1381. [48] K.S. Park, Y.Y. Lee, J.H. Song, J.J. Koh, C.S. Shin, D.J. Park, B.H. Yoon, J.K. Jun, H.K. Lee, The relationship of birth weight to insulin-like growth factors (IGF) and the amount of mitochondrial DNA in umbilical cord blood (Abstract), Diabetes 47 (Suppl 1) (1998) A321. [49] H. Kaneda, J. Hayashi, S. Takahama, C. Taya, K.F. Lindahl, H. Yonekawa, Elimination of paternal mitochondrial DNA in intraspecific crosses during early mouse embryogenesis, Proc. Natl. Acad. Sci. USA 92 (1995) 4542 – 4546.
.
135
[50] W.W. Hauswirth, P.J. Laipis, Transmission genetics of mammalian mitochondria: a molecular model and experimental evidence, in: E. Quagliariello, et al. (Eds.), Achievements and Perspectives of Mitochondrial Research: Biogenesis, vol. 2, Elsevier, Amsterdam, 1985. [51] G.S. Michaels, W.W. Hauswirth, P.J. Laipis, Mitochondrial copy number in bovine oocytes and somatic cells, Dev. Biol. 94 (1982) 246 – 251. [52] L. Piko, L. Matsumoto, Number of mitochondria and some properties of mitochondrial DNA in the mouse egg, Dev. Biol. 49 (1976) 1 – 10. [53] M.P. Stern, Diabetes and cardiovascular disease: the common soil hypothesis, Diabetes 44 (1995) 369 – 374. [54] J.L. Leahy, Impaired b-cell function with chronic hyperglycemia: overworked b-cell hypothesis, Diab. Rev. 4 (1996) 298 – 319. [55] J.V. Neel, The thrifty genotype revisited, in: J. Kobberling, R. Tattersal (Eds.), The Genetics of Diabetes Mellitus, Academic Press, New York, 1982. [56] R.N. Lightowlers, P.F. Chinnery, D.M. Turnball, N. Howell, Mammalian mitochondrial genetics: heredity, heteroplasmy and disease, Trends Genet. 13 (1997) 450 – 455. [57] R.A. Schultz, S.J. Swoap, L.D. McDaniel, B. Zhang, E.C. Koon, D.J. Garry, K. Li, R.S. Williams, Differential expression of mitochondrial DNA replication factors in mammalian tissues, J. Biol. Chem. 273 (1998) 3447 – 3451.