Characterization of thoracal fat depots – expression of adipokines and remodeling factors and impact of adipocyte conditioned media in fibroblast scratch assays

Abstract

Adipose tissue is not only a connective tissue but also an endocrine organ secreting adipokines like Leptin and Adiponectin, lipokines such as palmitoileic acid and extracellular vesicles. These factors and the expression of matrix remodeling enzymes impact surrounding tissues via paracrine effects. The expression of selected secretion factors and the effect of adipocyte conditioned media from four thoracal adipose tissue origins - subcutaneous, perivascular, pericardial and epicardial adipose tissues – in a fibroblast proliferation/wound healing scratch assay model were investigated. Results were compared directly and according to the type 2 diabetic mellitus (T2DM) status of the patients the tissues are originated from. Adipocyte conditioned media from non-diabetic patients resulted in a significant higher scratch closure rate compared to the media with T2DM background. Linoleic acid incubation in scratch assay resulted in a reduced scratch closure rate. Leptin, Adiponectin and Visfatin/Nampt expression and MMP2, MMP9 and FSTL1 mRNA levels did not vary according to T2DM subgroups directly, leading to the assumption that these factors are not causal for scratch assay effects observed. In contrast significant mRNA expression differences were monitored between the thoracal tissue origins implying variations in the local effects of the different adipose tissue depots.

Keywords

Thoracal adipose tissue depots adipokines MMP scratch wound healing assay fatty acids type 2 diabetes mellitus

1 Introduction

For decades the function of adipose tissue was mainly assigned as energy storage and supporting or connective tissue and a depot for nutrients [1, 2]. The fact that numerous factors are secreted by adipocytes has become scientific focus in the 1990s and adipose tissue has been reclassified as an endocrine organ [1 , 4]. A decisive role in this process played the discovery of Leptin that is secreted by adipose tissue and regulates processes such as food uptake and appetite and energy expenditure by signaling the nutritional status [2 , 5]. Low levels of Leptin result in increased food uptake and reduced energy expenditure and Leptin levels in obese persons are reduced or a leptin resistance can be monitored [1, 6]. Also the adipokine Adiponectin is decreased in obesity and beneficial metabolic effects such as insulin sensitivity and inflammation processes are reduced [1 , 5]. Elevated levels of plasma adiponectin are associated with low risk for type 2 diabetes [7]. A growing number of adipokines are published and function is described to be involved in multiple organ systems like brain, pancreas, heart and vasculature and immune system [1, 4]. An adipokine with enzymatic activity is Visfatin/Nampt that has been discussed to be a visceral-enriched activator of insulin receptors [8] but that is also expressed by other adipose tissues [9]. Not only adipokines are secreted by adipose tissue but also cytokines (e.g. TNF-α and IL-6), growth factors (e.g. VEGF), lipokines (e.g. palmitoileic acid) and extracellular vesicles [1 , 10]. Protocols for the generation of adipose tissue conditioned media and the challenges are described in literature [11 –14] and secreted factors are investigated in conditioned media resulting from different fat depots such as visceral fat, omental adipose tissue, breast tissue or perivascular fat tissue [11 , 16]. An impact of secreted factors on various organ systems and (patho)physiological processes can be assumed [1 , 10]. Closely associated are metabolic syndrome/disease with type 2 diabetes mellitus (T2DM), cardiovascular risk and wound healing processes for all of which it has been shown that conditioned media and therewith adipose tissue secreted factors are relevant mediators [10]. More than 2 billion people in the world are obese accompanied by diseases like T2DM and metabolic syndrome, atherosclerosis and myocardial infarction [17]. But not the Body Mass Index (BMI) per se is representative to define adiposity with risk for secondary diseases but the abdominal vs. gluteal localization and visceral vs. subcutaneous distribution of adipose tissue and therefore the waist to hip circumference is of importance [18 –21]. Due to the endocrine function of adipose tissue the accumulation of fat in obesity leads to a pathophysiological secretion pattern depending on the adipose tissue depot increased [17, 19, 22]. Beside visceral and subcutaneous fat depots which exert systemic (and locally) effects by the release of adipokines, lipokines and cytokines there are also locally acting fat depots such as pericardial fat and perivascular fat [23 –26]. Perivascular fat is located near the vascular wall of large to small diameter arteries and impacts processes of atherosclerosis by secretion of adipokines and inflammatory factors [18, 25]. Epicardial fat tissue is a visceral fat depot between the myocardial surface and the visceral pericardium without boundary to myocardium [27, 28]. Therefore adipokines and other secreted factors may directly influence the myocardium and epicardial fat has key roles in the regulation of cardiovascular (patho)physiology [27, 28]. The developmental origin of epicardial fat is the splanchnopleuric mesoderm whereas the pericardial adipose tissue has its (ectodermal) origin in the primitive thoracic mesenchyme and is separated from the heart by the pericardium [24 , 27–30]. Thoracic fat depots from different origins and from non-diabetic (ND) vs. T2DM patients may vary in their secretion profile and their impact on surrounding tissues, cell proliferation and wound healing processes.

2 Objective

The objective of the study was to discover differences of mRNA-expression levels of adipokines and tissue remodeling factors such as MMP (matrix metalloprotease) 2 and 9 as well as FSTL1 (Follistatin-like 1) in subcutaneous, pericardial, epicardial and perivascular (here A. thoracica interna) fat depots originated from ND vs. T2DM patients. In addition it was hypothesized that adipocyte conditioned media generated from these four human fat tissues impact human fibroblast cell line BJ in a wound healing scratch assay according to patients T2DM status and this can be associated to expression profile.

