Mitochondrial Homeostasis in Obesity-related Hypertriglyceridemia.

CONTEXT: The prevalence of obesity and hypertriglyceridemia is an alarming worldwide health issue. Mitochondria play a central role in these disorders as they control cell metabolism.
OBJECTIVE: The aim of the present study was to characterize mitochondrial homeostasis in subcutaneous and visceral adipose tissue (SAT and VAT) in grade III obese patients with and without hypertriglyceridemia. Moreover, this study presents the evaluation of mitochondrial fitness as a marker for hypertriglyceridemia improvement. PATIENTS: Eight control and 12 hypertriglyceridemic (HTG) grade III obese subjects undergoing bariatric surgery were included. MAIN OUTCOME MEASURES: Anthropometric and biochemical data were obtained before and 3 months after surgery. Mitochondrial homeostasis was evaluated by mitochondrial DNA (mtDNA), gene expression and protein abundance in SAT and VAT.
RESULTS: Mitophagy-related gene expression was increased in HTG SAT and VAT, while mitochondrial marker gene expression and mtDNA were decreased, indicating an altered mitochondrial homeostasis in HTG. Mitophagy protein abundance was increased in VAT of those subjects that did not improve their levels of triglycerides after bariatric surgery, whereas mitochondrial protein was decreased in the same tissue. Indeed, triglyceride levels positively correlated with mitophagy-related genes and negatively with mitochondrial content markers. Moreover, mitochondria content and mitophagy markers seem to be significant predictors of hypertriglyceridemia and hypertriglyceridemia remission.
CONCLUSIONS: Mitochondrial homeostasis of adipose tissue is altered in hypertriglyceridemic patients. At the protein level, mitochondria content and mitophagy are potential markers of hypertriglyceridemia remission in obese patients after bariatric surgery. These results may contribute to the implementation of a clinical approach for personalized medicine.

Obesity is one of the most challenging health issues for our society because of its high world prevalence and its association with an increased risk of morbidity and mortality. Obesity is related to a great number of comorbidities such as hypertriglyceridemia (HTG), type 2 diabetes, cardiovascular diseases, and insulin resistance, among others (1). HTG is a metabolic disorder with increasing prevalence worldwide, ~10% in the adult population with interregional variation (2). HTG is associated with an increased risk of acute pancreatitis and cardiovascular events and with increased mortality (3). Clearly, obesity and HTG are serious public health problems.

It is well known that the main source of fatty acids in the liver come from white adipose tissue (WAT) (4), which are assembled with glycerol and apo B (5). WAT, besides its role in energy management by storing energy excess after feeding or by releasing lipids during periods of energy deprivation, acts as endocrine organ that participates in body energy homeostasis. The failure in any of these tasks leads to body metabolic disruption. However, the precise molecular phenomena occurring during WAT dysfunction remain unclear. Two types of WATs can be distinguished according to its location, visceral adipose tissue (VAT) and subcutaneous adipose tissue (SAT), both display different functional characteristics (6). VAT is more metabolically active, more sensitive to lipolysis. and has a greater capacity to generate free fatty acids (FFA) than SAT, whereas SAT is more avid in circulating FFA and triglyceride absorption.

One of the main features of obesity is the increase of adipocyte number and size in WAT because of an increase in triglyceride accumulation leading to hyperplasia and/or adipocyte hypertrophy (7, 8). These expanded adipocytes become frequently dysregulated (9), in response to unfavorable changes in the systemic endocrine milieu, in part through an increased rate of local necrosis, autophagy, apoptosis, and pro-inflammatory responses. An abnormal enlargement of the adipocyte size could cause lipotoxicity in surrounding organs because of the release of FFA into the bloodstream (10-12). Because the adipocyte is a highly active secretory cell, the secretory machinery gets easily overloaded, leading to altered release of adipokines (13). Mitochondria, which are at the heart of metabolism, become dysfunctional under these conditions. Mitochondrial dysfunction enhances FFA release into the blood streams, which results in lipotoxicity. Moreover, imbalanced lipid metabolism can, in turn, alter mitochondria homeostasis (14). Furthermore, it is well known that mitochondrial dysfunction triggers oxidative stress because the cell is incapable of restoring the cellular metabolism and controlling the cell survival. Therefore, it is not surprising that an accumulation of dysfunctional mitochondria is linked to different diseases such as cardiovascular disease, lung disease, and metabolic disorders (11, 15-19). To restore the energy balance occurring under oxidative stress, cells activate autophagy as a defense mechanism (20). Dysfunctional mitochondria are eliminated by mitophagy (21), a specific autophagic response confined to mitochondria, which represents 1 of the major pathways of mitochondrial quality control. Because WAT represents around 10% of total body weight in lean adults, but can achieve > 50% in obese subjects (22), it is not surprising that any obesity-induced changes in WAT mitochondria can substantially disrupt whole-body energy homeostasis, including changes in serum triglycerides.

