Abstract
The mitochondria are the major cellular site of energy production and respiration. Recent research has focused on investigating the role of mitochondria in disease development and it has become increasingly evident that mitochondrial dysfunction contributes to a variety of human diseases. Mitochondrial DNA (mtDNA) quantity is very important for maintaining mitochondrial function and meeting the energy needs of the body. We have measured mitochondrial content in 1259 Mexican American individuals (from 42 extended families) and have shown that mtDNA quantity (a surrogate measure of mitochondrial integrity) has a large genetic component. We performed a genome scan and a genome-wide quantitative transcriptomic scan to identify QTLs influencing mitochondrial content. A variance components linkage-based genome scan utilizing 439 STR markers was used to localize a QTL for mitochondrial content on chromosome 10q (LOD = 3.83). Significant linkage to the mitochondrial genome was also detected for mitochondrial transmission (LOD = 3.39). For replication, we measured mitochondrial content in an independent Caucasian population (1088 individuals) finding evidence for linkage in these same regions. As part of the San Antonio Family Heart Study, we obtained genome-wide quantitative transcriptional profiles from 1240 individuals. Using lymphocyte samples, we quantitated 20 413 transcripts and examined correlations between the expression levels of these transcripts and mitochondrial content using the variance components method. Using regression analysis allowing for residual genetic components, we identified 829 transcripts (including many novel genes) influencing mitochondrial content that vary in their general biological actions, from cell signaling to cell trafficking and ion binding.
INTRODUCTION
The primary function of the mitochondria is cellular respiration and energy metabolism for intracellular metabolic pathways including the Krebs cycle, B-oxidation and cholesterol synthesis. They are also involved in cell signaling for apoptosis ( 1 , 2 ). The mitochondrial genome (mtDNA) is vital for maintaining normal mitochondrial function and energy production for the body. The 16.6 kb genome encodes 37 genes: 22 tRNAs, 2 rRNAs and 13 proteins involved in oxidative phosphorylation. mtDNA encodes only a small number of mitochondrial functioning proteins; however, most of the proteins found in the mitochondria are encoded in the nucleus, synthesized in the cytosol and transported into the mitochondria ( 1 , 2 ).
The mitochondria are double membrane bound organelles containing four sub-compartments: the outer membrane, the intermembrane space, the inner membrane and the matrix. These mitochondrial membranes contain complexes that are responsible for the importing, sorting and assembly of nuclear encoded proteins into the mitochondria ( 3 ). The density of cellular mitochondria varies according to tissue type, with tissues that require energy through oxidative phosphorylation having the highest density, e.g. skeletal muscle ( 1 ).
Variations within the mitochondrial genome contribute to disease susceptibility. Additionally, the quantity of mtDNA is very important for normal mitochondrial function. It has become increasingly evident that mitochondrial dysfunction contributes to diabetes as well as a variety of other human disorders that include cancer, obesity, multiple sclerosis, several psychiatric disorders and a wide range of age-related disorders ( 4–9 ). Mitochondrial oxidative stress appears to modulate the rate at which mitochondrial and mtDNA damage occurs in cells, with the degree of damage being central to the development of age-related metabolic and degenerative diseases, aging and cancer ( 5 ).
Given the important role of mitochondria in several metabolic pathways, much of the recent attention has been focused on studies investigating the role of mitochondrial dysfunction in the development of insulin resistance and type 2 diabetes. Results have shown reductions in bioenergetic capacity of the mitochondria in insulin resistance and type 2 diabetes ( 10–14 ). In individuals with a family history of type 2 diabetes, a reduction in the number and size of mitochondria in insulin-resistant skeletal muscle has been observed ( 11 , 12 ). Evidence also indicates that a decline in mitochondrial function with age contributes to changes in disease susceptibility in the elderly ( 13 ). mtDNA copy number has been seen to decrease by up to 50% in skeletal muscle of type 2 diabetics ( 15 ) and is lower in offspring of type 2 diabetic patients when compared with age, sex and BMI matched controls without a family history of disease ( 16 ).
Variations in both mtDNA and mitochondrial dysfunction contribute to multiple human diseases. New mutations are introduced to the mitochondrial genome at a rate 5–10 times that of the nuclear genome, resulting in a greater proportion of mitochondrial dysfunction being due to mtDNA mutations ( 17 ). This makes mitochondrial-related diseases among the most common of genetic disorders today and imposes a major burden on society.
RESULTS
Heritability of mitochondrial content
Using the variance component-based quantitative genetic methods available in our computer package, SOLAR, we estimated the heritability of mitochondrial content to be 0.329 ( P = 4.7 × 10 −20 ) in our Mexican American families. All analyses initially allowed for the effects of sex and age. Mitochondrial content was also seen to decrease significantly with age ( P = 0.004) as expected ( 5 ). Age variation in the sample accounted for ∼2% of the total phenotypic variation in this phenotype. Additionally, there was no marginal effect of sex on mitochondrial content in this sample. To our knowledge, this is the first estimate of human mitochondrial content and indicates a substantial genetic component to this trait. This estimate suggests that it is reasonable to search for the genes that are responsible for the observed genetic variance in this trait.
An autosomal QTL influencing mitochondrial content
Using our genome scan data, we searched for QTLs influencing mitochondrial content. Variance component-based quantitative trait linkage analysis, implemented in SOLAR, was used to perform a genome scan utilizing 439 autosomal short tandem repeat (STR) polymorphisms. Because the mitochondrial content phenotype was normally distributed (kurtosis = −0.03), we employed the standard version of variance component analysis under a multivariate normal model. We have previously shown that this conformation of a quantitative trait to normality is sufficient to yield the usual asymptotic distribution of the linkage test statistic ( 18 ). Using this approach, we localized a single nuclear genome QTL for mitochondrial content to chromosome 10q11 at 72 cM (LOD = 3.83, nominal P = 1.3 × 10 −5 , genome-wide P = 0.005) very near the STR marker D10S1220. The 1-LOD support interval covers a region of ∼24 Mb. Figure 1 shows the LOD function for the chromosome 10q QTL. An additional suggestive linkage (LOD = 2.18, nominal P = 0.0008, genome-wide P = 0.300) was found on chromosome 8 at 41 cM. No other region in the nuclear genome provided evidence for linkage.
Localization of chromosome 10q QTL for mitochondrial content.
Localization of chromosome 10q QTL for mitochondrial content.
A mitochondrial QTL influencing mitochondrial content
Because it is reasonable to assume that mitochondrial content may also be influenced by variation in the mtDNA, we tested for potential linkage of this phenotype to the mitochondrial genome. The effective lack of recombination in the mitochondrial genome makes it possible to test this hypothesis without typing mitochondrial genetic variants, given that sufficient pedigree generational depth and complexity is available. The Mexican American families have substantial power for detecting mitochondrial effects. In the sample, there are 299 mitochondrial lineages represented. Within these lineages, there are 6289 pairs of individuals. Because the correlation between mitochondrial lineage members is effectively constant (except for mutational variance), this provides outstanding contrast with models of autosomal (or sex-linked) transmission. We used the variance component methods implemented in SOLAR to test for significant mitochondrial effects ( 19 , 20 ). Significant linkage (LOD = 3.39, P = 3.9 × 10 −5 ) was detected for mitochondrial transmission. This significant linkage is maintained even when a ‘two-locus’ model that includes the autosomal chromosome 10q QTL is evaluated.