3 Material and methods

3.1 Patient’s material

Adipose tissues of 29 patients with an average age of 72.6±9.4 years were implemented in the experimental setup. The study was approved by the ethics committee of the Dresden University (Ethikkommission an der TU Dresden, registration number EK190062013). Tissue was extracted according to routine aortocoronary bypass intervention in cardiac surgery. 25 of the patients were male and four female. 16 of the patients suffered from a T2DM disease and six of these patients were long term treated with insulin before the surgical intervention. 13 patients were grouped as non-diabetics (ND). The average age of the T2DM patients was 75.9±6.3 years vs. 68.6±10.6 years in the ND group.

3.2 Preparation of adipose tissue samples

Adipose tissue samples (subcutaneous, perivascular, pericardial and epicardial depots) were directly transferred into 10 ml Ringer’s solution during surgical procedure in the theatre. A second washing step in Ringer’s solution followed to eliminate blood cells. After macroscopically dissection of connective tissue with scalpels and forceps under sterile conditions, tissue aliquots for qRT-PCR were branched off and shock frozen in liquid nitrogen. These samples were stored at –80°C until RNA isolation was performed. Average masses of 412±238 mg, 713±449 mg, 275±252 mg and 199±123 mg were obtained from subcutaneous, pericardial, perivascular and epicardial locations. For enzymatic release of adipocytes from the fat tissue sample the remaining tissue probe was incubated in 8.2 ml of modified Ringer’s solution (plus 1.5 M NaH₂PO₄ and 0.7 M Na₂HPO₄) containing 0.34 mPZUnits/ml collagenase NB4 (Serva, Heidelberg, Germany) and 0.93 M glucose for 3 h at 37°C resulting in an almost completely homogeneous suspension of cells. Suspension was filtered using a nylon filter to separate adipocytes from remaining/connective tissue; filter was washed twice with modified Ringer’s solution to include adipocytes preliminary stuck in the filter. The volume of cell suspension was determined.

3.3 Isolation of RNA

RNA isolation from adipose tissue was performed with the RNeasy Fibrous Tissue Mini Kit (Qiagen, Hilden, Germany) according to the manufacturer’s instructions. RNA concentration was determined using Qubit Flourometer after probes were diluted in 200 μl TAE-buffer containing 1 μl QuantiFluor^® RNA Dye.

3.4 Quantitative real-time-RT-PCR

3.4.1 cDNA synthesis

mRNA template was adopted according to concentration and transcribed into cDNA using Maxima First Strand cDNA Synthesis Kit for RT-qPCR (Thermo Fisher Scientific, Waltham, USA) according to manufacturer’s instructions. Oligo(dT)₁₈ and random hexamer primers are implemented and reaction was performed as follows: denaturation 94°C for 15 s, primer annealing 55°C for 15 s and elongation at 72°C for 1.5 h.

3.4.2 qRT-PCR

Complementary DNA was amplified using GoTaq^® qPCR Master Mix (Promega, Madison, USA) according to manufacturer’s instructions. 7.5 μl reaction master mix (2x) were used for a total volume of 15 μl containing template and 500 nM of forward and reverse primer, respectively. The reaction was performed in a Rotor-Gene-Q (Qiagen) each sample was performed in triplicates. For calibration curves different dilutions up to 1:1000 were measured and efficiencies were calculated using Rotor-Gene Q software. As reference genes YWHAZ (tyrosine 3/tryptophan 5-monooxygenase activation protein), PGK1 (phosphoglycerate kinase 1) and HPRT (hypoxanthine phosphoribosyl – transferase) were used to relativize expression of genes of interest: Leptin, Adiponectin Visfatin/NAMPT, MMP2 (Matrix Metalloproteinase 2), MMP9 and FSTL1 (Follistatin-like 1). Primer sequences are listed in Table 1. In addition in each reaction an internal control was implemented to calculate a correction factor. On the basis of the standard curves a virtual concentration of cDNA quantity of each sample was determined.

Table 1
Primer sequences used for qRT-PCR

Gene Primer Sequence (5’-3’) Annealing temperature (°C)