Despite the crucial role that mitochondria play in WAT physiology, the processes involved in WAT mitochondrial homeostasis and the role in mitophagy in human obesity and hypertriglyceridemia are still poorly understood. Therefore, the aim of the present study is to characterize the status of mitochondrial homeostasis in SAT and VAT in morbidly obese patients with and without HTG. Moreover, this study presents the evaluation of basal mitochondrial homeostasis as a marker for hypertriglyceridemia improvement after weight loss through bariatric surgery.

Materials and Methods

Subjects

A total of 20 grade III obese patients, who were scheduled for bariatric surgery at the “Virgen de la Victoria Clinical University Hospital,” were included in the present study between 2016 and 2019. Participants were grouped according to their triglyceride levels in subjects with or without HTG (defined as triglyceride levels > 200 mg/dL) and paired with body mass index (BMI). Eight patients were non-HTG subjects (controls) and 12 were HTG patients (Table 1). Three months after surgery, HTG subjects were evaluated to assess HTG remission, which was triglyceride levels lower than 150 mg/mL (Table 2). Seven subjects improved their triglyceride levels (IMP HTG) and 5 subjects did not improve their levels (non-IMP HTG).

Table 1.

Clinical and metabolic variables for each study group

	Non-HTG (n = 8)	HTG (n = 12)	P value
Sex (M/W)	2/6	3/9	0.704
Age (y)	41 ± 6	48 ± 9	0.058
BMI (kg/m2)	48 ± 7	51 ± 8	0.483
Waist (cm)	128 ± 19	138 ± 14	0.198
Hip (cm)	137 ± 23	149 ± 12	0.158
Waist to hip ratio	0.94 ± 0.13	0.93 ± 0.09	0.840
Glucose (mg/dL)	93 ± 5	103 ± 11	0.035
Triglycerides (mg/dL)	113 ± 18	272 ± 62	n/a
Cholesterol (mg/dL)	196 ± 3.6	209 ± 38	0.481
HDL-c (mg/dL)	54 ± 9	37 ± 4	0.001
LDL-c (mg/dL)	114 ± 27	125 ± 40	0.505
Total proteins (g/dL)	7.2 ± 0.5	7.0 ± 0.4	0.417
Uric acid (mg/dL)	5.0 ± 0.7	5.2 ± 1.2	0.670
Urea (mg/dL)	30 ± 6	32 ± 6	0.354
Creatinine (mg/dL)	0.72 ± 0.16	0.77 ± 0.17	0.479
SBP (mmHg)	141 ± 10	131 ± 17	0.213
DBP (mmHg)	82 ± 4	83 ± 8	0.834
GOT (U/L)	18 ± 6	22 ± 7	0.167
GPT (U/L)	38 ± 17	40 ± 14	0.771
GGT (U/L)	33 ± 20	43 ± 28	0.238

Data are means ± SD.

Abbreviations: BMI, body mass index; DBP, diastolic blood pressure; GOT, glutamic oxaloacetic transaminase; GGT, gamma-glutamyl transferase; GPT, glutamic pyruvic transaminase; HDL-c, high-density lipoprotein cholesterol; LDL-c, low-density lipoprotein cholesterol; n/a, not available; SBP, systolic blood pressure.

Table 2.

Clinical and metabolic parameters for hypertriglyceridemia group after bariatric surgery divided in 2 groups related to the improvement or not in their levels of triglycerides

	IMP HTG			Non-IMP HTG
	Basal	3 months	P	Basal	3 months	P
Sex (M/W)	2/5			1/4
Age (y)	50 ± 8			45 ± 10
BMI (kg/m2)	52 ± 9	41 ± 6	0.000	50 ± 9	42 ± 8	0.000
Triglycerides (mg/dL)	245 ± 33	145 ± 31	0.003	308 ± 78	296 ± 57	0.818
Cholesterol (mg/dL)	198 ± 40	151 ± 44	0.017	223 ± 32	227 ± 59	0.923
HDL-c (mg/dL)	37 ± 4	38 ± 10	0.819	37 ± 4	39 ± 9	0.686
LDL-c (mg/dL)	118 ± 36	83 ± 35	0.030	127 ± 48	143 ± 51	0.566

Data are means ± SD.

Abbreviations: BMI, body mass index; DBP, diastolic blood pressure; GOT, glutamic oxaloacetic transaminase; GGT, gamma-glutamyl transferase; GPT, glutamic pyruvic transaminase; HDL-c, high-density lipoprotein cholesterol; LDL-c, low-density lipoprotein cholesterol; SBP, systolic blood pressure.

