Decreased Mitochondrial Dynamics Is Associated with Insulin Resistance, Metabolic Rate, and Fitness in African Americans

Abstract

Context

African American women (AAW) have a higher incidence of insulin resistance and are at a greater risk for the development of obesity and type 2 diabetes than Caucasian women (CW). Although several factors have been proposed to mediate these racial disparities, the mechanisms remain poorly defined. We previously demonstrated that sedentary lean AAW have lower peripheral insulin sensitivity, reduced maximal aerobic fitness (VO₂max), and lower resting metabolic rate (RMR) than CW. We have also demonstrated that skeletal muscle mitochondrial respiration is lower in AAW and appears to play a role in these racial differences.

Objective

The goal of this study was to assess mitochondrial pathways and dynamics to examine the potential mechanisms of lower insulin sensitivity, RMR, VO₂max, and mitochondrial capacity in AAW.

Design

To achieve this goal, we assessed several mitochondrial pathways in skeletal muscle using gene array technology and semiquantitative protein analysis.

Results

We report alterations in mitochondrial pathways associated with inner membrane small molecule transport genes, fusion–fission, and autophagy in lean AAW. These differences were associated with lower insulin sensitivity, RMR, and VO₂max.

Conclusions

Together these data suggest that the metabolic racial disparity of insulin resistance, RMR, VO₂max, and mitochondrial capacity may be mediated by perturbations in mitochondrial pathways associated with membrane transport, fission–fusion, and autophagy. The mechanisms contributing to these differences remain unknown.

African American women (AAW) have a 2-fold greater incidence of developing type 2 diabetes than Caucasian women (CW) (1–4). This racial disparity is also observed in African American and Caucasian men, but to a lesser extent (5). Although the mechanisms remain unclear, there is growing evidence that insulin sensitivity is lower in African Americans controlling for factors known to contribute to impaired insulin action like obesity. Paradoxically, visceral and hepatic fat, generally considered strong correlates of lower insulin sensitivity (6) are lower in AAW (7–9). In support of these observations, we reported that lean (body mass index [BMI] 22.7 ± 3.1 kg/m²) AAW have lower peripheral insulin sensitivity, lower maximal aerobic fitness (VO₂max), and impaired skeletal muscle mitochondrial capacity than BMI-matched CW (10). Lower resting metabolic rate (RMR), which may contribute to the development of obesity, has been consistently shown in AAW (11, 12) than CW, and may be associated with mitochondrial impairment (13). Yet, our understanding of the mechanisms potentially contributing to the racial disparity in mitochondrial function is unclear.

The association between mitochondrial function and insulin resistance is controversial (14–16). Several key observations suggest that mitochondrial pathways including metabolic inflexibility (17, 18), oxidative stress (19), and chronic inflammation (20) mediate skeletal muscle insulin resistance. On the other hand, acute high-fat feeding in rodents (21) and humans (22) results in mitochondrial biogenesis and increased oxidative capacity. Nevertheless, pathways associated with mitochondrial remodeling (eg, fission and fusion) are correlated with levels of insulin sensitivity in both humans (23, 24) and rodents (25, 26). Moreover, proteins in the fission and fusion pathways respond to increased levels of physical activity (27, 28) and changes in diet (29). Together these data suggest that multiple mitochondrial pathways likely contribute to insulin sensitivity and may contribute to the observed racial disparity of insulin resistance.

The primary goal of this study was to determine if skeletal muscle mitochondrial dynamics are impaired in lean AAW compared with lean CW. To achieve this goal, we examined several mitochondrial pathways including membrane polarization and potential, translocation, and transport in human skeletal muscle using gene array technology. In addition, expression of mitochondrial proteins of the fission and fusion, localization, and apoptosis pathways were examined. These pathways were then correlated with measures of peripheral insulin sensitivity, aerobic fitness, resting metabolic rate, and mitochondrial capacity.