Leptin forward primer GGCTTTGGCCCTATCTTTTC 58

reverse primer ACCGGTGACTTTCTGTTTGG

ADIPOQ forward primer CCTGGTGAGAAGGGTGAGAAAG 61

reverse primer AGATCTTGGTAAAGCGAATGGGCAT

NAMPT forward primer GCCAGCAGGGAATTTTGTTA 60

reverse primer TGATGTGCTGCTTCCAGTTC

MMP2 forward primer GACATCAAGGGCATTCAGGAGC 60

reverse primer CGTCACAGTCCGCCAAATGAAC

MMP9 forward primer TGCGTCTTCCCCTTCACTTTC 59

reverse primer GAACAAACTGTATCCTTGGTCCG

FSTL1 forward primer GCAGTCACAGAGAAAGGGGA 58

reverse primer GCAGTGTCCATCGTAATCAACCTG

YWHAZ forward primer GGTCTGGCCCTTAACTTCTCTG 56

reverse primer TCTCCTTGGGTATCCGATGTCC

PGK1 forward primer AGAGCCCAGAGCGACCCT 58

reverse primer AAGTGCCAATCTCCATGTTGTTGAGC

HPRT forward primer GGCTCCGTTATGGCGACC 58

reverse primer AGCCCCCCTTGAGCACAC

Gene	Primer	Sequence (5’-3’)	Annealing temperature (°C)
Leptin	forward primer	GGCTTTGGCCCTATCTTTTC	58
	reverse primer	ACCGGTGACTTTCTGTTTGG
ADIPOQ	forward primer	CCTGGTGAGAAGGGTGAGAAAG	61
	reverse primer	AGATCTTGGTAAAGCGAATGGGCAT
NAMPT	forward primer	GCCAGCAGGGAATTTTGTTA	60
	reverse primer	TGATGTGCTGCTTCCAGTTC
MMP2	forward primer	GACATCAAGGGCATTCAGGAGC	60
	reverse primer	CGTCACAGTCCGCCAAATGAAC
MMP9	forward primer	TGCGTCTTCCCCTTCACTTTC	59
	reverse primer	GAACAAACTGTATCCTTGGTCCG
FSTL1	forward primer	GCAGTCACAGAGAAAGGGGA	58
	reverse primer	GCAGTGTCCATCGTAATCAACCTG
YWHAZ	forward primer	GGTCTGGCCCTTAACTTCTCTG	56
	reverse primer	TCTCCTTGGGTATCCGATGTCC
PGK1	forward primer	AGAGCCCAGAGCGACCCT	58
	reverse primer	AAGTGCCAATCTCCATGTTGTTGAGC
HPRT	forward primer	GGCTCCGTTATGGCGACC	58
	reverse primer	AGCCCCCCTTGAGCACAC

3.5 Cell culture

3.5.1 Culture of adipocytes and generation of conditioned media

100 μl of isolated adipocytes were cultured per well of a 24 well plate containing 1 ml DMEM/HAM-F12 (4.5 % Glucose, 2 mM L-Glutamine, 1% Penicillin/Streptomycin and 10% FCS; Gibco/Thermo Fisher Scientific). Three wells were prepared, remaining adipocytes were stored at –80°C. If volume of generated adipocytes was lower than 100 μl the respective media volume was adopted. Adipocytes were incubated in a humidified atmosphere in a cell culture incubator with 5% CO₂. Medium was conditioned for 16 hours and was directly after incubation time used in scratch assay experiments or it was stored at –80°C until experiment was performed.

3.5.2 Culture of human fibroblast line BJ

The human fibroblast cell line BJ (ATCC, Manassas, USA) was cultured in EMEM containing 1% non-essential amino acids and 10% FCS (EMEM-NEAA-FCS; Gibco/Thermo Fisher Scientific) in a humidified atmosphere in a cell culture incubator at 5% CO₂. Passages six to ten were used for scratch assay experiments.

3.6 Scratch assay

BJ cells were seeded with a defined cell number of 22×10³ cells/cm² in 24 well plates and cultured for 48 h in EMEM containing 1% non-essential amino acids and 10% FCS. If a uniform confluency of 95% was reached wells were used for scratch assay. Scratch was set using a pipet tip as a cross through the whole well. Medium including cells and clusters that were scraped of was replaced by probe of interest or control media. Adipocyte conditioned media were diluted 1:4 with serum-free EMEM-NEAA. Assay was performed in triplicates (three independent wells per conditioned medium sample). Final serum content was 2.5%. Two different control conditions were investigated in parallel: On the one hand EMEM-NEAA without serum and on the other hand medium containing 2.5% serum. For the analysis of the impact of insulin concentrations of 0.1, 1 and 25 μM (stock solution 1.76 mM; Sigma Aldrich, Darmstadt, Germany) were diluted in EMEM-NEAA. Before EMEM-NEAA containing insulin was added scratch was set and the well was washed twice with PBS to wash out remaining FCS. To determine the effect of relevant fatty acids these were resolved in EMEM-NEAA containing 5% BSA (Serva). These stock solutions were further diluted to a final concentration of 50 μM. Relevant substance concentration levels were determined using CellTiter-Blue^® Cell Viability Assay (Promega) and Rotitest^®-Vital assay (Carl Roth GmbH, Karlsruhe, Germany) in 96 well format (not shown). As controls for the insulin condition EMEM-NEAA containing equivalent concentrations of HCl and for the fatty acid setup equivalent concentrations of BSA were investigated and used to calculate relative changes in the scratch closure. Using a CK 30–F200 (Olympus, Tokyo, Japan) microscope and a AM7023 Dino-Eye Eyepiece Camera (Dino-lite Digital Microscope, Hsinchu, Taiwan) scratch area was monitored and measured with the Fiji (https://fiji.sc/) software at time points 0 h, directly after the media of interest were added and after 8 h incubation. To define the same regions the scratched cross in the well was used for orientation. Area of scratch was determined at 0 h and the area of scratch closure was calculated in relation to the not cell covered area at 8 h. In addition the scratch closure rate was related on the basis of the internal control.

3.7 Statistical analysis

Differences between the thoracal adipose tissue depots within one group were tested by One-Way-ANOVA (Kruskal-Wallis- and Dunn’s multiple comparison tests). Direct comparison of gene expression or scratch closure rate was analyzed by Two-Way-ANOVA (Sidak‘s multible comparison test). Analyses were performed in Graph Pad Prism 7 software (GraphPad Software, Inc.).