The patients completed a structured interview to obtain the following data: sex, age, medical history, and drug consumption. All subjects underwent a standardized anthropometric examination: weight, height, blood pressure, waist and hip circumferences, and biochemical parameters (Table 1). Exclusion criteria included diabetes, lipid-lowering medication, cardiovascular disease, acute disease or chronic inflammatory disease, severe coagulopathy, or tobacco or alcohol abuse. All participants gave their informed consent, and the study was reviewed and approved on December 28, 2015, by the ethics and research committee of Virgen de la Victoria Clinical University Hospital (Malaga, Spain).

Human Sample Collection

Peripheral venous blood samples were collected after a 12-hour overnight fast before surgery and 3 months after surgery. Serum and plasma were separated in aliquots and immediately frozen at -80°C until their analysis. Biochemical parameters were quantified as part of routine patient management. Some parameters were not measured in all the samples; for those variables whose missing data was less than 20%, multiple imputation by chained equations was performed. The calculations were obtained using the R package “mice” version 3.14.0.

SAT and VAT samples were obtained during sleeve gastrectomy after a 12-hour overnight fast. The surgeon aimed to obtain adipose tissue biopsies at similar anatomic places in all subjects. Adipose tissue biopsies were washed in 0.9% saline solution, immediately frozen by immersion into liquid nitrogen, and stored at -80°C until submitted to RNA, DNA, and protein extraction.

Gene Expression

RNA was extracted from 100 mg of SAT and VAT samples by using manually extraction with TRIzol Reagent (Invitrogen, Carlsbad, CA, USA). RNA was reversed transcribed into cDNA by using “PrimeScript RT Master Mix (Perfect Real Time)” (TAKARA, Kusatsu, Shiga Prefecture, Japan) in a 2720 Thermal Cycler (Applied Biosystems, Foster City, CA, USA).

Real-time quantitative PCR (RT-qPCR) was carried out with 10 ng cDNA per each sample of subcutaneous and visceral adipose tissue for genes related to mitophagy (PTEN-induced kinase 1 [PINK1] PINK1, parkin [PARK2], FUN14 domain-containing 1 [FUNDC1], autophagy related 13 [ATG13], voltage-dependent anion channel 3 [VDAC3]), mitochondrial fusion (mitofusin 1 [MFN1], mitofusin 2 [MFN2]), and mitochondrial biogenesis and function (cardiolipin synthase 1 [CRLS1], citrate synthase [CS], nuclear respiratory factor 1 [NRF1), peroxisome proliferator activated receptor gamma coactivator 1 alpha), WAT development (homeobox C8 [HOXC8], transcription factor 21 [TCF21]), and adipose tissue metabolism (leptin [LEP]). All primers were synthesized by Integrated DNA Technologies, Inc. The sequence of the primers used in this study can be found in Table 3. RT-qPCR was performed using SYBR green (TB Green Premix Ex Taq, Tli RNaseH Plus, TAKARA), carried out on 7500 Fast Real-Time PCR System (Applied Biosystems). Quantification of transcript level by RT-PCR was done by using the relative Ct (ΔΔCt) method. mRNA transcripts were normalized to L7.

Table 3.

List of primers

Primer	Forward	Reverse
ATG13	AACCAGTGAGGGAAGCACTG	CTTCAGAAGGGTCCTGGCTC
CRLS1	ATGACGAGAATTGGCTTGGC	TTTGATTGGCCCAGTTTCGA
CS	GGCCATTGACTCTAACCTGG	CACTTACATTGCCACCCTCA
FUNDC1	AAAGCAGCACCTGAAATCAACA	AAGCCTCCCACAAATCCACTG
HOXC8	TCCCAGCCTCATGTTTCCAT	TGAGAGACTTCAATCCGACG
L7	ACCTGCAGAACCCAAATTGG	TTGACGAAGGCGAAGAAGCT
LEP	TGTGGCTTTGGCCCTATCTT	ACCGGTGACTTTCTGTTTGGA
MFN1	ATGACCTGGTGTTAGTAGACAGT	AGACATCAGCATCTAGGCAAAAC
MFN2	CTCTCGATGCAACTCTATCGTC	TCCTGTACGTGTCTTCAAGGAA
mt143bp	CCACTGTAAAGCTAACTTAGCATTAACC	GTGATGAGGAATAGTGTAAGGAGTATGG
NRF1	GCTGATGAAGACTCGCCTTCT	TACATGAGGCCGTTTCCGTTT
PARK2	GTGTTTGTCAGGTTCAACTCCA	GAAAATCACACGCAACTGGTC
PINK	CCCAAGCAACTAGCCCCTC	GGCAGCACATCAGGGTAGTC
PPARGC1A	GCTTTCTGGGTGGACTCAAGT	GAGGGCAATCCGTCTTCATCC
RNU2	GGATTTTTGGAGCAGGAG	CTGCAATAGCAGGTCGATGC
TCF21	CAGATCCTGGCTAACGACAA	CGGTCACCACTTCTTTCAGG
VDAC3	AACGATGGCACTGAATTTGGA	AACGGGTGTTGTTACTCCCAG