Materials and Methods

Subjects

Twenty-two AAW and 22 CW women aged 18.7 to 38.3 years were recruited for the study from print advertisements in the Pittsburgh area. All included subjects were sedentary by self-report (≤20 minutes of intentional exercise, 3 times per week), not pregnant or lactating, and weight stable (<±3 kg in the previous 6 months). Subjects were excluded for unstable medical conditions, smoking, metabolic disease (eg, diabetes mellitus), or medication use that would affect the primary outcome measures. Prior to the initiation of the testing procedures, all subjects provided informed consent, and were medically screened and cleared. The protocol was approved by the University of Pittsburgh Institutional Review Board.

Study protocol

The details of the study protocol and the subject characteristics have been published elsewhere (10). Briefly, after enrollment in the study, subjects reported to the Endocrinology and Metabolism Research Center for a body composition assessment using dual-energy X-ray absorptiometry (Lunar iDXA; GE Healthcare) followed by a graded exercise test conducted on an electronically braked cycle ergometer (Lode, Groningen, The Netherlands). Maximum effort was confirmed using criteria from the American College of Sports Medicine (30). Multisensor activity monitors (SenseWear MF Armband, BodyMedia) were used to determine free-living physical activity.

Skeletal muscle biopsy and insulin sensitivity

Approximately 4 days following the exercise test, subjects reported to the Clinical and Translational Research Center at ~5.00 pm, were fed a standard meal (10 kcal/kg; 50% carbohydrate, 15% protein, 35% fat), and then fasted until the completion of study procedures. Prior to the clamp study and in the fasted state, a muscle biopsy of the vastus lateralis was performed under local anesthesia (10, 31). The biopsy sample was cleaned of adipose tissues, dried, and flash frozen in liquid nitrogen. Hepatic and peripheral insulin sensitivity were determined using a 2-step (2-hour infusion of insulin (Humulin-R) at 15 mU/m²/min, followed by 2 hours at 40 mU/m²/min) hyperinsulinemic euglycemic clamp method in the fasted state (10). Briefly, euglycemia (target 90 mg/dL) was maintained with a variable 20% dextrose infusion enriched with [6,6-²H₂]glucose. Rates of glucose disposal and endogenous glucose production were calculated using nonsteady-state equations from plasma [6,6-²H₂]glucose enrichment determined by gas chromatography mass spectrometry (10). Indirect calorimetry was used prior to the clamp procedures (baseline) and at the end of each step of the clamp (steady state).

Tissue analyses

In addition to measures of oxidative capacity and mitochondrial respiration (10), portions of the biopsy samples were used for gene expression and protein content.

Gene expression:

Approximately 15 to 20 mg of muscle tissue was used for the determination of gene expression using quantitative real time polymerase chain reaction. RNA was extracted using the RNeasy mini kit (Qiagen 74104, Qiagen, Valencia, CA). cDNA was prepared from 1 µg of RNA using the RT² First Strand kit (Qiagen 330421; Qiagen, Valencia, CA). The real-time RT²-PCR gene array for human mitochondria biogenesis and function (Qiagen 330321, Qiagen, Valencia, CA) was used to quantify gene expression. Data were normalized to the arithmetic mean of the housekeeping genes B2M, GAPDH, HPRT1, and RPLP0.

Protein expression.

Total protein was extracted from ~30 mg of skeletal muscle tissue using cell lysis buffer (Cell Signaling, Boston, MA) with protease inhibitors (#11836153001, Sigma-Aldrich), as previously described (32). Protein expression was semiquantified by western blot using the following antibodies: FIS1 (#sc-98900, Santa Cruz), MFN2 (#M6444, Sigma-Aldrich), DLP1 (#611112, BD Biosciences), LC3AB (#84557, Cell Signaling), Beclin1 (#3495, Cell Signaling), OPA1 (#612606, BD Biosciences), and BNIP3 (#B7431, Sigma-Aldrich). Proteins were normalized to β-actin (#sc-47778, Santa Cruz), and an internal loading control was used to adjust for gel-to-gel variability. Blots were visualized using Immun-Star WesternC chemiluminescence (BioRad) and imaged with ChemiDoc XRS + (BioRad). Densitometry was completed using ImageJ software.