4 Results

4.1 Expression patterns of thoracic adipose tissues

The relative mRNA expression levels of adipokines Leptin, Adiponectin, Visfatin/NAMPT and remodeling proteins MMP2, MMP9, FSTL1 was determined in the four thoracic adipose tissue depots – subcutaneous, perivascular, pericardial and epicardial fat. Numerous significant differences were detected in this analysis (Supplementary Figure 1). Adipokine Leptin was with 0.8±0.6 significantly lower expressed in perivascular fat tissue compared to epicardial adipose tissues (1.2±0.5; p≤0.05), whereas Adiponectin was higher expressed in pericardial fat tissue (1.5±0.5) compared to all other three tissue origins (p≤0.05 subcutaneous and epicardial; p≤0.0001 perivascular fat tissue). Visfatin/NAMPT was significantly lower expressed in the peri- and epicardial adipose tissue depots (1.3±0.4; p≤0.0001 and 1.5±0.5; p≤0.05 and p≤0.01, respectively) compared to subcutaneous (2.3±1.1) and perivascular (2.3±0.9) fat samples, which exhibit comparable expression levels. In contrast in peri- and epicardial adipose tissue matrix remodeling factor MMP2 was with 1.1±0.5 (p≤0.01) and 1.0±0.5 (p≤0.05) significantly higher expressed than in the perivascular adipose tissues (0.6±0.5) but in comparable amounts as in subcutaneous fat depot. In general very low levels of MMP9 expression were detected with highest amount in subcutaneous fat (0.4±0.8) but due to the high variance with no significant differences when compared to the other adipose tissues. The matrix remodeling factor FSTL1 was with 0.5±0.2 significantly lower expressed in epicardial fat depots compared to pericardial (p≤0.0001), perivascular (p≤0.05) and subcutaneous (p≤0.001) adipose tissues.

The patients investigated were grouped according to their diabetic disease status. 16 of the 29 patients exhibit a T2DM, 13 were non-diabetic (ND) patients. No direct significance between adipose tissue depots of the same anatomic localization from ND vs. T2DM patients was determined (Fig. 1). However the significant difference between perivascular and epicardial Leptin expression was not observed in T2DM samples but only in the ND sample set (0.7±0.4 vs. 1.3±0.5; p≤0.05). Pericardial adipose tissues of both groups, T2DM and ND patients, expressed significantly higher levels of Adiponectin than perivascular fat depots, but this expression levels in the subgroups did not differ significantly from levels for subcutaneous and epicardial adipose tissues as in the dataset investigating all patients together. Also there were no differences in the Visfatin/NAMPT expression levels of the different adipose tissue origins of T2DM vs. ND patients. In both groups Visfatin/NAMPT expression was lower in adipose tissues close to the heart compared to subcutaneous and perivascular tissue. Regarding the remodeling enzymes MMP perivascular fat depot expressed significant lower levels (0.5±0.5) of MMP2 compared to subcutaneous (1.2±0.9; p≤0.05) and pericardial (1.2±0.4; p≤0.01) adipose tissue in samples of ND patients but expression did not differ in T2DM specimen. MMP9 expression was very low with highest levels in the subcutaneous fat depots. A significant difference was detected between subcutaneous and pericardial adipose tissue. FSTL1 was significantly lower expressed in epicardial adipose tissue compared to pericardial fat independent from diabetic disease.

Fig.1

Relative mRNA expression of Leptin, Adiponectin, Visfatin/NAMPT, MMP2, MMP9 and FSTL1 in subcutaneous, perivascular, pericardial and epicardial adipose tissue from non-diabetic (ND) vs. patients with diabetes mellitus type 2 (T2DM). Expression was related to expression of reference genes YWHAZ, PGK1 and HPRT and calculated according to standard curve. Results in diagram were tested by Kruskal-Wallis- and Dunn’s multiple comparison tests. No significant differences were detected between the expression levels of thoracic depots itself when grouped according to diabetic status tested with Two-Way-ANOVA.

Patients were also grouped according the BMI of 27 and therefore adiposity. Leptin expression was significantly different in the BMI>27 subgroup comparing subcutaneous (1.4±0.7), pericardial (1.6±0.9; p≤0.01) and epicardial (1.1±0.4; p≤0.01) adipose tissue with the lower expressing perivascular (0.6±0.4) samples (Fig. 2). Leptin mRNA levels did not differ in the group with BMI lower than 27. Adiponectin was expressed with lowest levels in perivascular adipose tissue samples and for patients with BMI>27 this value (0.6±0.4) did not only differ significantly comparing pericardial (1.4±0.3; p≤0.0001) adipose tissue (also observed in the group with BMI<27) but also subcutaneous tissue (1.1±0.4; p≤0.05). Visfatin/NAMPT mRNA expression was independent from the subdivision according to the BMI. MMP2 differed significantly (p≤0.05) between the perivascular and the pericardial adipose tissues of patients with BMI>27 but not in patients with a BMI<27 and no differences were detected for MMP9 expression. FSTL1 expression was not significantly different in the BMI<27 subgroup. Comparable to the T2DM group in the group exhibiting a BMI>27 a significant lower expression in the epicardial (0.4±0.2) vs. the pericardial (1.3±0.8; p≤0.001) and subcutaneous (1.1±0.6; p≤0.01) adipose tissue depots were monitored.

Fig.2

Relative mRNA expression after BMI subclassification according to Fig. 1.