Mitochondrial DNA

DNA was isolated from 40 mg of SAT and VAT samples by using the QIAamp DNA Mini Kit (QIAGEN, Germany) according to the manufacturer’s instruction. Relative amounts of mtDNA (mt143bp) and nuclear DNA were determined by RT-qPCR. The RT-qPCR was performed with 4 ng of total DNA using SYBR green (TB Green Premix Ex Taq, Tli RNaseH Plus), carried out on QuantStudio 6 Pro Real-Time PCR System (Applied Biosystems). Primers were designed by Oligo 7 software and synthesized by Integrated DNA Technologies, Inc; their sequences are shown in Table 3. Results are expressed as the ratio (ΔCt) between nuclear DNA and mitochondrial DNA (mt143bp).

Western Immunoblotting

Samples of SAT and VAT homogenate (100 mg) were prepared for western immunoblotting in protein lysis buffer (Tris-HCl 1M pH 8.5, NaCl 5M, 2% SDS, EDTA 0.5M, 1% Triton X-100, containing 1% each of phosphatase inhibitor cocktail I and II, and protease inhibitor; Sigma, UK.) Samples were centrifuge at 12 000g for 5 minutes and the fat layer created on the top was removed. Total protein quantification was carried out using DC Protein Assay (BioRad, US). All the samples were equalized for protein, added to 4× SDS sample buffer (composition: Tris-HCl 100 mM, pH 6.8, 4% SDS, 2% bromophenol blue, 20% glycerol; Sigma, UK), boiled (95°C, 5 minutes) and applied to 10% to 15% SDS gels. Proteins were transferred to polyvinylidene difluoride membrane, nonspecific binding was blocked (5% Blotto non-fat dry [Santa Cruz Biotech, US] in Tris-buffered saline [TBS] containing 0.05% Tween 20 [TBS-T]) and membranes were incubated overnight at 4 °C with the antibodies raised against TOMM20 (1:1000; Cat# ab186735, RRID:AB_2889972), FUNDC1 (1:1000; Cat# ab272627, RRID:AB_2915943), phospho-PARKIN2 (1:1000; Cat# ab73016, RRID:AB_1269675), total-PARKIN2 (1:1000; Cat# ab77924, RRID:AB_1566559) and SMC3 (1:1000;Cat# 5696, RRID:AB_10705575); all IgG, raised in rabbit or mouse, diluted in 5% Blotto nonfat dry milk/TBS-T. Membranes were washed and incubated (room temperature, 2 hours) with a secondary horseradish peroxidase-linked anti-rabbit/anti-mouse antibody (1:2000 in 5% Blotto nonfat dry in TBS-T). Immunoreactive bands were detected using WesternBright enhanced chemiluminescent substrate (BioRad, US). Images were captured using the BioRad ChemiDoc MP Imaging System and densitometry analysis was carried out using ImageJ (http://rsb.info.nih.gov/). These analyses were running in parallel in both adipose tissues (SAT and VAT) and SMC3 marker was used as the loading control.

Statistical Analysis

The results are expressed as mean values ± SD unless otherwise stated. Clinical and metabolic data were analyzed by means comparison (Mann-Whitney U) between non-HTG and HTG subjects. Gene expression from adipose tissue samples were submitted to univariate repeated-measures general linear models followed by post hoc Bonferroni test. After testing the normal distribution of the continuous variables by the Shapiro-Wilk test, we applied logarithmic transformation as needed to ensure normality of skewed variables in repeated-measures general linear models analysis. Adipose tissue depots (SAT and VAT) were introduced as a within-subjects effect, group (non-HTG and HTG) was considered a between-subjects effect, and glucose levels as a covariable because the HTG group presented higher levels of glucose. Differences of protein abundance between HTG and non-HTG groups, gene expression, and protein abundance IMP HTG and non-IMP HTG groups were evaluated by independent samples t test. The level of significance was set at P < 0.05 for the main effects and the interaction. Relationships between gene expression, mtDNA or protein abundance and clinical/biochemical variables were analyzed by Spearman correlation test. To analyze the predictive value of mitochondrial markers of significant correlations, a simple linear regression analysis was carried out including groups of subjects (with and without HTG) as a dependent variable, and gene expression levels and BMI, and triglyceride (TG) levels were included as covariates. The statistical powers were calculated for the different approach, so gene and mtDNA expression showed a statistical power of 99% (error B = 0.01) assuming safety values of 80% and minimum detectable effect of 0.5 and bilateral design. On the other hand, protein abundance showed a statistical power of 90% (B = 0.1), assuming safety values of 80% and minimum detectable effect of 1. We used SPSS statistical software 26.0 (SPSS Inc., Chicago, IL, USA) for analyses.