Replication of QTL linkages in an independent sample
Similar to our original results in the Mexican American samples, mitochondrial content was significantly heritable ( h2 = 0.52, P = 5.5 × 10 −28 ). In this population of European ancestry, mitochondrial content significantly decreased with age ( P = 0.016) in a manner similar to that seen in the Mexican American sample. However, in this sample, we also detected significantly increased mitochondrial content in females ( P = 0.0007). However, both age and sex combined accounted for only 3.2% of the total phenotypic variation in mitochondrial content. Using available genetic marker data, we tested directly for evidence of linkage at the region on chromosome 10q observed in Mexican Americans. Quantitative trait linkage analysis confirmed the presence of a QTL ( P = 0.0022) on chromosome 10q within 3 cM of the peak in Mexican Americans. Similarly, we observed evidence for mitochondrial transmission of mitochondrial content in this sample ( P = 0.0275). Combining this linkage information between the two independent studies provides a LOD score of 5.0 ( P = 8 × 10 −7 ) for the chromosome 10q QTL and a combined LOD of 3.7 ( P = 2 × 10 −5 ) for linkage to the mitochondrial genome.
Transcriptomic variation and mitochondrial content
Genome scanning using linkage or association, while useful for identifying relatively large genetic effects, may miss important genetic determinants or interactants. As another independent investigation into genes that may play a role in influencing mitochondrial content, we tested for correlations of gene expression in lymphocytes with mitochondrial content, using the genome-wide quantitative transcriptional profile data set. Using standard regression methods (employed in SOLAR) and allowing for non-independence among related family members, we identified 829 transcripts (many novel) that potentially correlate with mitochondrial content [nominal P -value < 0.01, false discovery rate (FDR) = 0.25]. This FDR was chosen in advance of analysis as a compromise between finding as many novel transcriptional correlates with mitochondrial content as possible but limiting the relative percentage of false discoveries to a rate not exceeding 1 in 4. These genes vary in their general biological actions, from cell signaling to cell trafficking and ion binding. Table 1 shows the genes exhibiting the strongest relationships with mitochondrial content.
Transcripts displaying the strongest correlations with mitochondrial content
| Gene ID | P -value | Location | Putative function |
|---|---|---|---|
| NCOA1 | 7.4 × 10 −8 | 2p23 | Viral transactivation, DNA synthesis |
| TESC | 9.96 × 10 −7 | 12q24 | Ion binding, signal transduction |
| ZNF703 | 1.54 × 10 −6 | 8p12 | Ion binding |
| IMPDH1 | 4.45 × 10 −6 | 7q31 | Retinitis pigmentosa |
| hmm3998 | 7 × 10 −6 | 15 | Unknown |
| MAPK7 | 9.24 × 10 −6 | 17p11 | Cellular proliferation, morphogenesis |
| PTPN18 | 1.13 × 10 −5 | 2q21 | Cell signaling, protein interactions |
| C19orf10 | 1.15 × 10 −5 | 19p13 | Cellular proliferation, hypertrophy, eosinophilia |
| PLOD3 | 1.37 × 10 −5 | 7q22 | Amino acid hydroxylation |
| APOB48R | 1.38 × 10 −5 | 16p11 | Lipoprotein binding |
| CAPNS1 | 1.73 × 10 −5 | 19q13 | Cellular migration and movement |
| CDC42EP4 | 2.6 × 10 −5 | 17q24 | Cellular signaling |
| ARF1 | 2.98 × 10 −5 | 1q42 | Cellular formation, lipid metabolism |
| ELF4 | 3.29 × 10 −5 | Xq26 | Transcription factor |
| CXX1 | 3.62 × 10 −5 | Xq26 | Post-translational modification |
| FLII | 3.77 × 10 −5 | 17p11 | Cellular biogenesis, embryonic development |
| LOC389203 | 4.54 × 10 −5 | 4p15 | Unknown |
| ITGB2 | 5.55 × 10 −5 | 21q22 | Cellular signaling and interaction, function and maintenance |
| FXYD5 | 5.72 × 10 −5 | 19q12 | Ion channel binding |
| FEM1A | 6.06 × 10 −5 | 19p13 | Prostaglandin receptor |
| RAC1 | 6.97 × 10 −5 | 7p22 | Lipid metabolism, tumorigenesis, cell death |
| CDC14A | 8.14 × 10 −5 | 1p21 | Cell morphology |
| TCEB3 | 8.23 × 10 −5 | 1p36 | Cell mitosis |
| SLC30A9 | 8.36 × 10 −5 | 4p13 | Solute carriers, DNA excision repair |
| KIAA1109 | 8.47 × 10 −5 | 4q27 | Endopeptidase activity, proteolysis |
| RABEP1 | 8.58 × 10 −5 | 17p13 | GTPase binding protein |
| ADAMTS6 | 8.84 × 10 −5 | 5q12 | Peptidase activity, ion binding |
| PPGB | 9.77 × 10 −5 | 20q13 | Carboxypeptidase, intracellular protein transport |
| OAZ1 | 0.0001038 | 19p13 | Cell biosynthesis |
| ARHG | 0.0001107 | 11p15 | Cellular trafficking |
| Gene ID | P -value | Location | Putative function |
|---|---|---|---|
| NCOA1 | 7.4 × 10 −8 | 2p23 | Viral transactivation, DNA synthesis |
| TESC | 9.96 × 10 −7 | 12q24 | Ion binding, signal transduction |
| ZNF703 | 1.54 × 10 −6 | 8p12 | Ion binding |
| IMPDH1 | 4.45 × 10 −6 | 7q31 | Retinitis pigmentosa |
| hmm3998 | 7 × 10 −6 | 15 | Unknown |
| MAPK7 | 9.24 × 10 −6 | 17p11 | Cellular proliferation, morphogenesis |
| PTPN18 | 1.13 × 10 −5 | 2q21 | Cell signaling, protein interactions |
| C19orf10 | 1.15 × 10 −5 | 19p13 | Cellular proliferation, hypertrophy, eosinophilia |
| PLOD3 | 1.37 × 10 −5 | 7q22 | Amino acid hydroxylation |
| APOB48R | 1.38 × 10 −5 | 16p11 | Lipoprotein binding |
| CAPNS1 | 1.73 × 10 −5 | 19q13 | Cellular migration and movement |
| CDC42EP4 | 2.6 × 10 −5 | 17q24 | Cellular signaling |
| ARF1 | 2.98 × 10 −5 | 1q42 | Cellular formation, lipid metabolism |
| ELF4 | 3.29 × 10 −5 | Xq26 | Transcription factor |
| CXX1 | 3.62 × 10 −5 | Xq26 | Post-translational modification |
| FLII | 3.77 × 10 −5 | 17p11 | Cellular biogenesis, embryonic development |
| LOC389203 | 4.54 × 10 −5 | 4p15 | Unknown |
| ITGB2 | 5.55 × 10 −5 | 21q22 | Cellular signaling and interaction, function and maintenance |