Statistical analyses

Data are presented as mean ± standard error of the mean. Gene expression data were not considered independent, and were therefore analyzed with partial least-squares correlation (PLSC) (33). The correlation between data projected on the principal singular vectors provides a multivariate generalization of the Pearson product–moment correlation, which can be assessed with a permutation test. For our data, PLSC was performed between a vector for race, mitochondrial function, or physiological variable, x, and the matrix of gene/protein expression data, Y. Initially, PLSC was performed for all gene expression data. Those variables with a partial correlation of 0.2 (absolute value) were retained and PLSC was recalculated and tested with 1000 random permutations of the data. The criterion of a partial correlation of 0.2 was chosen because (1) it tended to yield results that were consistent with finding significant Pearson product–moment correlations, variable by variable, and (2) it was rather conservative (did not inflate the importance of gene expression variables). Missing data were estimated via prediction from linear regression prior to performing the singular value decomposition of the matrix cross-product. The resulting correlation is the same as a standardized (multivariate) mean difference between the groups, the same way a Pearson product–moment correlation is analogous to a 2-sample t-test or analysis of variance, when 1 variable is binary. Results concerning race are presented as means ± standard error of the mean, with latent variables from PLSC noted. For the continuous mitochondrial function or physiological variables, the partial correlations for latent variables along with PLSCs are presented. All analyses were performed with the statistical software, R (34). Correlational analyses of protein expression and mitochondrial function as well as other physiological variables were performed using Pearson product–moment analyses. Statistical significance was assumed at P ≤ .05.

Results

Insulin sensitivity, resting metabolic rate, and maximal aerobic fitness are lower in AAW

Subjects for this study represent a subset of the total participant pool as previously reported (10) owing to limits of tissue availability for gene/protein expression. As demonstrated in the parent study (10), AAW and CW were similar in age, BMI, and weight by design. Peripheral insulin sensitivity, normalized to circulating insulin, was lower in AAW compared to CW, in agreement with our previous report (10). Levels of moderate and vigorous physical activity were similar between the groups. Maximal aerobic fitness and resting metabolic rate, normalized to FFM, were lower in AAW (Table 1) (10).

Table 1.

Subject characteristics.

	AAW (n = 10)	CW (n = 11)	P
Age	21.65 ± 0.90	23.45 ± 1.76	.43
Body composition
BMI, kg/m2	24.30 ± 1.02	23.45 ± 1.01	.57
Wt, kg	66.67 ± 3.49	65.36 ± 3.00	.78
FFM, kg	47.40 ± 2.33	43.01 ± 1.42	.12
FM, kg	19.27 ± 1.84	22.34 ± 2.00	.27
Peripheral insulin sensitivity (M/I), mg/min/kgFFM/µU insulin • mL	0.26 ± 0.05	0.36 ± 0.06	.01
RMR, kcal/d adjusted for FFM	1247 ± 37	1374 ± 35	.02
Physical activity and fitness
Moderate activity, min/day	96.56 ± 11.08	70.67 ± 12.56	.14
Vigorous activity, min/day	18.13 ± 4.69	35.04 ± 16.62	.36
VO2max, mL/kgFFM/min	46.48 ± 1.49	53.64 ± 2.07	.01

All data are mean ± standard error of the mean.