Results

Basal Characteristics of Patients

The differences between non-HTG and HTG patients in clinical and metabolic variables are summarized in Table 1. Regarding anthropometric and clinical variables, as expected from our design, there were no differences in age in BMI between non-HTG and HTG patients. HTG group presented higher triglyceride levels, accompanied by a decrease in high-density lipoprotein (HDL) cholesterol levels and an increase in glucose levels.

Altered genes related to mitochondrial fitness and white adipose markers in HTG

The effect of HTG on the expression levels of genes related to mitochondrial fitness and adipose was analyzed (Fig. 1). We observed an effect of HTG increasing gene expression of PARK2 in SAT and PINK1 in both tissues (SAT and VAT), whereas ATG13 in VAT was decreased in the HTG group. No significant differences in FUNDC1 expression levels were observed in our studied groups (Fig. 1A).

Figure 1.

Expression profiles of genes related to mitochondrial and white adipose tissue processes (n = 8-12). RNA was extracted from VAT (visceral adipose tissue) and SAT (subcutaneous adipose tissue) of obese people without/with hypertriglyceridemia (non-HTG/HTG; n = 8-12). Gene expression levels were measured by RT-qPCR and relative quantification of the expression levels was performed by the comparative threshold cycle (Ct) method. (A) Genes involved in mitophagy (PARK2, PINK1, FUNDC1, and ATG13). (B) Genes involved in mitochondrial fusion (MFN1 and MFN2). (C) Genes involved in mitochondrial biogenesis and function (CS, CRLS1, NRF1, PPARGC1, and VDAC3). (D) Genes involved in adipose tissue development and metabolism (HOXC8, LEP, and TCF21). Data are expressed as the mean ± SD (*P < 0.05; ***P < 0.005).

No significant changes in 2 genes involved in the mitochondrial fusion process, MFN1 and MFN2, were found (Fig. 1B).

Several genes involved in mitochondrial biogenesis and function were analyzed. The expression levels of NRF1 and VDAC3 in both tissues were increased in the HTG group. There was a decrease in CRLS1 in both tissues of the HTG group compared with the non-HTG group. No significant differences in CS and PPARGC1A gene expression were observed in our studied groups (Fig. 1C).

The expression levels of 3 white adipose tissue development and metabolism markers was analyzed (Fig. 1D). There was an HTG effect in TCF21 gene expression in SAT, where the levels were lower in the HTG group compared with the non-HTG group. No effect was found in LEP and HOXC8 gene expression.

Moreover, the effect of the type of adipose tissue (SAT and VAT) on the expression of mitochondrial and adipose related genes was also analyzed (Fig. 2). VDAC3, PARK2, MFN2, LEP, and HOXC8 decreased in VAT compared with SAT, whereas TCF21 increased.

Figure 2.

Main differences found in gene expression between both adipose tissues (subcutaneous: SAT; visceral: VAT; n = 19-20). RNA was extracted from VAT and SAT of obese people without/with hypertriglyceridemia (non-HTG/HTG; n = 8-12). Gene expression levels were measured by RT-qPCR and relative quantification of the expression levels was performed by the comparative threshold cycle (Ct) method. Genes involved: mitochondrial fusion, mitophagy and function (MFN2, PARK2, VDAC3), and adipose tissue development and metabolism (HOXC8, LEP, TCF21). Data are expressed as the mean ± SD (*P < 0.05; *** P < 0.005).

mtDNA in HTG and its association with genes related to mitochondrial and adipose tissue fitness

The analysis of mtDNA expression in both experimental group (non-HTG and HTG) showed a decrease in VAT of HTG compared with the non-HTG group (Fig. 3). Among the significant correlations between mtDNA and gene expression, a negative association between mtDNA and NRF1 in SAT was found.

Figure 3.

Mitochondrial DNA (mtDNA) expression and its correlation with different genes relate with mitochondrial dynamics, biogenesis and function, and adipose tissue development and metabolism of obese without/with hypertriglyceridemia (non-HTG/HTG; n = 5-10). mtDNA copy number was measured by RT-qPCR in both adipose tissue (subcutaneous: SAT; visceral: VAT). Data are expressed as the mean ± SD (*P < 0.05). The heat maps represent the mtDNA and individual gene expression in rows and columns, respectively, from different adipose tissues (S: subcutaneous; V: visceral). Spearman correlation ratio is displayed on a color scale from blue (negative correlation) to red (positive correlation). Statistical significance is expressed as: *P < 0.05; **P < 0.01.

Altered proteins related to mitochondrial fitness in HTG

Several proteins related to mitochondrial abundance and mitophagy (FUNDC, TOMM20 p/tPARKIN2, and PINK1) were also evaluated in HTG. Although mitochondria abundance (TOMM20) was decreased in both tissues in HTG and mitophagy markers were increased in SAT of HTG group, these differences were not statistically significant (Fig. 4).

