No evidence for change in expression of TBC1D1 and TBC1D4 genes in cultured human adipocytes stimulated by myokines and adipokines

ABSTRACT TBC1D1 and TBC1D4 proteins play analogous, but not identical role in governing insulin-signalling pathway. Little is known about changes in expression levels of TBC1D1 and TBC1D4 genes in mammals, including humans. Number of factors were studied, but data remain controversial. The aim of this study was to evaluate the effect of selected cytokines, adipokines and myokines with known or putative insulin sensitivity regulation activity (adiponectin, irisin, omentin, interleukin 6, leptin, resistin, and tumour necrosis factor) on TBC1D1 and TBC1D4 expression levels in cultured differentiated human adipocytes. No significant differences were found between the adipocytes treated with different stimuli and this effect was determined not dose dependent. It is reasonable to conclude that relative shortage of data showing no change in TBC1D1 and TBC1D4 from literature results from publication bias; therefore, our finding provides additional insight into the role of both genes.


Introduction
Two members of Tre-2/BUB2/CDC16 domain family -TBC1D1 (TBC1 domain family member 1) and TBC1D4 (TBC1 domain family member 4, also known as Akt substrate of 160 kDa, AS160) proteins play analogous, but not identical role in governing insulin-dependent docking and membrane fusion of glucose transporter GLUT4 containing vesicles, sequestered in intracellular compartment [1,2]. TBC1D1 and TBC1D4 are key regulators of GLUT4 redistribution and serve as Rab GTPase activating proteins (Rab GAPs) [3], sharing Rab substrate specificity [4,5]. Unphosphorylated, basal state TBC1D1 and TBC1D4 remain active, promoting their Rab substrates hydrolytic activity and keeping them in inactive (GDPbound) forms, leading to GLUT4 sequestration. Phosphorylation, in response to glucose uptake stimuli, as insulin and muscle contraction, cause inhibition by phosphoprotein-binding protein 14-3-3 capture, and allows Rabs to become GTP-bound (active) and facilitate GLUT4 liberation to the plasma membrane. Recent data from murine model suggests co-operation between Tbc1d1 and Tbc1d4 in GLUT4 trafficking, with dominant role of Tbc1d1, and Tbc1d4 serving as precise modulator of insulin sensitivity [6].
TBC1D1 and TBC1D4 exhibit distinct abundance among insulin responsive tissues. Both proteins are highly, but variably, expressed in skeletal muscles [2,, 7]. Rodent Tbc1d4 protein is expressed in low levels in preadipocytes, but is induced during differentiation [8], whereas Tbc1d1 is present in murine adipocytes in much less abundance than its paralogue [9], and Tbc1d1 gene shows different splicing pattern than in muscle cells [7]. Human subcutaneous adipose tissue exhibits opposite patterns of TBC1D1 and TBC1D4 expression than rodents -levels of TBC1D1 mRNA was shown to be threefold greater than TBC1D4 [10].
Little is known about changes in expression levels of TBC1D1 and TBC1D4 genes. Number of factors were studied, but data remains controversial. Diabetes was shown to down-regulate the expression of TBC1D4 in skeletal muscle and adipose tissue of humans and rodents [11,12,13,14]. Tbc1d1 was shown to be up-regulated in diabetic rat muscle [15]. One study showed no significant change in both TBC1D1 and TBC1D4 protein abundance in diabetic patients [16]. To the best of our knowledge, the expression regulation of TBC1D1 in mammalian adipose tissue was not studied before. Therefore, we aimed to investigate the effect of selected cytokines, adipokines and myokines with known or putative insulin sensitivity regulation activity (adiponectin, cleaved FNDC5/irisin, intelectin 1/omentin, interleukin 6, leptin, resistin, and tumour necrosis factor) on TBC1D1 and TBC1D4 expression levels in cultured differentiated human adipocytes.

Tissue handling, primary cells isolation and culture
Abdominal subcutaneous tissue samples were obtained, with informed consent, from 5 non-obese, non-diabetic and cancer-free male adult subjects during elective abdominal surgery for cholelithiasis. None of the subjects presented active inflammatory process, assessed by normal blood C-reactive protein. The protocol was approved by local Ethics Review Board of Łódź Medical University number RNN/215/16/KE.
All tissue samples were transported to the cell culture unit in HBSS with antibiotics (penicillin, streptomycin and amphotericin B), and were subjects of immediate processing. All samples were mechanically minced and digested with 0.2% collagenase from Clostridium histolyticum in HBSS for 45 min in 37°C. The reaction was stopped by dilution 1:10 with PBS-EDTA and all mixtures were passed through 250 µm mesh, followed by filtering through 100 µm cell strainer and centrifugation. The pellets were resuspended in 1 ml of RBCL Buffer, diluted 1:10 with PBS and filtered through 70 µm cell strainer. Following second centrifugation, cell pellets were resuspended in 5 ml of DMEM-HG/Ham's F-12 with 10% FBS, antibiotics (penicillin, streptomycin and amphotericin B; after the first day of initial culture, amphotericin B was removed), and plated on T-25 dishes, preincubated for 2 hours in culture medium. Cells were maintained in for 2 weeks, until the first passage. After reaching 80% confluence, all cell lines were cryopreserved in culture medium supplemented with 10% DMSO for further procedures.