4.2 Scratch assay closure during incubation with adipocyte conditioned media

Adipocytes were obtained from thoracic tissue samples with yields of 0.48±0.23 μl/mg for subcutaneous and 0.54±0.33 μl/mg for pericardial fat depot and lower values for epicardial (0.31±0.22 μl/mg) and perivascular adipose tissue (0.25±0. 23 μl/mg). Conditioned media were generated by incubation of a defined adipocyte volume in DMEM/HAM-F12 for 16 h. The scratch assay was performed with human fibroblasts BJ. In general a slightly but insignificant higher scratch closure was monitored for all incubations with adipocyte conditioned media (115.9±30.6% subcutaneous, 119.8±37.3% perivascular, 115.1±30.3% pericardial and 107.2±21.0 % epicardial adipose tissue; Supplementary Figure 2). The patients were grouped according to their T2DM status. Thirteen individuals did not suffer from diabetic disease and sixteen patients were T2DM-patients. Investigating the ND patients for all thoracic adipose tissues a significant higher scratch closure was detected reaching from 132.1±34.9% (p≤0.001) after incubation with conditioned media resulting from adipocytes generated from subcutaneous fat to 128.6±32.7% (p≤0.01) from pericardial, 120.0±16.5% (p≤0.05) from perivascular and 119.4±33.8% (p≤0.05) from epicardial fat tissue. This positive impact on the scratch closure was not observed after incubation with media conditioned by adipocytes from T2DM patients adipose tissues (Fig. 3A). Values from 112.7±40.3% after incubation with conditioned media from pericardial adipose tissue to 96.8±18.7% for perivascular fat were determined with no significant changes compared to control condition. Scratch closure was significantly higher after incubation with conditioned media from subcutaneous (132.1±34.9 % vs. 102.7±19.1%; p≤0.0001) and or perivascular (120.0±16.5 % vs. 96.8±18.7 %; p≤0.01) adipocytes from ND patients, respectively and directly compared to the particular T2DM patient group (Fig. 3B). By further subclassification of T2DM group into patients with insulin medication (116.5±17.5%) vs. patients that were not treated with insulin (94.5±15.3%; p≤0.0001) the significant difference in scratch closure between subcutaneous fat conditioned media from ND vs. T2DM patients was abrogated when tissue from insulin treated patients was used. In other words the advantage for scratch closure was observed after incubation with conditioned media resulting from subcutaneous adipose tissue from ND patients and T2DM patients with insulin medication but not for T2DM patients that were not treated with insulin. In contrast investigating perivascular adipose tissue with 96.25±21.2% (p≤0.01) for T2DM patients not treated with insulin and 97.8±15.4% (p≤0.05) for T2DM patients with insulin medication comparable scratch closure rates were detected for both arms (Fig. 3C).

Fig.3

Closure of BJ fibroblast monolayer scratch eight hours after scratch setting and incubation with adipocyte conditioned media in relation to control media incubation. A) Scratch closure rate after incubation with adipocyte conditioned media from subcutaneous, perivascular, pericardial and epicardial origin respectively, on the one hand from non-diabetic patients and on the other hand from patients with T2DM. Statistical testing was performed using Kruskal-Wallis- and Dunn’s multiple comparison test. B) Direct comparison of subcutaneous and perivascular adipose tissue depots from ND vs. T2DM patients. Two-Way-ANOVA with Sidak’s multiple comparison test was applied. C) T2DM patients were further subclassified according to their medication with insulin and scratch closure rates of subgroups were tested according to B).

4.3 Scratch assay closure during incubation with insulin and fatty acids

To investigate a potential direct effect of insulin on the scratch closure in scratch assay using BJ human fibroblast insulin at concentrations of 0.1, 1 and 25 μM were analyzed in an assay setup not containing FCS. A significant increase in scratch closure was detected at a concentration of 0.1 μM with a scratch closure of 148.9±30.4% (p≤0.05; n = 4) but not at 1 and 25 μM (120.2±25.1% and 129.9±45.8%, respectively; Fig. 4A).

Fig.4

Scratch closure rates in BJ monolayer cultures eight hours after scratching and incubation with Insulin and fatty acids. A) Scratch closure rate after incubation with different concentrations of insulin in comparison to control media. Statistical significance was tested with Kruskal-Wallis- and Dunn’s multiple comparison test. B) Impact of fatty acid incubation (elaidic, linoleic, palmitic, palmitelaidic and stearic acid) on scratch closure rate was investigated and statistical significance was tested with Kruskal-Wallis- and Dunn’s multiple comparison test.

In order to detect an impact of potential lipokine function on BJ scratch assay a preliminary set of fatty acids (palmitic [C16:0], palmitelaidic [C16:1], stearic [C18:0], elaidic [C18:1] and linoleic acid [C18:2]) was investigated for effects on scratch closure. On the basis of cell viability assays (not shown) the concentration for the fatty acid incubation was set to 50 μM for all fatty acids implemented. A non-significant reduction of BJ scratch closure was detected after incubation of palmitic (63.3±12.0%, n = 5), palmitelaidic (65.4±16.3%) and elaidic acid (78.4±23.6%). Incubation with linoleic acid resulted with 33.6±9.3% (p≤0.01) in a significant reduction of scratch closure rate. Merely stearic acid incubation did not reduce the scratch closure (103.5±46.8%; Fig. 4B).