Figure 4.

Mitochondrial and mitophagy protein abundance in subcutaneous and visceral adipose tissue (SAT and VAT, respectively) of obese people without/with hypertriglyceridemia (non-HTG/HTG; n = 4-9). Western immunoblotting was used to analyze several mitochondrial proteins. FUNDC, PINK1, pPARKIN2, and tPARKIN2 were divided by TOMM20 to assess the quantity of mitophagy per mitochondria abundance. The left panel represent the data analyzed in SAT and the right panel data analyzed in VAT. A representative blot for every measure is displayed at the bottom of the figure. Data are expressed as the mean ± SD.

Relationship of mitochondrial fitness and white adipose markers with clinical and biochemical parameters

The relation between the levels of gene expression related to mitochondrial fitness and white adipose and the clinical and biochemical profile of the individuals were analyzed (Fig. 5A). BMI positively correlated with SAT PARK2. SAT HOXC8, MFN2, and PARK2 positively correlated with waist circumference, whereas VAT MFN1 was negatively correlated. A positive correlation was found between VAT MFN2 and PARK2 with the hip circumference. The waist hip ratio was negatively correlated with VAT MFN1 and positively correlated with SAT mtDNA. Systolic blood pressure was negatively correlated with SAT FUNDC1, VDAC3, and VAT MFN1. However, there was a positive correlation between VAT Park2 with diastolic blood pressure. Glucose levels negatively correlated with VAT CRLS, MFN1, SAT mtDNA and positively correlated with VAT HOXC8. The 2 parameters with more significant correlation were TG and HDL. On the 1 hand, TG negatively correlated with ATG13 (VAT), CRLS (both tissues), MFN1 (SAT), mtDNA (both tissues), and TCF21 (SAT), whereas the correlation was positive with FUNDC1 and HOXC8 in VAT, NRF1, PINK1, and VDAC3 in both tissues. On the other hand, a negative correlation was found in HDL with FUNDC1, HOXC8, and NRF1 in VAT, PPARGC1A in SAT, PINK1 and VDAC3 in both tissues, whereas CRLS (both tissues) and TCF21 (SAT) was positively correlated. Uric acid was negatively correlated with MFN1 in VAT. A negative correlation was found between SAT FUNDC1 and mtDNA urea, whereas VAT HOXC8 showed a positive correlation. Creatinine showed a positive correlation with VAT HOXC8. Glutamic oxaloacetic and glutamic pyruvic transaminase (GPT) were positively correlated with SAT MFN2 and PARK2, whereas VAT MFN1 was negatively correlated with GPT.

Figure 5.

Correlations between genes and protein related to mitochondria and white adipose tissue with clinical parameters. The heat map represents the clinical parameters and individual adipose tissue (S, subcutaneous; V, visceral) gene (A) and protein (B) expression in columns and rows, respectively. Spearman correlation ratio is displayed on a color scale from blue (negative correlation) to red (positive correlation). Statistical significance is expressed as: *P < 0.05; **P < 0.01.

To analyze the impact of the expression levels of genes related to mitochondrial biology, biogenesis and mitophagy, which were relevant to hypertriglyceridemia status (reported in Fig. 2), we used linear regression models. We included the expression of each of the relevant genes as independent variables in separate models and having or not hypertriglyceridemia as dependent variable. The regression model showed that gene expression of ATG13-V (R2 = 0.235, F = 6.9, β = -0.525; P = 0.017), CRLS1-S (R2 = 0.698, F = 44.9, β = -0.845; P = 0.000), CRLS1-V (R2 = 0.793, F = 73.9, β = -0.897; P = 0.000), NRF1-S (R2 = 0.326, F = 9.7, β = 0.603; P = 0.006), NRF1-V (R2 = 0.216, F = 6.0, β = 0.510; P = 0.026), PINK1-S (R2 = 0.473, F = 18.1, β = 0.708, P = 0.000), PINK1-V (R2 = 0.266, F = 7.9, β = 0.552; P = 0.012), VDAC3-S (R2 = 0.570, F = 26.1, β = 0.770; P = 0.000), and VDAC3-V (R2 = 0.318, F = 9.9, β = 0.595; P = 0.006) are significant predictors of hypertriglyceridemia. Moreover, mtDNA-visceral is also a significant predictor of hypertriglyceridemia (R2 = 0.328, F = 8.3, β = -0.611; P = 0.012).