Abbreviations: BMI, body mass index; Wt, body weight; FFM, fat free mass; FM, fat mass; RMR, resting metabolic rate

Selective genes in pathways associated with inner mitochondrial membrane transport, mitochondrial transport, and outer membrane transport are lower in skeletal muscle from AAW

Our previous data suggest that impaired mitochondrial function is a key variable in predicting the lower insulin sensitivity (10), as well as decreased resting metabolic rate and maximal aerobic fitness (13). Of the 79 genes assessed from a commercially available gene array platform, only 6 were associated with differentiation by race (Fig. 1A–J). These genes included TAZ (partial r = –0.425), TIMM17B (partial r = –0.468), TIMM44 (partial r = –0.315), TIMM8A (partial r = –0.433), TOMM34 (partial r = –0.409), and TSPO (partial r = –0.382). The PLSC r comprising these variables was significant (PLSC r = 0.631; P = .012). Four of the genes downregulated in AAW are associated with inner membrane translocation (TAZ, TIMM44, TIMM17B, and TIMM8A; Fig. 1A). The other 2 genes, TSPO and TOMM34, are associated with mitochondrial transport (Fig. 1B) and outer membrane translocation (Fig. 1E), respectively. There were no racial differences in gene content within the pathways associated with small molecule transport, membrane localization, mitochondrial fission and fusion, targeting proteins to the mitochondria, apoptosis, membrane polarization and potential, or mitochondrial protein import. Thus, selective genes associated with inner mitochondrial membrane transport, mitochondrial transport, and outer membrane transport are decreased in skeletal muscle from AAW.

Figure 1.

Skeletal muscle gene expression analysis. Gene expression was determined from skeletal muscle biopsy samples as described in Materials and Methods from lean (BMI < 25 kg/m²) Caucasian women (CW, n = 10, white bars) and BMI-matched African American women (AAW, n = 11, black bars). Gene associated with (A) Inner membrane translocation, (B) Mitochondrial transport, (C) small molecule transport, (D) membrane localization, (E) outer membrane translocation, (F) mitochondrial fission and fusion, (G) targeting proteins to mitochondria, (H) apoptosis, (I) membrane polarization and potential, and (J) mitochondrial protein import. Data are mean ± standard error of the mean and normalized to CW. BMI, body mass index. *Latent variables retained by PLSC (partial r between—0.32 and -0.47); PLSC r = 0.631; P = .012.

Proteins associated with mitochondrial fission–fusion and autophagy are altered in skeletal muscle from AAW

The mitochondrial network is dynamic, reacts to various stimuli, and is regulated by the fission–fusion and autophagy pathways. Of the fission–fusion proteins, FIS1 (Fig. 2A) and MFN2 (Fig. 2B) were lower, while OPA1 (Fig. 2C) was significantly higher (all P ≤ .05) in skeletal muscle from AAW than from CW. DLP1 (Fig. 2D) was similar between the groups. Regarding the autophagy pathway, LC3A/B (Fig. 2E) was lower, while Beclin1 (Fig. 2F) was higher in AAW than in CW (both P ≤ .05). Thus, proteins in the mitochondrial regulatory pathways of fission–fusion and autophagy are altered in skeletal muscle from AAW.

Figure 2.

Skeletal muscle protein expression. Proteins were extracted from skeletal muscle biopsy samples and quantified as described in Materials and Methods from lean (BMI < 25 kg/m²) Caucasian women (CW, n = 10, white bars) and BMI-matched African American women (AAW, n = 11, black bars). (A) Fission 1 (FIS1), (B) Mitofusin 2 (MFN2), (C) Optic atrophy 1 (OPA1), (D) Dynamin 1 like (DLP1), (E) Light chain 3 A/B (LC3A/B), and (F) Beclin-1. Data are normalized to protein content of β-actin and relative to CW, arbitrarily set to 1. Representative immunoblots (right) for CW and AAW. *P ≤ 0.05, group difference.

Insulin sensitivity is associated with small molecule transport, fission–fusion, and autophagy pathways

Peripheral insulin sensitivity is lower in AAW than in CW and the racial difference is associated with decreased mitochondrial capacity (10). When all subjects were pooled for correlational analyses, genes associated with mitochondrial fusion (MFN1 and MFN2) tended to be positively associated with peripheral insulin sensitivity. In support of the gene expression data, MFN2 protein expression was positively correlated with peripheral insulin sensitivity. With respect to proteins involved in autophagy, Beclin1 was negatively associated with insulin sensitivity, while LC3A/B was positively associated with insulin sensitivity. Several genes associated with the small molecule transport pathway were positively associated with insulin sensitivity (SLC24A4, SLC24A15, SLC24A16, SLC24A27, and SLC24A30). Thus, insulin sensitivity is associated with small molecule transport, fission–fusion, and autophagy pathways (Table 2).