Adipogenic terminal differentiation and stimulation
Thawed cells were cultured in the same conditions as primary isolates. For adipogenic terminal differentiation, cells were seeded on T-25 dishes, media were replenished thrice a week. After reaching 80-90% confluence, adipogenic phenotype was induced in serumfree DMEM-HG/Ham's F-12-based differentiation medium, containing 0.5 mM IBMX, 100 nM human insulin, 100 nM dexamethasone, 2 nM T 3 , 1 µM rosiglitazone, 33 µM biotin, 17 µM pantothenic acid and 10 µg/mL iron-saturated transferrin. Cells were maintained in differentiation medium until all visible cells acquired adipocyte morphology (this time was individually variable from 14 to 21 days). Adipogenic differentiation was assessed by visual inspection and confirmed by qPCR for LEP gene expression. After adipocyte phenotype induction, cells were either maintained in serum-free DMEM-HG/Ham's F-12-based medium, supplemented with 10 nM human insulin, 10 nM dexamethasone and antibiotics for control samples or stimulated with chosen cytokine, myokine and adipokine for 24 h in the same maintenance medium. Cell lines, isolated from two individuals, replicated in two biological replicates, were used to determine most stable reference genes. Cell lines isolated from remaining three individuals, replicated in three biological replicates, were used to determine genes of interest (GOI) -TBC1D1 and TBC1D4 relative expression. For stable reference gene determination, stimulants (adiponectin, cleaved FNDC5/irisin, intelectin 1/ omentin, interleukin 6, leptin, resistin and tumour necrosis factor) were used in one (highest) concentration -100 ng/μL, for assessing GOI relative expression, every stimulant was used in three concentrations -25 ng/μL, 50 ng/μL and 100 ng/μL, concurrently with own control cell culture.

RNA purification, processing and qPCR
Total RNA from cultured cells was extracted using TRI Reagent. Purity and yield assessment was performed by spectrophotometric absorbance measurement at 230 nm, 260 nm and 280 nm. For qPCR analysis, RNA samples (1 to 3 µg) were digested with DNase I and reverse transcribed (iScript Advanced cDNA Synthesis Kit for RT-qPCR) according to manufacturer's instructions. cDNA samples (2 μl RT each reaction product) were amplified in triplicates, along with No-Template and No-Reverse Transcription controls, using iTaq Universal SYBR Green Supermix, according to manufacturer's recommendations on CFX96 Touch Platform (Bio-Rad). Primers, spanning exon-exon junction, were designed using free Primer-BLAST tool [17]. All primers used are listed with their sequence in Table  1. Five reference genes (B2M, UBC, TFRC, SDHA, and HPRT1) were tested to determine two most stably expressed in experimental conditions.

Data analysis
Quantification cycles (C q ) were calculated automatically by Bio-Rad CFX Manager Software v. 3.1. Both GOI were normalized to two most stably expressed genes in experimental condition (B2M and SDHA), as determined by geNorm algorithm [18]. Relative quantities were log 2 transformed. The relation between relative expression and stimulant concentration was assessed using ANOVA (all samples) and Student t test for Pearson correlation coefficient r (to determine dose dependency). P-value below 0.05 was considered significant. Results were reported as fold change. Statistical analysis was carried out using R (The R Project for Statistical Computing, www. r-project.org). Figures 1 and 2 present the effect of adiponectin, cleaved FNDC5 (irisin), intelectin 1 (omentin), interleukin 6, leptin, resistin and tumour necrosis factor on relative expression of TBC1D1 gene ( Figure 1) and TBC1D4 gene (Figure 3). No significant differences were found between the cultures treated with different stimuli. We observed a high dispersion of expression change. All effects were determined as not dose dependent.
Murine Tbc1d1 is expressed in adipocytes in lesser abundance than its paralogue Tbc1d4 [9]. Much less is known about changes in TBC1D1 expression in different physiological setups. Sáinz et al. described diminished expression of Tbc1d1 in ob/ob mice treated with leptin [20]. Tbc1d1 expression was significantly increased in diabetic Goto-Kakizaki rats muscle [15]. Preventive and therapeutic exercise group of rats (in comparison to sedentary group) exhibited increased Tbc1d1 expression in soleus and gastrocnemius muscles [32]. Myocardial infarct induces the expression of Tbc1d1 in murine heart muscle both in infarct zone and border zone [25]. Differences in Tbc1d1 expression in both adipose and muscle tissues were also noted between sexes in chicken model [31]. Expression changes of discussed genes remain controversial, and data from different studies might not be comparable due to differences in biological setups, including different stimuli as well as different material used (either biopsy specimen or cultured cells), interspecies variability and methodological issues. Being regulators of metabolism, both genes may be governed by different mechanisms in different species. It was shown that TBC1D1 variants are linked to obesity in many species, including human, rabbit and chicken, while deletion of murine Tbc1d1 suppressed obesity in ob/ob mice [35]. Kinase activation pattern, including Tbc1d1 phosphorylation in murine skeletal muscles, did not show the same pattern in human muscles [36]. Different adipose tissue depots exhibit different expressional pattern, this study was performed on in vitro differentiated subcutaneous adipocytes. Further work is needed to confirm the observation on different adipocytes populations. Nevertheless changes in TBC1D1 and TBC1D4 expression in human adipocytes were not subject of thorough studies. Changes in expression levels were rather noted in addition in studies of regulation of other genes, involved in insulin signalling pathway regulation. It is reasonable to conclude that relative shortage of data showing no change in TBC1D1 and TBC1D4 results from publication bias; therefore, our findings provide additional insight into the role of both genes. TBC1D1 as