Regarding the relation between mitophagy-related protein levels and clinical parameter measures in our individuals, the heat map shows marked correlations of several mitophagy proteins in the adipose tissue of the participants with their clinical parameters, with obvious tissue-related differences (Fig. 5B). BMI negatively correlated with different mitophagy markers in SAT: FUNDC, PINK, and TOMM20. A negative correlation was found between waist and hip circumference with FUNDC and TOMM20 in SAT. Cholesterol was correlated negatively with FUNDC and TOMM20, whereas low-density lipoprotein was negatively correlated with PINK1 in VAT. GPT negatively correlated with SAT PINK1 and positively with SAT pParkin2. Gamma-glutamyl transferase negatively correlated with FUNDC in VAT and TOMM20 in both tissues.

The Relationship Between Basal Mitochondrial Homeostasis and Clinical and Metabolic Changes 3 Months After Bariatric Surgery

The differences in clinical and metabolic variables of patients’ HTG 3 months after bariatric surgery between subjects who improved and did not improve are summarized in Table 2. Regarding clinical variable, both groups significantly decreased their BMI. Regarding metabolic parameters, triglyceride and cholesterol levels were decreased in IMP HTG group, whereas no significant changes were in non-IMP HTG group.

To evaluate the significance of basal mitochondrial gene expression and protein abundance on TG improvement, we analyzed those markers in adipose tissue of non-IMP HTG and IMP HTG groups. Regarding gene expression, just SAT MFN1 was decreased significantly in the non-IMP HTG group compared with the IMP HTG group. Regarding protein abundance, VAT TOMM20 was decreased, whereas VAT PINK1 and tPARKIN were increased in non-IMP HTG compared with IMP HTG group (Fig. 6).

Figure 6.

Mitochondrial and mitophagy protein abundance in subcutaneous and visceral adipose tissue (SAT and VAT, respectively) of obese people who improved or not improved hypertriglyceridemia after 3 months of bariatric surgery (IMP HTG; no-IMP HTG; n = 4-5). Western immunoblotting was used to analyze several mitochondrial proteins. FUNDC, PINK1, pPARKIN2, and tPARKIN2 were divided by TOMM20 to assess the quantity of mitophagy per mitochondria abundance. The left panel represents the data analyzed in SAT and the right panel data analyzed in VAT. A representative blot for every measure is display at the bottom of the figure. Data are expressed as the mean ± SD (***P < 0.005).

To analyze how predictive our gene and protein data can be used to determine improvement or not after the surgery, a Spearman’s correlation test was run with our data set. In this case, changes of triglycerides, HDL, low-density lipoprotein, cholesterol, and BMI 3 months after the surgery was used for the clinical variables. Figure 7A shows the heat map resulting from the analysis of gene expression. A positive correlation was found between HDL and SAT ATG13 and FUNDC1. Cholesterol was negatively correlated with VAT NRF1 and PARK2. A negative correlation was found between BMI and SAT MFN2, PINK1 and VDAC3. The same analysis was carried out at protein levels (Fig. 7B), where only a negative correlation was found between triglycerides and VAT FUNDC1 and a positive correlation between BMI and SAT tPARKIN.

Figure 7.

Correlations between genes and proteins related to mitochondria and white adipose tissue with clinical changes 3 months after bariatric surgery. The heat map represents the clinical parameters and individual adipose tissue (S, subcutaneous; V, visceral) gene (A) and protein (B) expression in columns and rows, respectively. Spearman correlation ratio is displayed on a color scale from blue (negative correlation) to red (positive correlation). Statistical significance is expressed as: *P < 0.05; **P < 0.01.

To further analyze the impact of the abundance of relevant mitochondrial and mitophagy protein on hypertriglyceridemia improvement after 3 months of bariatric surgery (reported in Fig. 7), we used linear regression models. These models proved that visceral protein abundance of TOMM20 (R2 = 0.741, F = 23.9, P = 0.002; β=–0.879, P = 0.002), tPARKIN (R2 = 0.456, F = 7.7, P = 0.027; β = 0.724, P = 0.027), and PINK1 (R2 = 0.443, F = 7.4, P = 0.030; β = 0.716, P = 0.030) are significant predictors of hypertriglyceridemia improvement.