| FXYD5 | 5.72 × 10 −5 | 19q12 | Ion channel binding |
| FEM1A | 6.06 × 10 −5 | 19p13 | Prostaglandin receptor |
| RAC1 | 6.97 × 10 −5 | 7p22 | Lipid metabolism, tumorigenesis, cell death |
| CDC14A | 8.14 × 10 −5 | 1p21 | Cell morphology |
| TCEB3 | 8.23 × 10 −5 | 1p36 | Cell mitosis |
| SLC30A9 | 8.36 × 10 −5 | 4p13 | Solute carriers, DNA excision repair |
| KIAA1109 | 8.47 × 10 −5 | 4q27 | Endopeptidase activity, proteolysis |
| RABEP1 | 8.58 × 10 −5 | 17p13 | GTPase binding protein |
| ADAMTS6 | 8.84 × 10 −5 | 5q12 | Peptidase activity, ion binding |
| PPGB | 9.77 × 10 −5 | 20q13 | Carboxypeptidase, intracellular protein transport |
| OAZ1 | 0.0001038 | 19p13 | Cell biosynthesis |
| ARHG | 0.0001107 | 11p15 | Cellular trafficking |
Transcripts displaying the strongest correlations with mitochondrial content
| Gene ID | P -value | Location | Putative function |
|---|---|---|---|
| NCOA1 | 7.4 × 10 −8 | 2p23 | Viral transactivation, DNA synthesis |
| TESC | 9.96 × 10 −7 | 12q24 | Ion binding, signal transduction |
| ZNF703 | 1.54 × 10 −6 | 8p12 | Ion binding |
| IMPDH1 | 4.45 × 10 −6 | 7q31 | Retinitis pigmentosa |
| hmm3998 | 7 × 10 −6 | 15 | Unknown |
| MAPK7 | 9.24 × 10 −6 | 17p11 | Cellular proliferation, morphogenesis |
| PTPN18 | 1.13 × 10 −5 | 2q21 | Cell signaling, protein interactions |
| C19orf10 | 1.15 × 10 −5 | 19p13 | Cellular proliferation, hypertrophy, eosinophilia |
| PLOD3 | 1.37 × 10 −5 | 7q22 | Amino acid hydroxylation |
| APOB48R | 1.38 × 10 −5 | 16p11 | Lipoprotein binding |
| CAPNS1 | 1.73 × 10 −5 | 19q13 | Cellular migration and movement |
| CDC42EP4 | 2.6 × 10 −5 | 17q24 | Cellular signaling |
| ARF1 | 2.98 × 10 −5 | 1q42 | Cellular formation, lipid metabolism |
| ELF4 | 3.29 × 10 −5 | Xq26 | Transcription factor |
| CXX1 | 3.62 × 10 −5 | Xq26 | Post-translational modification |
| FLII | 3.77 × 10 −5 | 17p11 | Cellular biogenesis, embryonic development |
| LOC389203 | 4.54 × 10 −5 | 4p15 | Unknown |
| ITGB2 | 5.55 × 10 −5 | 21q22 | Cellular signaling and interaction, function and maintenance |
| FXYD5 | 5.72 × 10 −5 | 19q12 | Ion channel binding |
| FEM1A | 6.06 × 10 −5 | 19p13 | Prostaglandin receptor |
| RAC1 | 6.97 × 10 −5 | 7p22 | Lipid metabolism, tumorigenesis, cell death |
| CDC14A | 8.14 × 10 −5 | 1p21 | Cell morphology |
| TCEB3 | 8.23 × 10 −5 | 1p36 | Cell mitosis |
| SLC30A9 | 8.36 × 10 −5 | 4p13 | Solute carriers, DNA excision repair |
| KIAA1109 | 8.47 × 10 −5 | 4q27 | Endopeptidase activity, proteolysis |
| RABEP1 | 8.58 × 10 −5 | 17p13 | GTPase binding protein |
| ADAMTS6 | 8.84 × 10 −5 | 5q12 | Peptidase activity, ion binding |
| PPGB | 9.77 × 10 −5 | 20q13 | Carboxypeptidase, intracellular protein transport |
| OAZ1 | 0.0001038 | 19p13 | Cell biosynthesis |
| ARHG | 0.0001107 | 11p15 | Cellular trafficking |
| Gene ID | P -value | Location | Putative function |
|---|---|---|---|
| NCOA1 | 7.4 × 10 −8 | 2p23 | Viral transactivation, DNA synthesis |
| TESC | 9.96 × 10 −7 | 12q24 | Ion binding, signal transduction |
| ZNF703 | 1.54 × 10 −6 | 8p12 | Ion binding |
| IMPDH1 | 4.45 × 10 −6 | 7q31 | Retinitis pigmentosa |
| hmm3998 | 7 × 10 −6 | 15 | Unknown |
| MAPK7 | 9.24 × 10 −6 | 17p11 | Cellular proliferation, morphogenesis |
| PTPN18 | 1.13 × 10 −5 | 2q21 | Cell signaling, protein interactions |
| C19orf10 | 1.15 × 10 −5 | 19p13 | Cellular proliferation, hypertrophy, eosinophilia |
| PLOD3 | 1.37 × 10 −5 | 7q22 | Amino acid hydroxylation |
| APOB48R | 1.38 × 10 −5 | 16p11 | Lipoprotein binding |
| CAPNS1 | 1.73 × 10 −5 | 19q13 | Cellular migration and movement |
| CDC42EP4 | 2.6 × 10 −5 | 17q24 | Cellular signaling |
| ARF1 | 2.98 × 10 −5 | 1q42 | Cellular formation, lipid metabolism |
| ELF4 | 3.29 × 10 −5 | Xq26 | Transcription factor |
| CXX1 | 3.62 × 10 −5 | Xq26 | Post-translational modification |
| FLII | 3.77 × 10 −5 | 17p11 | Cellular biogenesis, embryonic development |
| LOC389203 | 4.54 × 10 −5 | 4p15 | Unknown |
| ITGB2 | 5.55 × 10 −5 | 21q22 | Cellular signaling and interaction, function and maintenance |
| FXYD5 | 5.72 × 10 −5 | 19q12 | Ion channel binding |
| FEM1A | 6.06 × 10 −5 | 19p13 | Prostaglandin receptor |
| RAC1 | 6.97 × 10 −5 | 7p22 | Lipid metabolism, tumorigenesis, cell death |
| CDC14A | 8.14 × 10 −5 | 1p21 | Cell morphology |
| TCEB3 | 8.23 × 10 −5 | 1p36 | Cell mitosis |
| SLC30A9 | 8.36 × 10 −5 | 4p13 | Solute carriers, DNA excision repair |
| KIAA1109 | 8.47 × 10 −5 | 4q27 | Endopeptidase activity, proteolysis |
| RABEP1 | 8.58 × 10 −5 | 17p13 | GTPase binding protein |
| ADAMTS6 | 8.84 × 10 −5 | 5q12 | Peptidase activity, ion binding |
| PPGB | 9.77 × 10 −5 | 20q13 | Carboxypeptidase, intracellular protein transport |
| OAZ1 | 0.0001038 | 19p13 | Cell biosynthesis |
| ARHG | 0.0001107 | 11p15 | Cellular trafficking |
Pathway analysis of transcripts correlated with mitochondrial content
The 829 transcripts meeting the FDR of 0.25 criteria (nominal P < 0.01) were imported into Ingenuity Pathways Analysis (IPA) and matched based on their GenBank accession number. After excluding cases of multiple significant transcripts within a single gene, there were 689 unique genes identified by IPA from the list of 829 significant transcripts. There were 140 transcripts that were not identified by IPA, predominantly because their GenBank ID had been retired or corresponded to a hypothetical protein not yet described in the literature. Of the identified transcripts, 453 genes were eligible for network analysis in IPA based on available connectivity information from the published literature. A total of 423 of the mapped genes included information on functions and/or canonical pathways from the literature.