5 Discussion

The study presented herein investigates four thoracic fat depots for differences in expression of adipokines and matrix remodeling factors and the impact of adipocyte-conditioned media in wound healing relevant experimental setup of scratch assay implementing fatty acids and the direct effect of insulin. Differences in expression profile and regarding effects of adipocyte secreted factors between the adipose tissue depots and according to the diabetic disease status were hypothesized. This is one of the first studies directly comparing four various thoracic adipose tissue depots. Therefore and accompanying cardiac surgery including sternotomy adipose tissues specimen from epicardial, pericardial, perivascular and subcutaneous localization were extracted. Aliquots of each tissue specimen were used to isolate mRNA on the one hand. On the other hand and regarding the relevance of adipose tissue for aspects of fibroblast proliferation and migration and therewith on wound healing, adipocyte-conditioned media were generated and used to investigate their impact on a scratch-assay setup with human fibroblast cell line BJ. A non-significant increase was monitored for scratch closure rate 8 h after addition of the conditioned media, independent from the thoracic adipose tissue depot. But when individuals were grouped according their T2DM status conditioned media of all adipose tissue depots of ND patients resulted in a significantly beneficial effect on scratch closure, a higher scratch closure rate. This positive effect was not detected if conditioned media of T2DM patients was analyzed leading to the conclusion that adipose tissue of T2DM patients is altered. The scratch closure rate after incubation with conditioned media from subcutaneous and perivascular adipose tissue from ND patients was significantly higher than after incubation with material from T2DM patients in a direct comparison. In subcutaneous adipose tissue depot this difference was dependent on the patient’s medication with insulin. It is hypothesized that the secretion profile of adipokines, remodeling factors or fatty acids changed due to disease T2DM status. These changes might lead to a reduction of the positive proliferative effect on human fibroblasts. These results are in contrast to literature data comparing conditioned media from adipose tissue, adipocytes and adipose-derived stem cells in experimental scratch assay setups where no significant impact of adipocyte conditioned media on different skin cell fractions was observed [14]. Nevertheless adipose tissue origin, extraction and differences in experimental procedure but also in the incubation times can be causal for varying results. Adipocyte secretome may include adipokines, lipokines or enzymatic mediators. To mimic advantageous effects, in a first step in the same assay (BJ scratch assay) initial experiments were performed to investigate the direct influence of fatty acids. The incubation with linoleic acid resulted in a significantly reduced scratch closure rate in BJ cells. Other fatty acids co-incubated did not change the scratch closure rate significantly leading to the conclusion that these fatty acids are not relevant for the monitored effect in this certain experimental setup. The selection of fatty acids was done according to literature data describing relevant changes in the fatty acid profiles comparing epicardial and subcutaneous adipose tissue depots from ND and T2DM patients [31]. Pezeshkian et al. published a difference of fatty acids such as palmitic, stearic and linoleic acid in ND vs. T2DM patients profile [31].

Subsequent to investigate expression profiles of selected adipokines and matrix remodeling factors and relate these differences to the scratch assay results, mRNA levels of adipokines Leptin, Adiponectin and Visfatin/NAMPT and matrix remodeling factors MMP2/9 and FSTL1 were quantified via qRT-PCR. Leptin, Adiponectin and Visfatin/NAMPT are differently expressed in the thoracic depots especially the epicardial and perivascular tissues but also the pericardial data show different patterns in the analysis presented herein. Literature comparisons of expression pattern for epicardial and perivascular adipose tissue depots are summarized in Schäfer et al 2017 [26]. Adiponectin expression is generally reduced in obesity and in coronary artery disease in epicardial or perivascular adipose tissue [26, 29]. Leptin expression was increased in coronary artery disease and reduced in general in these adipose tissues and Visfatin/NAMPT was increased in coronary artery disease [26]. Metaanalysis (n = 601) of adiponectin expression in epicardial fat revealed a 0.9-fold decrease compared to expression level in subcutaneous adipose tissues [32]. In contrast it has been shown that diabetic and ND subjects express similar epicardial and subcutaneous adipose tissue adiponectin and leptin levels [33]. Recently the epicardial adipose tissue transcriptome of patients with or w/o T2DM and coronary artery disease was investigated in comparison to subcutaneous depot [34]. In this study, diabetic epicardial fat demonstrates a transcriptome markedly different from that of subcutaneous adipose tissue in the same subjects showing enrichment in genes involved in inflammation, innate immune response and endothelium damage [34]. In the analysis presented herein a significant lower expression of Leptin was detected comparing perivascular and epicardial adipose tissue, adiponectin was significantly higher in pericardial adipose tissue compared to subcutaneous, perivascular and epicardial fat and Visfatin/NAMPT was significantly lower expressed in peri-and epicardial fat tissue than in the subcutaneous and perivascular counterparts. But after subgrouping the patients according to T2DM disease status and statistical analysis no direct differences between the adipose tissue depots itself were monitored. This is a similar result as in literature where no significant differences in adiponectin and leptin levels comparing epicardial and subcutaneous fat according to diabetic disease were detected [33]. Also leptin expression in ND vs. T2DM adipose tissue depots could not be associated with the higher scratch closure rate detected for ND patients in scratch assay although leptin has been described to be relevant for wound healing processes analyzed in animal model and keratinocyte scratch assay [35, 36]. In addition the adiponectin expression did not differ according to T2DM status and could therefore not be the reason for scratch assay differences, although also for adiponectin a positive contribution on wound healing effects has been described previously [37, 38]. However the differences detected between the thoracic adipose tissue depots but also the tissue quantity might influence fibroblast proliferation and the individual patient wound healing process.