Table 2.

Correlations and partial correlations between insulin sensitivity and mitochondrial gene expression and protein content.

Parameter	r	P (2-tailed)
Gene expression	0.64	.047
FXC1	0.31
HSPD1	0.31
MFN1	0.29
MPV17	0.30
SLC25A2	0.42
SLC25A13	0.31
SLC25A19	-0.37
SLC25A2	0.42
SLC25A23	-0.32
UCP2	0.36
Protein content
MFN2	0.51	.05
Beclin1	-0.59	.02
LC3A/B	0.46	.09

Correlations for gene expression are partial least-squares correlations, with partial correlations for latent variables indicated. Correlations for protein content are independent Pearson product–moment correlations.

Resting metabolic rate is associated with mitochondrial small molecule transport, inner and outer membrane transport, and mitochondrial transport

Resting metabolic rate (RMR) is lower in AAW compared to CW across the lifespan (10, 35) and the difference may be mediated through impaired mitochondrial capacity (13). Genes expressed from 2 pathways positively associated with RMR were small molecule transport (SLC25A4, SLC25A27, and SLC25A30) and inner membrane transport (TIMM17A, TIMM17B, TIMM23, and TIMM50). Thus, genes within the small molecule transport, inner and outer membrane transport, and mitochondrial transport pathways are associated with RMR (Table 3).

Table 3.

Correlations and partial correlations between resting metabolic rate and mitochondrial gene expression and protein content.

Parameter	r	P (2-tailed)
Gene expression	0.81	.001
MIPEP	0.41
MTX2	0.33
SLC25A4	0.36
SLC25A27	0.52
TIMM50	0.42
TOMM34	0.34
Protein content
DLP1	–0.49	.08

Correlations for gene expression are partial least-squares correlations, with partial correlations for latent variables indicated. Correlations for Protein content are independent Pearson product-moment correlations.

Maximal aerobic fitness is associated with multiple mitochondrial pathways

VO₂max is lower in AAW than in CW (10) and is correlated with mitochondrial content (36) and function (37). In contrast to the associations with RMR, the only gene in the small molecule transport pathway associated with VO₂max was SLC25A4. Several genes associated with inner membrane translocation (TAZ, TIMM8A, TIMM17B, and TIMM50) were positively correlated with VO₂max. Other genes involved in apoptosis (BCL2 and SOD2) and fission and fusion (MFN1) were correlated with VO₂max. Thus, VO₂max is associated with multiple mitochondrial pathways (Table 4).

Table 4.

Correlations and partial correlations between maximal aerobic fitness and mitochondrial gene expression.

Parameter	r	P (2-tailed)
Gene expression	0.75	.003
MFN1	0.34
SOD2	0.37
TAZ	0.34
TIMM8A	0.39
TIMM17B	0.40
TIMM50	0.44
UXT	0.35

Correlations for gene expression are partial least-squares correlations, with partial correlations for latent variables indicated.

Mitochondrial respiration is associated with mitochondrial small molecule and inner membrane transport

Our previous data suggest that mitochondrial respiration is lower in AAW under basal, maximal coupled, and maximal uncoupled respiration independent of mitochondrial content and muscle fiber type (10). Under conditions of carbohydrate-supported respiration, basal (5 genes) and maximal coupled respiration (6 genes) were associated with genes within the inner mitochondrial membrane transport, mitochondrial transport, small molecule transport, and apoptotic pathways. In contrast, under conditions of carbohydrate with palmitate, basal (8 genes), maximal coupled (8 genes), and maximal uncoupled respiration (8 genes) were associated with genes within the inner membrane transport, mitochondrial transport, small molecule transport, and apoptotic pathways. These data further support the observations of associations linking mitochondrial gene expression with insulin sensitivity, RMR, and VO₂max (Table 5).