There were 44 significant functional categories ( P < 0.05) and 27 highly significant functional categories ( P < 0.01). The 27 highly significant categories are shown in Figure 2 . The 10 most significant functional categories were: cancer, cell cycle, cell death, reproductive system disease, respiratory disease, cell-to-cell signaling and interaction, cellular movement, connective tissue development and function, cell signaling and cellular function and maintenance.
The 27 most significant functional categories from IPA. The significance threshold shown in yellow represents P > 0.05. The set of functions shown below represent P < 0.01.
The 27 most significant functional categories from IPA. The significance threshold shown in yellow represents P > 0.05. The set of functions shown below represent P < 0.01.
Of our set of significant genes, there were 11 genes involved in cell senescence ( P = 0.0096). Of this set, ARID3A , BRCA1 , CDKN1A , CDKN1C , CEBPB , DUSP1 , GPI , JUNB , RAC1 and RAF1 were up-regulated while NPM1 was down-regulated. There were 11 genes ( ABL1 , CAPNS1 , LRP , MRA , PRKACA , PRKCD , RAC1 , RAF1 , RAPGEF1 , STAT3 , TIMP2 ) involved in the migration for fibroblasts ( P = 0.0023) that were all up-regulated.
There were 23 significant canonical pathway categories ( P < 0.05) and 12 highly significant canonical pathway categories ( P < 0.01). The 12 highly significant categories are shown in Figure 3 , and includes PTEN signaling, integrin signaling and ERK/MAPK signaling. There were 18 genes correlating with mitochondrial content involved in the integrin signaling canonical pathway that were all significantly upregulated ( P = 6.5 × 10 −5 ; Table 2 ). The integrin signaling gene network is shown in Figure 4 .
The set of significant canonical pathways from IPA. The significance threshold shown in yellow represents P > 0.05.
The set of significant canonical pathways from IPA. The significance threshold shown in yellow represents P > 0.05.
The integrin signaling gene network. Genes corresponding to significant transcripts correlated with mitochondrial content that are up-regulated are filled in red. The intensity of the red denoted the degree of up-regulation. Genes involved in integrin signaling are outlined in aqua.
The integrin signaling gene network. Genes corresponding to significant transcripts correlated with mitochondrial content that are up-regulated are filled in red. The intensity of the red denoted the degree of up-regulation. Genes involved in integrin signaling are outlined in aqua.
Up-regulated genes involved in the integrin signaling canonical pathway
| Name | Description | GenBank ID | Correlation coefficient | Location | Family |
|---|---|---|---|---|---|
| ABL1 | v-abl Abelson murine leukemia viral oncogene homolog 1 | NM_007313.1 | 0.081 | Nucleus | Kinase |
| ACTB | Actin, beta | NM_001101.2 | 0.098 | Cytoplasm | Other |
| ARF1 | ADP-ribosylation factor 1 | NM_001658.2 | 0.126 | Cytoplasm | Transporter |
| ARPC1A | Actin-related protein 2/3 complex, subunit 1A, 41 kDa | NM_006409.2 | 0.101 | Cytoplasm | Other |
| ARPC1B | Actin-related protein 2/3 complex, subunit 1B, 41 kDa | NM_005720.2 | 0.091 | Cytoplasm | Other |
| CAPN1 | Calpain 1, (µ/I) large subunit | NM_005186.2 | 0.097 | Cytoplasm | Peptidase |
| FYN | FYN oncogene related to SRC, FGR, YES | NM_153048.1 | 0.106 | Plasma membrane | Kinase |
| GRB2 | Growth factor receptor-bound protein 2 | NM_002086.3 | 0.090 | Plasma membrane | Other |
| ILK | Integrin-linked kinase | NM_004517.1 | 0.092 | Plasma membrane | Kinase |
| ITGB2 | Integrin, beta 2 (complement component 3 receptor 3 and 4 subunit) | NM_000211.1 | 0.120 | Plasma membrane | Other |
| MAP2K2 | Mitogen-activated protein kinase 2 | NM_030662.2 | 0.084 | Cytoplasm | Kinase |
| RAC1 | Ras-related C3 botulinum toxin substrate 1 (rho family, small GTP binding protein Rac1) | NM_006908.3 | 0.118 | Cytoplasm | Enzyme |
| RAF1 | v-raf-1 murine leukemia viral oncogene homolog 1 | NM_002880.1 | 0.082 | Cytoplasm | Kinase |
| RAPGEF1 | Rap guanine nucleotide exchange factor (GEF) 1 | NM_005312.2 | 0.085 | Cytoplasm | Other |
| RHOA | Ras homolog gene family, member A | NM_001664.1 | 0.090 | Cytoplasm | Enzyme |
| RHOG | Ras homolog gene family, member G (rho G) | NM_001665.1 | 0.115 | Cytoplasm | Enzyme |
| VASP | Vasodilator-stimulated phosphoprotein | M_003370.1 | 0.096 | Plasma membrane | Other |
| ZYX | Zyxin | M_003461.3 | 0.111 | Plasma membrane | Other |
| Name | Description | GenBank ID | Correlation coefficient | Location | Family |
|---|---|---|---|---|---|
| ABL1 | v-abl Abelson murine leukemia viral oncogene homolog 1 | NM_007313.1 | 0.081 | Nucleus | Kinase |
| ACTB | Actin, beta | NM_001101.2 | 0.098 | Cytoplasm | Other |
| ARF1 | ADP-ribosylation factor 1 | NM_001658.2 | 0.126 | Cytoplasm | Transporter |
| ARPC1A | Actin-related protein 2/3 complex, subunit 1A, 41 kDa | NM_006409.2 | 0.101 | Cytoplasm | Other |
| ARPC1B | Actin-related protein 2/3 complex, subunit 1B, 41 kDa | NM_005720.2 | 0.091 | Cytoplasm | Other |
| CAPN1 | Calpain 1, (µ/I) large subunit | NM_005186.2 | 0.097 | Cytoplasm | Peptidase |
| FYN | FYN oncogene related to SRC, FGR, YES | NM_153048.1 | 0.106 | Plasma membrane | Kinase |
| GRB2 | Growth factor receptor-bound protein 2 | NM_002086.3 | 0.090 | Plasma membrane | Other |
| ILK | Integrin-linked kinase | NM_004517.1 | 0.092 | Plasma membrane | Kinase |
| ITGB2 | Integrin, beta 2 (complement component 3 receptor 3 and 4 subunit) | NM_000211.1 | 0.120 | Plasma membrane | Other |
| MAP2K2 | Mitogen-activated protein kinase 2 | NM_030662.2 | 0.084 | Cytoplasm | Kinase |