The significances detected between the four thoracic fat depots varied. Leptin was significantly lower expressed in perivascular vs. epicardial fat solely in the ND patients group but not in the T2DM subgroup, due to a lower expression in epicardial tissue. Adiponectin differs in both groups significantly comparing perivascular and pericardial adipose tissue and Visfatin/NAMPT expression pattern is comparable. MMP2 expression that is significantly lower in perivascular adipose tissue compared to peri-and epicardial fat when all specimen were analyzed together showed no significant difference when T2DM patients were statistically tested. In the ND patients group perivascular MMP2 expression was lower in perivascular depot and significantly different when compared with subcutaneous and pericardial fat. MMP9 mRNA expression varied between perivascular and pericardial fat, expressing higher amount in the perivascular depot in ND patients and lower amounts in the T2DM group, although expression of MMP9 per se was low. Both enzymes MMP2 and 9 also called gelatinase A and B are involved in wound healing processes and overexpression has been shown in wound fluids [39, 40]. Overexpression of MMP2 and 9 in endothelial cells with diabetic background and higher serum level of these enzymes are described in literature and additional experimental data relate the two gelatinases MMP2 and 9 to diabetic disease [41].

Taken together in this initial number of samples significant differences in the expression of relevant adipokines and matrix remodeling factors in the different adipose tissues were defined but the data reveal with a patient’s number of 29 the main limitation of the study – the number of specimen included. A larger number of samples would also enable further sub-classification according to cardiovascular disease or other disease relevant parameters. In addition more adipokines and remodeling factors but also transcription factors can be implemented in ongoing research projects [34]. Nevertheless a paracrine impact of adipocytes via secretion in culture media on a human fibroblast culture was demonstrated.

Adipose tissue, its paracrine and endocrine secretion of mediators and effectors attach more and more importance for diseases such as metabolic syndrome and therewith adiposity, T2DM and cardiovascular disease. To clearly define expression profiles and secreted factors as well as effects on cellular systems may contribute to understand the complete clinical disease pattern and as a result to define supportive treatment strategies.

Footnotes

Acknowledgments

The excellent technical assistance of Maria Feilmeier and Tina Kapell is acknowledged.

The supplementary figures are available in the electronic version of this article: .

References

Choi

CHJ

, Cohen

. Adipose crosstalk with other cell types in health and disease. Exp Cell Res. 2017.

Trujillo

, Scherer

. Adipose tissue-derived factors: Impact on health and disease. Endocr Rev. 2006;27(7):762–78.

Gonzalez

, Moreno-Villegas

, Gonzalez-Bris

, Egido

, Lorenzo

. Regulation of visceral and epicardial adipose tissue for preventing cardiovascular injuries associated to obesity and diabetes. Cardiovasc Diabetol. 2017;16(1):44.

Bluher

, Mantzoros

. From leptin to other adipokines in health and disease: Facts and expectations at the beginning of the 21st century. Metabolism. 2015;64(1):131–45.

Cao

. Adipocytokines in obesity and metabolic disease. J Endocrinol. 2014;220(2):T47–59.

Maffei

, Halaas

, Ravussin

, Pratley

, Lee

, Zhang

, et al. Leptin levels in human and rodent: Measurement of plasma leptin and ob RNA in obese and weight-reduced subjects. Nat Med. 1995;1(11):1155–61.

, Shin

, Ding

, van Dam

. Adiponectin levels and risk of type 2 diabetes: A systematic review and meta-analysis. JAMA. 2009;302(2):179–88.

Fukuhara

, Matsuda

, Nishizawa

, Segawa

, Tanaka

, Kishimoto

, et al. Visfatin: A protein secreted by visceral fat that mimics the effects of insulin. Science. 2005;307(5708):426–30.

Berndt

, Kloting

, Kralisch

, Kovacs

, Fasshauer

, Schon

, et al. Plasma visfatin concentrations and fat depot-specific mRNA expression in humans. Diabetes. 2005;54(10):2911–6.

10.

Dai

, Zhang

, Yu

, Tian

. Therapeutic applications of conditioned medium from adipose tissue. Cell Prolif. 2016;49(5):561–7.

11.

Fletcher

, Sacca

, Pistone-Creydt

, Colo

, Serra

, Santino

, et al. Human breast adipose tissue: Characterization of factors that change during tumor progression in human breast cancer. J Exp Clin Cancer Res. 2017;36(1):26.

12.

Alvarez-Llamas

, Szalowska

, de Vries

, Weening

, Landman

, Hoek

, et al. Characterization of the human visceral adipose tissue secretome. Mol Cell Proteomics. 2007;6(4):589–600.

13.

Skalnikova

, Motlik

, Gadher

, Kovarova

. Mapping of the secretome of primary isolates of mammalian cells, stem cells and derived cell lines. Proteomics. 2011;11(4):691–708.

14.

Kober

, Gugerell

, Schmid

, Zeyda

, Buchberger

, Nickl

, et al. Wound healing effect of conditioned media obtained from adipose tissue on human skin cells: A comparative in vitro study. Ann Plast Surg. 2016;77(2):156–63.

15.

Maury

, Ehala-Aleksejev

, Guiot

, Detry

, Vandenhooft

, Brichard

. Adipokines oversecreted by omental adipose tissue in human obesity. Am J Physiol Endocrinol Metab. 2007;293(3):E656–65.