Discussion

There is an increasing interest in studying mitochondrial fitness in metabolic diseases because of its role in energy homeostasis. In the present study, we corroborate its implication in HTG and how this could be an important biomarker to predict the benefits of a bariatric surgery in a weight-loss intervention. We analyzed SAT and VAT in parallel because of the differences found in the literature associated to metabolic disease. Although SAT seems to be the most affected in terms of quantity during the disease process, it has been shown that changes in VAT composition are linked to metabolic disease (23). An increase in blood TG levels is one of the main components in the development of metabolic disease (24, 25) and, therefore, we studied both tissues in parallel for this study. The importance of studying both tissues was clear after finding differences in leptin and TCF21 mRNA expression, typical markers of WAT, depending on the tissue. As reported by other authors, leptin mRNA levels were lower in VAT, whereas TCF21 mRNA levels were higher (26-28). Besides, we report differences between both adipose tissues in mitochondrial fusion and mitophagy gene expression with a decrease of their levels in VAT. However, when we analyzed how HTG could affect the expression of those mitochondria-related genes, we found a similar effect of HTG in both tissues, although it was more significant in SAT. HTG seems to increase mitophagy in the adipose tissue of obese individuals as indicated by the increasing in PARK2 and PINK expression, two important genes related to mitophagy, especially PARK2 because it has been shown in different in vitro models as a requirement for starting the mitophagy process (29, 30). In fact, the lack of this gene (PARK2) in animal studies protects against the deleterious effects of a high-fat diet such as weight gain, fat mass, and insulin level increase (31), suggesting the importance of this gene in obesity development. This was corroborated by the increasing in NRF1, which regulates the expression of both genes (32). In addition, Hang et al showed how an increase in TG levels through a high-fat diet or a cell treatment with fatty acids increased the expression of both markers in an animal model that again corroborates our results (33). Although we found a decrease in ATG13 mRNA expression, 1 of the main autophagosome component to carry out autophagy, it has been shown by other authors that this is not essential for the PINK1-Parkin-dependent mitophagy process (34). It seems clear that most of the mitophagy-related genes analyzed in this study are positively correlated with the TG levels suggesting that mitochondrial impairment in adipose tissue is related to the metabolic state of the obese individual. In fact, the regression analysis illustrated the predictive value of some mitochondrial-related genes (ATG13, CRLS1, NRF1, PINK1, VDAC3) in hypertriglyceridemia development. It is known that lipids are accumulated in the adipose tissue and liver as TG when they are not needed. An excess of TG in the adipose tissue may cause an increase in oxidative stress, which further affects the mitochondrial well-being. In fact, some mutations are found in mtDNA related with higher TG levels in healthy donors (35). Although we only reached the statistical significance in VAT, mtDNA levels were lower in those patients with HTG. This implies a decrease in the number of mitochondria that could be related with an increase in mitophagy (21). Indeed, we found a negative correlation of mtDNA with TG levels suggesting an increase in mitophagy in those patients with higher TG levels and, therefore, a worse mitochondrial response. Different studies have shown changes of mtDNA levels in obesity and related diseases in human and rodent models, showing a decrease in the number of copies in the adipose tissue (36-40). Although mtDNA levels only changed in VAT under HTG condition, it was negatively correlated with the expression of NRF1 in SAT, as mentioned previously, the master regulator of PARK-PINK-pathway for starting the mitophagy process. Moreover, our data from the regression analysis proof that mtDNA was also a hypertriglyceridemia predictor.

To check the predictor value of beneficial improvement after the bariatric surgery, we investigated the correlation between the perioperative mitochondrial markers with the biochemical changes 3 months after surgery. Nevertheless, there were no correlation between the levels of TG with the mitochondrial-related genes; however, it was a negative correlation with FUNDC protein expression. Analyzing the protein abundance in HTG group depending on their improvement in TG levels, we found lower TOMM20 abundance in VAT of those individual with no HTG improvement. TOMM20 was also lower in the adipose tissue of obese individuals by other authors (37), but its abundance under HTG condition was never investigated before. Besides, TOMM20, tPARKIN2, and PINK1 was corroborated as a predictor of hypertriglyceridemia improvement by the lineal regression analysis. Therefore, our result suggests a reduction in mitochondria number following the previous results in mtDNA, where we also showed a decrease in its expression. The increasing of PINK1 and tPARKIN in VAT of non-IMP HTG group corroborate the increase in mitophagy. The increase in mitophagy of patients who did not improve HTG after bariatric surgery could be explained as a compensatory effect of mitochondrial failure because PARKIN-PINK1-pathway has been shown as an important player in different animal models for preventing metabolic disorders through mitophagy activation (41, 42). On the other hand, different studies showed how a mitochondrial impairment increase TG levels (43, 44) so that might shift the question toward the level of mitochondrial failure and its relationship with obesity. Thus, it seems that HTG causes mitochondrial fitness impairment and that the effect of bariatric surgery on the improvement of TG levels could be related to the basal mitochondrial fitness of each person. Unfortunately, these adipose samples were unavailable. Further studies would be needed to understand if there is a dysfunctional mitophagy process in those individuals who did not improve their TG levels after the surgery or in the contrary, the mitophagy was effective but without a mitochondrial turnover. Therefore, analyzing mitochondrial fitness after the surgery in those individuals would shed light on this issue.

Conclusions

Mitochondrial fitness of adipose tissue is altered in hypertriglyceridemic obese patients. Protein level, mitochondria content, and mitophagy are potential markers of hypertriglyceridemia remission in obese patients after bariatric surgery. Hence, our results may contribute to the implementation of a clinical approach for a personalized medicine of obesity and hypertriglyceridemia.