| RAC1 | Ras-related C3 botulinum toxin substrate 1 (rho family, small GTP binding protein Rac1) | NM_006908.3 | 0.118 | Cytoplasm | Enzyme |
| RAF1 | v-raf-1 murine leukemia viral oncogene homolog 1 | NM_002880.1 | 0.082 | Cytoplasm | Kinase |
| RAPGEF1 | Rap guanine nucleotide exchange factor (GEF) 1 | NM_005312.2 | 0.085 | Cytoplasm | Other |
| RHOA | Ras homolog gene family, member A | NM_001664.1 | 0.090 | Cytoplasm | Enzyme |
| RHOG | Ras homolog gene family, member G (rho G) | NM_001665.1 | 0.115 | Cytoplasm | Enzyme |
| VASP | Vasodilator-stimulated phosphoprotein | M_003370.1 | 0.096 | Plasma membrane | Other |
| ZYX | Zyxin | M_003461.3 | 0.111 | Plasma membrane | Other |
Up-regulated genes involved in the integrin signaling canonical pathway
| Name | Description | GenBank ID | Correlation coefficient | Location | Family |
|---|---|---|---|---|---|
| ABL1 | v-abl Abelson murine leukemia viral oncogene homolog 1 | NM_007313.1 | 0.081 | Nucleus | Kinase |
| ACTB | Actin, beta | NM_001101.2 | 0.098 | Cytoplasm | Other |
| ARF1 | ADP-ribosylation factor 1 | NM_001658.2 | 0.126 | Cytoplasm | Transporter |
| ARPC1A | Actin-related protein 2/3 complex, subunit 1A, 41 kDa | NM_006409.2 | 0.101 | Cytoplasm | Other |
| ARPC1B | Actin-related protein 2/3 complex, subunit 1B, 41 kDa | NM_005720.2 | 0.091 | Cytoplasm | Other |
| CAPN1 | Calpain 1, (µ/I) large subunit | NM_005186.2 | 0.097 | Cytoplasm | Peptidase |
| FYN | FYN oncogene related to SRC, FGR, YES | NM_153048.1 | 0.106 | Plasma membrane | Kinase |
| GRB2 | Growth factor receptor-bound protein 2 | NM_002086.3 | 0.090 | Plasma membrane | Other |
| ILK | Integrin-linked kinase | NM_004517.1 | 0.092 | Plasma membrane | Kinase |
| ITGB2 | Integrin, beta 2 (complement component 3 receptor 3 and 4 subunit) | NM_000211.1 | 0.120 | Plasma membrane | Other |
| MAP2K2 | Mitogen-activated protein kinase 2 | NM_030662.2 | 0.084 | Cytoplasm | Kinase |
| RAC1 | Ras-related C3 botulinum toxin substrate 1 (rho family, small GTP binding protein Rac1) | NM_006908.3 | 0.118 | Cytoplasm | Enzyme |
| RAF1 | v-raf-1 murine leukemia viral oncogene homolog 1 | NM_002880.1 | 0.082 | Cytoplasm | Kinase |
| RAPGEF1 | Rap guanine nucleotide exchange factor (GEF) 1 | NM_005312.2 | 0.085 | Cytoplasm | Other |
| RHOA | Ras homolog gene family, member A | NM_001664.1 | 0.090 | Cytoplasm | Enzyme |
| RHOG | Ras homolog gene family, member G (rho G) | NM_001665.1 | 0.115 | Cytoplasm | Enzyme |
| VASP | Vasodilator-stimulated phosphoprotein | M_003370.1 | 0.096 | Plasma membrane | Other |
| ZYX | Zyxin | M_003461.3 | 0.111 | Plasma membrane | Other |
| Name | Description | GenBank ID | Correlation coefficient | Location | Family |
|---|---|---|---|---|---|
| ABL1 | v-abl Abelson murine leukemia viral oncogene homolog 1 | NM_007313.1 | 0.081 | Nucleus | Kinase |
| ACTB | Actin, beta | NM_001101.2 | 0.098 | Cytoplasm | Other |
| ARF1 | ADP-ribosylation factor 1 | NM_001658.2 | 0.126 | Cytoplasm | Transporter |
| ARPC1A | Actin-related protein 2/3 complex, subunit 1A, 41 kDa | NM_006409.2 | 0.101 | Cytoplasm | Other |
| ARPC1B | Actin-related protein 2/3 complex, subunit 1B, 41 kDa | NM_005720.2 | 0.091 | Cytoplasm | Other |
| CAPN1 | Calpain 1, (µ/I) large subunit | NM_005186.2 | 0.097 | Cytoplasm | Peptidase |
| FYN | FYN oncogene related to SRC, FGR, YES | NM_153048.1 | 0.106 | Plasma membrane | Kinase |
| GRB2 | Growth factor receptor-bound protein 2 | NM_002086.3 | 0.090 | Plasma membrane | Other |
| ILK | Integrin-linked kinase | NM_004517.1 | 0.092 | Plasma membrane | Kinase |
| ITGB2 | Integrin, beta 2 (complement component 3 receptor 3 and 4 subunit) | NM_000211.1 | 0.120 | Plasma membrane | Other |
| MAP2K2 | Mitogen-activated protein kinase 2 | NM_030662.2 | 0.084 | Cytoplasm | Kinase |
| RAC1 | Ras-related C3 botulinum toxin substrate 1 (rho family, small GTP binding protein Rac1) | NM_006908.3 | 0.118 | Cytoplasm | Enzyme |
| RAF1 | v-raf-1 murine leukemia viral oncogene homolog 1 | NM_002880.1 | 0.082 | Cytoplasm | Kinase |
| RAPGEF1 | Rap guanine nucleotide exchange factor (GEF) 1 | NM_005312.2 | 0.085 | Cytoplasm | Other |
| RHOA | Ras homolog gene family, member A | NM_001664.1 | 0.090 | Cytoplasm | Enzyme |
| RHOG | Ras homolog gene family, member G (rho G) | NM_001665.1 | 0.115 | Cytoplasm | Enzyme |
| VASP | Vasodilator-stimulated phosphoprotein | M_003370.1 | 0.096 | Plasma membrane | Other |
| ZYX | Zyxin | M_003461.3 | 0.111 | Plasma membrane | Other |
Transcriptional analysis of genes in the 10q QTL region
To identify empirical candidate genes in the 1-LOD support interval for the chromosome 10q QTL, we performed a detailed examination of the transcriptional profiles of the 77 available transcripts in relation to their correlation with mitochondrial content. Analysis revealed a set of 11 genes whose expression levels exhibited nominally significant correlations with mitochondrial content. This is nearly three times the number that would be expected to be found by chance alone, suggesting that this region may include multiple genes related to mitochondrial content. Table 3 provides the central results of this analysis. All of the 11 transcripts are significantly heritable ( P < 0.0001); however, only two of these transcripts exhibit nominal evidence for cis -regulation when a linkage analysis is performed on their respective expression levels at their structural location. Both ZNF11B (LOD = 0.932, P = 0.019) and NCOA4 (LOD = 0.931, P = 0.019) show evidence for such cis -regulation and therefore become the strongest empirical candidates for influencing the 10q QTL based on these transcriptional data.