Table 5.

Correlations and partial correlations between mitochondrial respiration and mitochondrial gene expression.

Parameter	r	P (2-tailed)	Parameter	r	P (2-tailed)
Carbohydrate supported			Carbohydrate/palmitate supported
Basal	0.683	.009	Basal	0.682	.016
SLC25A4	–0.46		SLC25A1	–0.34
SOD2	0.42		SLC25A2	–0.29
TIMM17B	0.44		SLC25A4	0.29
TIMM50	0.53		SOD2	0.28
TIMM8A	0.38		TAZ	0.42
			TIMM17B	0.36
			TIMM23	0.30
			TIMM50	0.49
Maximal coupled	0.651	.050	Maximal coupled	0.620	.041
SLC25A14	–0.35		SLC25A1s	–0.29
SLC25A2	–0.43		SLC25A4	0.34
STARD3	–0.38		SOD2	0.44
TIMM17B	0.39		TAZ	0.32
TIMM8A	0.41		TIMM17B	0.37
TIMM8B	0.35		TIMM44	0.28
			TIMM50	0.44
			TIMM8A	0.30
Maximal uncoupled	0.630	.086	Maximal uncoupled	0.621	.039
			SLC25A10	–0.32
			SLC25A4	0.34
			SOD2	0.38
			TAZ	0.37
			TIMM17B	0.40
			TIMM44	0.27
			TIMM50	0.42
			TIMM8A	0.32

Correlations for gene expression are partial least-squares correlations, with partial correlations for latent variables indicated.

Discussion

The overall goal of this study was to determine if pathways associated with skeletal muscle mitochondrial dynamics are impaired in lean AAW compared with CW. A second goal was to examine the skeletal muscle gene expression and protein content across all subjects to better understand the relative contribution of these factors to the physiological variables known to be racially different (10, 13). This study revealed several novel findings. First, we present evidence that selective genes within the mitochondrial pathways of (1) inner membrane translocation, (2) mitochondrial transport, (3) outer membrane translocation, and (4) targeting proteins to mitochondria pathways are lower in skeletal muscle from lean AAW. Second, we demonstrate that several key proteins associated with fission–fusion and autophagy are altered in skeletal muscle from AAW. Finally, key phenotypic variables, demonstrated to be different between AAW and CW, were highly correlated with genes associated with several mitochondrial pathways, as well as proteins related to fission–fusion and autophagy. Together these data suggest that although there were no universal differences in the main pathways between lean AAW and CW, alterations in protein content for fission–fusion and autophagy between the racial groups may contribute to, or be reflective of, mitochondrial dysfunction in AAW predisposing them to increased prevalence of obesity and development of type 2 diabetes.

Mitochondrial function is regulated by a variety of factors encoded by mitochondrial and nuclear DNA (38) representing over 1000 different proteins (39). Recent evidence suggests that pathways associated with mitochondrial energy metabolism are altered in adipose tissue from African Americans (40). Further, mtDNA is reduced in prostate cancer cells in African American men (41) and may be linked to other cancer health disparities (42).

In our cohort of lean women of similar age and activity levels, we demonstrate very few significant racial differences in expression of the 79 mitochondrial genes examined in skeletal muscle. Nevertheless, there is evidence to suggest that perhaps alterations of these genes may contribute to decreased insulin sensitivity, resting metabolic rate, and maximal aerobic fitness all of which are lower in AAW (10). In support of this hypothesis, several genes, which may contribute to our observed racial phenotype, are correlated with mitochondrial function. Although this correlation does not imply causation, a better understanding of the contribution of altered mitochondrial gene expression in skeletal muscle to racial differences in phenotype is needed.