16.

Drosos

, Chalikias

, Pavlaki

, Kareli

, Epitropou

, Bougioukas

, et al. Differences between perivascular adipose tissue surrounding the heart and the internal mammary artery: Possible role for the leptin-inflammation-fibrosis-hypoxia axis. Clin Res Cardiol. 2016;105(11):887–900.

17.

Gonzalez-Muniesa

, Martinez-Gonzalez

, Hu

, Despres

, Matsuzawa

, Loos

RJF

, et al. Obesity. Nat Rev Dis Primers. 2017;3:17034.

18.

Akoumianakis

, Antoniades

. The interplay between adipose tissue and the cardiovascular system: Is fat always bad? Cardiovasc Res. 2017;113(9):999–1008.

19.

Kim

, Despres

, Koh

. Obesity and cardiovascular disease: Friend or foe? Eur Heart J. 2016;37(48):3560–8.

20.

Nishida

, Ko

, Kumanyika

. Body fat distribution and noncommunicable diseases in populations: Overview of the WHO Expert Consultation on Waist Circumference and Waist-Hip Ratio. Eur J Clin Nutr. 2010;64(1):2–5.

21.

Heitmann

, Lissner

. Hip Hip Hurrah! Hip size inversely related to heart disease and total mortality. Obes Rev. 2011;12(6):478–81.

22.

Vegiopoulos

, Rohm

, Herzig

. Adipose tissue: Between the extremes. EMBO J. 2017;36(14):1999–2017.

23.

Bays

. Central obesity as a clinical marker of adiposopathy; increased visceral adiposity as a surrogate marker for global fat dysfunction. Curr Opin Endocrinol Diabetes Obes. 2014;21(5):345–51.

24.

Iacobellis

. Epicardial adipose tissue in endocrine and metabolic diseases. Endocrine. 2014;46(1):8–15.

25.

Siegel-Axel

, Haring

. Perivascular adipose tissue: An unique fat compartment relevant for the cardiometabolic syndrome. Rev Endocr Metab Disord. 2016;17(1):51–60.

26.

Schafer

, Drosos

, Konstantinides

. Perivascular adipose tissue: Epiphenomenon or local risk factor? Int J Obes (Lond). 2017;41(9):1311–23.

27.

Patel

, Shah

, Verma

, Oudit

. Epicardial adipose tissue as a metabolic transducer: Role in heart failure and coronary artery disease. Heart Fail Rev. 2017.

28.

Nagy

, Jermendy

, Merkely

, Maurovich-Horvat

. Clinical importance of epicardial adipose tissue. Arch Med Sci. 2017;13(4):864–74.

29.

Iacobellis

. Local and systemic effects of the multifaceted epicardial adipose tissue depot. Nat Rev Endocrinol. 2015;11(6):363–71.

30.

Iacobellis

. Epicardial and pericardial fat: Close, but very different. Obesity (Silver Spring). 2009;17(4):625; author reply 6–7.

31.

Pezeshkian

, Mahtabipour

. Epicardial and subcutaneous adipose tissue Fatty acids profiles in diabetic and non-diabetic patients candidate for coronary artery bypass graft. Bioimpacts. 2013;3(2):83–9.

32.

Yim

, Rabkin

. Differences in gene expression and gene associations in epicardial fat compared to subcutaneous fat. Horm Metab Res. 2017;49(05):327–37.

33.

Teijeira-Fernandez

, Eiras

, Grigorian-Shamagian

, Salgado-Somoza

, Martinez-Comendador

, Gonzalez-Juanatey

. Diabetic and nondiabetic patients express similar adipose tissue adiponectin and leptin levels. Int J Obes (Lond). 2010;34(7):1200–8.

34.

Camarena

, Sant

, Mohseni

, Salerno

, Zaleski

, Wang

, et al. Novel atherogenic pathways from the differential transcriptome analysis of diabetic epicardial adipose tissue. Nutr Metab Cardiovasc Dis. 2017;27(8):739–50.

35.

Tadokoro

, Ide

, Tokuyama

, Umeki

, Tatehara

, Kataoka

, et al. Leptin promotes wound healing in the skin. PLoS One. 2015;10(3):e0121242.

36.

Murad

, Nath

, Cha

, Demir

, Flores-Riveros

, Sierra-Honigmann

. Leptin is an autocrine/paracrine regulator of wound healing. FASEB J. 2003;17(13):1895–7.

37.

Kawai

, Kageyama

, Tsumano

, Nishimoto

, Fukuda

, Yokoyama

, et al. Effects of adiponectin on growth and differentiation of human keratinocytes–implication of impaired wound healing in diabetes. Biochem Biophys Res Commun. 2008;374(2):269–73.

38.

Jin

, Xiao

, Ge

, Zhan

, Zhou

. Role of adiponectin in adipose tissue wound healing. Genet Mol Res. 2015;14(3):8883–91.

39.

Iuonut

A-MM

, Dindelegan

. C. C. Proteases As Biomarkers In Wound Healing. . Timisoara medical Journal. 2011; 61(1-2).

40.

Rohani

, Parks

. Matrix remodeling by MMPs during wound repair. Matrix Biol. 2015;44-46:113–21.

41.

Hopps

, Caimi

. Matrix metalloproteinases in metabolic syndrome. Eur J Intern Med. 2012;23(2):99–104.