Genes in the region of the chromosome 10q QTL whose expression is significantly correlated with mitochondrial content
| Gene | Mb | cM | Transcript, h2 | Association with mitochondrial content, P -value |
|---|---|---|---|---|
| ZNF11B | 42.4 | 65 | 0.537 | 0.0036 |
| ANK3 | 61.5 | 80 | 0.425 | 0.0055 |
| ARHGAP12 | 32.1 | 61 | 0.334 | 0.0116 |
| ZNF22 | 44.8 | 67 | 0.331 | 0.0162 |
| NCOA4 | 51.2 | 71 | 0.365 | 0.0174 |
| ZNF32 | 43.5 | 66 | 0.155 | 0.0205 |
| CCDC7 | 33.1 | 61 | 0.130 | 0.0242 |
| ALOX5 | 45.2 | 67 | 0.326 | 0.0291 |
| ZNF33A | 38.3 | 65 | 0.265 | 0.0420 |
| ZNF248 | 38.2 | 65 | 0.432 | 0.0472 |
| RASSF4 | 44.8 | 67 | 0.242 | 0.0497 |
| Gene | Mb | cM | Transcript, h2 | Association with mitochondrial content, P -value |
|---|---|---|---|---|
| ZNF11B | 42.4 | 65 | 0.537 | 0.0036 |
| ANK3 | 61.5 | 80 | 0.425 | 0.0055 |
| ARHGAP12 | 32.1 | 61 | 0.334 | 0.0116 |
| ZNF22 | 44.8 | 67 | 0.331 | 0.0162 |
| NCOA4 | 51.2 | 71 | 0.365 | 0.0174 |
| ZNF32 | 43.5 | 66 | 0.155 | 0.0205 |
| CCDC7 | 33.1 | 61 | 0.130 | 0.0242 |
| ALOX5 | 45.2 | 67 | 0.326 | 0.0291 |
| ZNF33A | 38.3 | 65 | 0.265 | 0.0420 |
| ZNF248 | 38.2 | 65 | 0.432 | 0.0472 |
| RASSF4 | 44.8 | 67 | 0.242 | 0.0497 |
Genes in the region of the chromosome 10q QTL whose expression is significantly correlated with mitochondrial content
| Gene | Mb | cM | Transcript, h2 | Association with mitochondrial content, P -value |
|---|---|---|---|---|
| ZNF11B | 42.4 | 65 | 0.537 | 0.0036 |
| ANK3 | 61.5 | 80 | 0.425 | 0.0055 |
| ARHGAP12 | 32.1 | 61 | 0.334 | 0.0116 |
| ZNF22 | 44.8 | 67 | 0.331 | 0.0162 |
| NCOA4 | 51.2 | 71 | 0.365 | 0.0174 |
| ZNF32 | 43.5 | 66 | 0.155 | 0.0205 |
| CCDC7 | 33.1 | 61 | 0.130 | 0.0242 |
| ALOX5 | 45.2 | 67 | 0.326 | 0.0291 |
| ZNF33A | 38.3 | 65 | 0.265 | 0.0420 |
| ZNF248 | 38.2 | 65 | 0.432 | 0.0472 |
| RASSF4 | 44.8 | 67 | 0.242 | 0.0497 |
| Gene | Mb | cM | Transcript, h2 | Association with mitochondrial content, P -value |
|---|---|---|---|---|
| ZNF11B | 42.4 | 65 | 0.537 | 0.0036 |
| ANK3 | 61.5 | 80 | 0.425 | 0.0055 |
| ARHGAP12 | 32.1 | 61 | 0.334 | 0.0116 |
| ZNF22 | 44.8 | 67 | 0.331 | 0.0162 |
| NCOA4 | 51.2 | 71 | 0.365 | 0.0174 |
| ZNF32 | 43.5 | 66 | 0.155 | 0.0205 |
| CCDC7 | 33.1 | 61 | 0.130 | 0.0242 |
| ALOX5 | 45.2 | 67 | 0.326 | 0.0291 |
| ZNF33A | 38.3 | 65 | 0.265 | 0.0420 |
| ZNF248 | 38.2 | 65 | 0.432 | 0.0472 |
| RASSF4 | 44.8 | 67 | 0.242 | 0.0497 |
DISCUSSION
Mutations in mtDNA have been associated with a wide variety of disorders from rare syndromes to more common diseases. A reduction in mitochondrial function as seen in these disease states is possibly the result of variations in the mtDNA sequence and a reduction in mtDNA levels. Here, we investigate the genetic variance influencing levels of mtDNA expression. Our results indicate a substantial genetic component to mitochondrial content (a surrogate measure of mitochondrial integrity), as observed in two independent populations, justifying the search for genes underlying the genetic variance in this trait.
Results from our study showed significant evidence for the identification of two genomic regions harboring genes influencing variations in mitochondrial content. Mitochondrial content was measured in a Mexican American cohort of 1259 individuals and in a replication cohort of 1088 individuals of European ancestry. Using genome scan data obtained previously, we have localized QTLs influencing mitochondrial content to both the nuclear and mitochondrial genomes in the Mexican American cohort. A single nuclear locus for mitochondrial content was observed on chromosome 10q11 (LOD > 3). Our replication study performed in the sample of European ancestry confirmed the presence of a QTL in the same region on chromosome 10q as that in the Mexican Americans. This region of chromosome 10 contains several strong candidate genes, identified through both bioinformatics and transcriptional profiling prioritization. On the basis of an analysis of information publicly available on the genomic databases, two likely candidate genes involved in mitochondrial processing, TFAM and TIMM23 , were identified.
Transcription factor A mitochondrial precursor ( TFAM ) plays multiple essential roles in maintaining mtDNA integrity. Given the importance of the mitochondria in cellular respiration, maintenance of the mitochondrial genome is critical for survival. TFAM is best known for its role as a key activator of mitochondrial transcription; however, it is also essential for replication, nucleoid formation, DNA damage sensing and repair ( 21 ). TFAM is a member of the high mobility group (HMG) protein family and contains two HMG box DNA-binding domains for DNA binding activity ( 21 ). For mtDNA transcription to occur, mitochondrial specific RNA polymerase, TFAM (containing the HMG box protein), and mitochondrial transcription factor B must all be present. TFAM binds immediately upstream of the mitochondrial transcriptional start sites to activate transcription ( 22 ). TFAM is required for accurate and efficient promoter recognition by the mitochondrial RNA polymerase. The dedicated mitochondrial polymerase (POLRMT) cannot interact with promoter DNA and initiate transcription without the presence of TFAM ( 22 ). TFAM is also able to efficiently bind, unwind and bend DNA without sequence specificity, a trait of the HMG domain family ( 22 , 23 ).
TIMM23 , translocase of inner mitochondrial membrane 23, is a key component of the TIM23 mitochondrial complex involved in the translocation of mitochondrial proteins from the cell nucleus into the mitochondria. Most mitochondrial proteins are encoded in the nucleus, synthesized in the cytosol on free ribosomes and then transported into or across the mitochondrial membranes. There are at least four main translocation complexes that mediate this process, ensuring that mitochondrial proteins are recognized and delivered, transported across membranes, sorted into the correct mitochondrial compartments and assisted with energy barriers ( 24 ). Protein import into the mitochondria is mediated by a single TOM (translocase of the outer membrane) complex and two functionally and structurally different TIM (translocase of the inner membrane) complexes, TIM23 and TIM22. Each of these TIM complexes is responsible for importing subsets of proteins into mitochondrial compartments. The TIM22 complex imports proteins with internal targeting signals and the TIM23 complex imports all proteins with amino-terminal presequence. The TIM23 complex consists of several membrane proteins, including Tim23, Tim17, Tim50 and Tim44, and provides a binding site for the mitochondrial heat shock protein, mtHSP70, and its cofactor Mge1.
Not only is Tim23 essential for protein import into the mitochondria, but several studies have also shown the importance of TIMM23 in cell viability ( 25 , 26 ). There is also some new evidence that the TIMM23 complex may be involved in diabetes ( 27 ) and is associated with the amelioration of oxidative stress damage induced by high glucose levels in the diabetic rat. The proposed mechanism of this relationship is the improved import of antioxidative enzymes such as superoxide dismutase and glutathione peroxides, facilitated by the TIMM23/TIMM44 complex ( 27 ).