The inner mitochondrial membrane contains the key complexes necessary for the generation of ATP. Yet, central to mitochondrial function is the membrane transport capacity and ultrastructure of the mitochondria. The contribution of changes in inner membrane translocation to mitochondrial function is supported by lower gene expression (TAZ, TIMM8A, and TIMM17B) in AAW than in CW. The transacylase Tafazzin (TAZ) modulates cardiolipin, the phospholipid necessary for mitochondrial structure, fission–fusion, and mitophagy (43). In humans (44) and rodents (45), inactivation of TAZ results in impaired mitochondrial structure and function (46). Phosphorylation of TAZ by GSK-3β, a kinase associated with type 2 diabetes and obesity, results in proteasomal degradation (47) and decreased phosphorylation of IRS1 in vitro (48). Conversely, activation of TAZ protects against lipid induced insulin resistance in skeletal muscle (49). The combination of lower TAZ gene expression in AAW combined with the correlation of TAZ and mitochondrial respiration across all subjects is strongly suggestive of a role for TAZ in mediating the racial differences in mitochondrial function (10).

The translocase proteins of the inner mitochondrial membrane are chaperones for proteins into the mitochondria (50). We demonstrate that the gene expression from 2 members of this family (TIMM8A and TIMM17B) are lower in skeletal muscle from AAW than in CW. TIMM17B forms a membrane pore with TIMM23 to facilitate the import of the proteome, thus contributing to mitochondrial biogenesis (51). A mutation in the TIMM8A protein has been associated with Mohr–Tranebjaerg syndrome characterized by progressive dystonia, decreased visual acuity, and dementia (52). Further study of TIMM8A has revealed that TIMM8A plays a role in the activation of dynamin 1-like protein (Dlp1 or Drp1)-mediated fission during apoptosis (53). Thus, decreased levels of TIMM17B may be reflective of attenuated capacity for mitochondrial fission/fusion and ultimately altered bioenergetics. In support of this hypothesis, a recent study demonstrated an upregulation of several translocase proteins in response to chronic exercise training in skeletal muscle from older male and female subjects (54). These data, in addition to our demonstration that several genes are correlated with aerobic fitness and mitochondrial respiration across all subjects, supports the notion that decreased expression of these genes in skeletal muscle may contribute to our demonstrated racial phenotype differences (10).

The mitochondrial translocator protein (TSPO) is involved in cholesterol transport and steroidogenesis (55). Although TSPO is best characterized in neuronal tissue, it is expressed in metabolically active tissues like skeletal muscle (56). TSPO has been demonstrated to play a role in mediating LPS-induced inflammation in microglia (57). A lack of TSPO in fibroblasts (58) and microglial cells (59) developed from TSPO knockout mice demonstrated lower oxygen consumption, mitochondrial membrane potential, and ATP synthesis. Further, TSPO gene expression is lower in peripheral blood mononuclear cells from patients with multiple sclerosis (60), a condition associated with mitochondrial dysfunction (61). Together, these data support the notion that altered TSPO gene expression in skeletal muscle from AAW may contribute to the mitochondrial dysfunction observed in these subjects.

The translocase of the outer mitochondrial membrane 34 (TOMM34) has been characterized as a cochaperone for heat shock proteins (HSPs) (HSP70 and HSP90) (62) and associated with cancer cell growth (63, 64). HSPs have been associated with insulin sensitivity in rodents and humans through the activation of inflammatory signaling pathways (eg, c-Jun) (65). Moreover, African American prostate cancer cells have been demonstrated to express lower levels of HSPs than Caucasian cells and exhibit greater mitochondrial dysfunction (66). Thus, it is tempting to speculate that lower TOMM34 as observed in our cohort of AAW may be associated with lower skeletal muscle HSP and an enhanced inflammatory milieu contributing to reduced insulin sensitivity and mitochondrial dysfunction. TOMM34 is transcriptionally regulated by nuclear respiratory factor 1 (NRF1) (67), a key factor mediating mitochondrial biogenesis (68). Thus, our demonstration of lower mitochondrial content in AAW (10) supports the observation of lower TOMM34 in skeletal muscle from AAW.