In addition to these mitochondrial processing genes identified through bioinformatics, our transcriptional profiling data set was used to identify genes that correlate with mitochondrial content within our 10q QTL. Of the 77 transcripts available within the chromosome 10 linkage region, 11 showed nominally significant correlations ( P < 0.05). These included several members of the zinc-finger family ( ZNF11B , ZNF22 , ZNF32 , ZNF33A and ZNF248 ); anykrin 3—an interconnecting protein in the nervous system; a nuclear receptor co-activator with steroid receptor transcription activity ( NCOA4 ); a RHO GTPase ( ARHGAP12 ); a novel coiled-coil domain containing protein ( CCDC7 ); a ras oncogene family member ( RASSF4 ) and 5-lipoxgenase—an enzyme involved in catalyzing the biosynthesis of leukotrienes ( ALOX5 ).
A whole genome analysis of our transcriptional profiling data set identified a total of 829 genes showing significant correlations with mitochondrial content (a complete list of these genes can be seen in Supplementary Material, Table S1). These genes varied in their biological functions and many were novel. Some insight into the biological functions and interactions of this complex data set were obtained through the use of the IPA application, which allowed us to identify biological networks of activity and potential upstream and downstream regulatory genes. Of these 829 genes, function and pathway information was available for 51%. The PTEN signaling and integrin signaling canonical pathways were the most significant of the 12 significant pathways identified. Within our table of the 30 transcripts showing the strongest correlations with mitochondrial content (Table 1 ), four of these genes were significantly up-regulated in the integrin signaling pathway ( ARF1 , ITGB2 , RAC1 and ARHG ).
Several of these transcripts most significantly correlated with mitochondrial content also fall within two of the most highly significant functional categories, tumorigenesis and cellular assembly and organization. C19orf10 , IMPDH1 , MAPK7 , and OAZ1 act in tumorigenesis pathways; ARHG , CAPNS1 , CDC14A , FXYD5 , ARF1 , CDC42EP4 , ITGB2 and FLII fall within the cellular assembly and organization functional category; RAC1 , NCOA1 and RABEP1 have functions within both of these categories.
Significant linkage to the mitochondrial genome was also detected for mitochondrial transmission, indicating a large genetic component that is inherited via the mitochondrial genome. Given that the linkage region is obligately small since it is limited to the mitochondrial genome, the mitochondrial linkage signal can be investigated comprehensively through resequencing of the mitochondrial genome in all of the maternal lineages of the Mexican American sample. Our replication analyses in an independent sample of European ancestry also showed evidence for mitochondrial transmission, as seen in the original Mexican American sample.
Our research has identified two QTLs (one nuclear and one mitochondrial) that influence mitochondrial content. Through bioinformatic and transcriptomic analyses, we have identified several plausible candidate genes that could underlie the variation observed. Further research is required to comprehensively dissect these QTL regions to identify those genes responsible for the observed genetic variance in mitochondrial content. A thorough investigation of the transcripts correlated with mitochondrial content could lead to the identification of other genes and pathways influencing this trait.
MATERIALS AND METHODS
Subject selection
Existing DNA samples and data from the San Antonio Family Heart Study (SAFHS) were utilized in this study. The study began in 1992 and was designed primarily to investigate the genetics of cardiovascular disease and its risk factors in Mexican Americans. The SAFHS, previously described in detail in MacCluer et al . ( 28 ), included 1431 individuals in 42 extended families at baseline. Ascertainment occurred by way of a single adult Mexican American proband selected at random, without regard to presence or absence of disease and almost exclusively from Mexican American census tracts in San Antonio. To ensure large, multigenerational pedigrees, probands had to have at least six age-eligible offspring and/or siblings living in San Antonio. All first, second and third degree relatives of the proband and of the proband's spouse, aged 16 or above, were eligible to participate in the study ( 28 ). Subsequently, approximately 850 participants were recalled for a 5-year follow-up (1997–2000), and 950 have been recalled again for a 10-year follow-up (2002-current) as part of the ongoing SAFHS investigations. All protocols were approved by the Institutional Review Board of the University of Texas Health Science Center at San Antonio (San Antonio, TX).
Genome scan
A total of 1348 subjects were genotyped for 439 highly polymorphic STR markers using the MapPairs version 6 Linkage Screening Set (Research Genetics, Carlsbad, CA). Genotype error detection and correction was performed using SIMWALK2 ( 29 ) and multipoint identity-by-descent (IBD) probability matrices have been calculated using Loki ( 30 ).
Mitochondrial content quantitation
There are no existing studies on the relative importance of genetic factors in the determination of mitochondrial content in humans. In order to address this deficiency, we performed the mitochondrial content assay (described in what follows) on samples from 1259 Mexican Americans who are members of 42 extended families. This sample of extended Mexican American families has substantial power to estimate relative genetic effects and to localize QTLs. Table 4 shows the breakdown of pair-wise relationships observed among the 1259 measured individuals. The complexity and depth of the pedigrees is revealed by the number of higher degree relative pairs. The average pedigree size is approximately 30 individuals. There are 748 females and 511 males in the sample, with a mean age of approximately 39 years at the time of initial recruitment (ranging from 16 to 94 years of age).We measured mtDNA content, relative to nuclear DNA, using an adaptation of the method of Walker et al . ( 31 ). Human nuclear and mtDNA samples were assayed in the same reaction plates with a CEPH DNA standard and real-time PCR performed using the Applied Biosystems 7900HT Sequence Detection System (Foster City, CA). Figure 5 shows the standard curves that we have obtained for DNA quantitation. The assay of the Mexican American samples was performed in a single batch and had an intra-assay coefficient of variation of ∼5%, suggesting sufficient accuracy for phenotyping.
Quantitation range of nuclear and mtDNA.
Quantitation range of nuclear and mtDNA.
Relative pairs in the 42 extended Mexican American families
| Type of relationship | Number of pairs |
|---|---|
| Identical twins | 3 |
| Parent–offspring | 1146 |
| Sibling | 1348 |
| Second degree | 3228 |
| Third degree | 4051 |
| Fourth degree | 3350 |
| Fifth degree | 379 |
| Sixth degree | 964 |
| Seventh degree | 326 |
| Other | 137 |
| Total | 14932 |
| Type of relationship | Number of pairs |
|---|---|
| Identical twins | 3 |
| Parent–offspring | 1146 |
| Sibling | 1348 |
| Second degree | 3228 |
| Third degree | 4051 |
| Fourth degree | 3350 |
| Fifth degree | 379 |
| Sixth degree | 964 |
| Seventh degree | 326 |
| Other | 137 |
| Total | 14932 |
Relative pairs in the 42 extended Mexican American families
| Type of relationship | Number of pairs |
|---|---|
| Identical twins | 3 |
| Parent–offspring | 1146 |
| Sibling | 1348 |
| Second degree | 3228 |
| Third degree | 4051 |
| Fourth degree | 3350 |
| Fifth degree | 379 |
| Sixth degree | 964 |
| Seventh degree | 326 |
| Other | 137 |
| Total | 14932 |
| Type of relationship | Number of pairs |
|---|---|
| Identical twins | 3 |
| Parent–offspring | 1146 |
| Sibling | 1348 |
| Second degree | 3228 |
| Third degree | 4051 |
| Fourth degree | 3350 |
| Fifth degree | 379 |
| Sixth degree | 964 |
| Seventh degree | 326 |
| Other | 137 |
| Total | 14932 |
Replication analyses
For replication analyses, we used existing DNA samples and data from the Metabolic Risk Complications of Obesity Genes (MRCOB) study ( 32 ). This project was initiated in 1994 and involved recruitment of families of European ancestry from the TOPS (Take Off Pounds Sensibly, Inc.) membership. These families largely reside in the US Midwest. Subjects were recruited to be healthy, except for obesity and its metabolic health consequences. For the current study, we employed a subset of the largest families consisting of 1088 individuals distributed over 170 families. Most of these families were nuclear; however, there were several large extended kindreds included in the sample. The sample is predominantly female with 291 males and 797 females. The mean age is 47.5 years, ranging from 13 to 94 years. For linkage analysis, we used existing genetic marker data and as with the Mexican American cohort, we measured mtDNA content using the method described earlier. Protocols were approved by the FMLH Institutional Review Board of the Medical College of Wisconsin (Milwaukee, WI).
Transcriptomic scan
Genome-wide quantitative transcriptional profiles have been obtained from 1240 Mexican American individuals in the SAFHS. For this study, lymphocyte samples were available from 1280 Mexican American individuals. Of the 1280 samples analyzed, acceptable transcriptional profiles were obtained from 1240 individuals. The sample consisted of 1154 individuals from 46 pedigrees and an additional 86 singletons. There are 734 females and 506 males in the sample, with a mean age of 39.3 years (SD = 16.7 years). Ages range between 15.5 and 94.2 years. Using lymphocyte samples collected from the first SAFHS examination period (January 1992–July 1996), we have obtained high-quality RNA and have been successful in quantitating large numbers of transcripts.
For each sample, 47 289 transcripts were quantitated using the Sentrix Human-6 version 1 expression BeadChip supplied by Illumina (San Diego, CA). Briefly, total RNA was extracted from 1280 lymphyocyte samples stored in RPMI-C freeze medium using a modified procedure of the QIAGEN RNeasy 96 protocol for isolation of total RNA from animal cells using spin technology (QIAGEN Inc.; Valencia CA). Total RNA yield (µg) and purity (260:280 nm) were determined spectrophotometrically using the NanoDrop ND-1000 (Wilmington, DE). Integrity of resuspended total RNA was determined by electrophoretic separation and subsequent laser-induced fluorescence detection using the RNA 6000 Nano Assay Chip Kit on the Bioanalyzer 2100 using the 2100 Expert software (Agilent Technologies, Germany). A total of 500 ng total RNA was synthesized, amplified and purified using the Ambion MessageAmp II Amplification Kit (Ambion; Austin TX) following the Illumina Sentrix Array Matrix 96-well expression protocol. Anti-sense RNA (aRNA) total yield (µg) and purity (260:280 nm) were again determined spectrophotometrically using the NanoDrop ND-1000 and a total of 1.5 µg of aRNA was stored for sample hybridization.
Hybridization of aRNA to Illumina Sentrix Human Whole Genome (WG-6) version 1 BeadChips and subsequent washing, blocking and detecting were performed using Illumina's BeadChip 6 × 2 protocol. Samples were scanned on the Illumina BeadArray 500GX Reader using Illumina BeadScan image data acquisition software (ver. 2.3.0.13). Illumina BeadStudio software (ver. 1.5.0.34) was used for preliminary data analysis, with a standard background normalization, to generate an output file for statistical analysis. To assess quality metrics of each day's run, several quality control procedures were implemented.
Statistical analyses
Estimation of the heritability of quantitative mtDNA levels
We performed quantitative genetic analysis of mtDNA levels to estimate the extent to which genetic factors influence this novel trait. There are no existing studies of the heritability of quantitative variation in mtDNA levels. We used the program, SOLAR ( 33 ), to obtain the maximum likelihood estimate of the heritability of quantitative mtDNA levels in the Mexican American families. This method uses all available familial information to obtain a statistically optimal estimate of heritability in the sample. Covariate effects including sex and age (and their interactions) were simultaneously estimated. To test the significance of the heritability of mtDNA levels, we employed a likelihood ratio test.
Linkage analysis of quantitative mtDNA levels
We used a pedigree-based multipoint variance components approach to test for linkage between marker loci and continuous traits such as mitochondrial content or quantitative transcript levels using a maximum likelihood method ( 33 , 34 ). In this method, the expected genetic covariances between relatives are specified as a function of the IBD relationships at a marked genomic location and the QTL-specific heritability which measures the genetic-effect size. The null hypothesis that the relative additive genetic variance (as estimated by the QTL-specific heritability) due to a QTL for a given phenotype equals zero (no linkage) can be tested using a likelihood ratio test. The likelihood of a restricted model in which the QTL-specific heritability is constrained to zero is compared with the likelihood of the general model in which genetic variance due to the QTL is estimated. The natural logarithm (ln) likelihood values of the general model and restricted model are then compared using the likelihood ratio test. Twice the difference between the ln likelihood values of these models yields a test statistic that is asymptotically distributed as a 1/2:1/2 mixture of a χ 2 distribution with one degree of freedom and a point mass at zero. The likelihood value can be converted into a logarithm (base 10) of odds to obtain an LOD score that is equivalent to the classical LOD score of linkage analysis.
Association of mitochondrial content with transcriptional expression
We tested for association between mitochondrial content levels and gene expression levels using a regression model that allows for residual genetic effects as implemented in the computer program, SOLAR. In this approach, mitochondrial content was modeled as a function of a given transcript's expression level. A likelihood ratio statistic was used to formally test the hypothesis that mitochondrial content levels were not correlated with gene expression levels. This test was performed conditionally upon other covariate effects including those of sex, age and their interactions as mentioned previously.
Ingenuity pathways analysis
After assessing the association between mitochondrial content and transcript levels, we filtered significantly associated transcripts using an FDR of 1 in 4 (so that 75% of chosen transcripts should truly be associated with mitochondrial content). These selected transcripts (and the levels of their association with mitochondrial content) were analyzed using IPA version 4.1 (Ingenuity ® Systems, www.ingenuity.com ). A functional analysis was conducted to identify the biological functions and/or diseases most significantly relevant to the data set. Genes from the data set that met the P -value cut-off of 0.01 (with an FDR of 0.25 or 1 in 4) and with documented data on biological functions and/or diseases in the Ingenuity Pathways Knowledge Base were considered for the analysis. The right-tailed Fisher's exact test was used to calculate a P -value determining the probability that each biological function and/or disease assigned to that data set was due to chance alone. Our genes of interest were overlaid onto a global molecular network developed from literature reported connectivity that is recorded in the Ingenuity Pathways Knowledge Base.
SUPPLEMENTARY MATERIAL
Supplementary Material is available at HMG Online.
ACKNOWLEDGEMENTS
This work was supported in part by a pilot grant from the San Antonio Area Foundation. Data collection for the San Antonio Family Heart Study was supported by a program project grant from the National Heart, Lung and Blood Institute (HL045222). We are grateful to the participants of the San Antonio Family Heart Study for their continued involvement. We also acknowledge the Fredric C. Barterr General Clinical Research Center, supported by M01-RR01346, which provides ongoing clinical support to this project. We sincerely thank the Azar/Shepperd families of San Antonio for their financial support of the transcriptional profiling study. Additional funds for transcriptional profiling, sequencing, genotyping and statistical analysis were provided by ChemGenex Pharmaceuticals Ltd, Australia. This investigation was conducted in facilities constructed with support from the Research Facilities Improvement Program Grant no. C06 RR013556 from the National Center for Research Resources, National Institutes of Health. The AT&T Genomics Computing Center supercomputing facilities used for this work were supported in part by a gift from the SBC Foundation. The statistical genetics computer package, SOLAR, is supported by a grant from the National Institutes of Mental Health (MH059490).
Conflict of Interest statement . None declared.