The opposing processes of mitochondrial fission and fusion are finely regulated by a number of proteins, including MFN1, MFN2, OPA1, DRP1, and FIS1. Alterations in the proper balance in these 2 processes are associated with mitochondrial dysfunction, oxidative stress and insulin resistance in skeletal muscle (25, 26, 58, 69). We previously demonstrated that when compared with sedentary subjects of varying ages, several proteins in the fission–fusion (FIS1, DLP1, MFN2, and OPA1) and autophagy (Beclin-1) pathways were higher in skeletal muscle from young physically active subjects who had concomitantly higher levels of mitochondrial respiration and maximal aerobic fitness (70). In support of those observations, we demonstrate that lower skeletal muscle protein content of FIS1 and MFN2 in AAW is concomitant with lower insulin sensitivity. In contrast, we demonstrate higher Beclin-1 protein content in skeletal muscle from AAW than from CW. This difference in skeletal muscle Beclin-1 and insulin sensitivity between the studies may be due to Beclin’s complex nature in terms of protein binding activity (71) and other as of yet defined functions. In support of our current observations, skeletal muscle Beclin-1 mRNA expression in ob/ob mice (72) and protein content following immobilization (73) are elevated concomitant with impaired insulin sensitivity. Thus, additional studies are necessary to better understand the association between Beclin-1 regulation, autophagy, and insulin resistance.

In rodents and humans, skeletal muscle MFN2 is associated with mitochondrial metabolism, oxidative stress, insulin resistance, obesity, and early onset type 2 diabetes (74–76). Further, skeletal muscle MFN2 is affected by acute (28) and chronic exercise (26), both demonstrated to increase insulin sensitivity. The combination of lower MFN2 protein content in AAW and positive correlation with insulin sensitivity across all subjects provides supportive evidence to the notion of a direct link between skeletal muscle MFN2 content and insulin sensitivity. Taken together, our data suggest a dysregulation of the proteins involved in fission, fusion, and autophagy in skeletal muscle from AAW. The alterations of these proteins likely contribute to the observed phenotype of decreased insulin sensitivity, resting metabolic rate, and maximal aerobic fitness.

This study is not without limitations. We acknowledge that the samples size is relatively low to establish any firm conclusions regarding the contributions of altered gene expression in skeletal muscle from AAW to the observed phenotype. However, the fact that our subjects were matched for BMI, physical activity, and age (10) significantly limits several confounding factors to this type of analysis. Moreover, racial differences in gene expression remained even after adjusting for levels of vigorous physical activity. It is accepted that there were very few genes demonstrated to be lower in AAW than in CW despite substantial differences in insulin sensitivity (26%) and mitochondrial respiration (25%). To bolster the evidence of altered gene expression in these lean AAW, we have added an analysis of complementary proteins known to contribute to alteration in insulin sensitivity and mitochondrial respiration. Although this is not an all-inclusive analysis, these data provide the foundation for additional studies to better understand the racial disparity in insulin resistance and increased risk for the development of type 2 diabetes in AAW.

In conclusion, the incidence of obesity, reduced insulin sensitivity, and risk for development of type 2 diabetes is higher in AAW than in CW. The etiology of this racial disparity is currently unknown. Our novel data provide evidence for alterations in mitochondrial gene expression and protein content in several pathways associated with mitochondrial function including small molecule, inner, and outer membrane transport as well as fission, fusion, and autophagy in skeletal muscle from lean AAW. Furthermore, these pathways are associated with RMR and maximal aerobic capacity, key factors associated with weight gain and insulin action. Although the exact mechanism contributing to the alterations in these genes and proteins is unknown, this may provide the foundation leading to the increased risk for development of type 2 diabetes in AAW.

Glossary

Abbreviations

AAW

African American women

BMI

body mass index

CW

Caucasian women

RMR

resting metabolic rate

PLSC

partial least-squares correlation

VO2max

maximal aerobic fitness

Additional Information

Disclosure Summary: The authors report no conflicts of interest.

Data availability: The datasets generated during and/or analyzed during the current study are not publicly available but are available from the corresponding author on reasonable request.