ITIH5 inhibits proliferation, adipogenic differentiation, and secretion of inflammatory cytokines of human adipose stem cells—A new key in treating obesity?

Inter‐α‐trypsin inhibitor heavy chain 5 (ITIH5) is widely expressed in the human body, and it is detected to be particularly abundant in adipose tissue. ITIH5 expression is increased in people with obesity compared to lean persons and is decreased by diet‐induced weight loss. This suggests that ITIH5 may be involved in the development of adiposity and clinical metabolic variables, although its exact function remains unknown. We measured the protein concentration of ITIH5 in adipose samples from patients undergoing abdominoplasty and tested for correlation with the subjects' BMI as well as inflammatory mediators. We stimulated human adipose stem cells (ASCs) with recombinant (r)ITIH5 protein and tested for an effect on proliferation, differentiation, and immunosuppressive properties when the cells were exposed to an artificial inflammatory environment. We found positive correlations between ITIH5 levels and the BMI (p < .001) as well as concentrations of inflammatory cytokines (TNF‐α, IL‐6, and MCP‐1) in adipose tissue (p < .01). Application of the rITIH5 protein inhibited both proliferation (p < .001) and differentiation of ASCs. Especially, the development of mature adipocytes was reduced by over 50%. Moreover, rITIH5 decreased the release of IL‐6 and MCP‐1 when the cells were exposed to TNF‐α and IL‐1β (p < .001). Our data suggest that ITIH5 is an adipokine that is increasingly released during human adipose tissue development, acting as a regulator that inhibits proliferation and adipogenic differentiation of ASCs. ITIH5 thus presents itself as a positive regulator of adipose tissue homeostasis, possibly protecting against both hyperplasia and hypertrophy of adipose tissue and the associated chronic inflammation.

adipose samples from patients undergoing abdominoplasty and tested for correlation with the subjects' BMI as well as inflammatory mediators.We stimulated human adipose stem cells (ASCs) with recombinant (r)ITIH5 protein and tested for an effect on proliferation, differentiation, and immunosuppressive properties when the cells were exposed to an artificial inflammatory environment.We found positive correlations between ITIH5 levels and the BMI (p < .001) as well as concentrations of inflammatory cytokines (TNF-α, IL-6, and MCP-1) in adipose tissue (p < .01).Application of the rITIH5 protein inhibited both proliferation (p < .001)and differentiation of ASCs.Especially, the development of mature adipocytes was reduced by over 50%.Moreover, rITIH5 decreased the release of IL-6 and MCP-1 when the cells were exposed to TNF-α and IL-1β (p < .001).Our data suggest that ITIH5 is an adipokine that is increasingly released during human adipose tissue development, acting as a regulator that inhibits proliferation and adipogenic differentiation of ASCs.ITIH5 thus presents itself as a positive regulator of adipose tissue homeostasis, possibly protecting against both hyperplasia and hypertrophy of adipose tissue and the associated chronic inflammation.

| INTRODUCTION
In the past, adipose tissue has been considered as energy storage and mechanic buffer, whereas today it is recognized as a highly complex metabolic organ involved in several physiological body functions, including systemic energy homeostasis. 1Furthermore, adipose tissue is also a major endocrine organ, which releases adipose-cytokines known as adipokines. 2 These molecules can act as endocrine mediators, released into the blood stream to affect distant organs, but they also modulate local signaling in a paracrine or autocrine manner.Leptin was the first discovered adipokine and is released from adipocytes proportionally to adipose mass.Under physiological conditions, high levels of leptin decrease appetite and food consumption, which is why it is considered an anorexigenic hormone that counteracts adipose tissue development. 3diponectin is another prominent adipokine, which is expressed conversely to fat mass and promotes adipogenesis. 4,5It has potent anti-inflammatory activity, that is, it inhibits tumor necrosis factor (TNF)-α release of M1macrophages. 6,7Thus, adipokines are promising candidates both for novel pharmacological treatment strategies and as diagnostic or prognostic biomarkers for the future treatment of dysbalanced homeostasis in adipose tissue, e.g., obesity and obesity-related diseases. 8uring vertebrate evolution, inter-αtrypsin inhibitor (ITI) family members are highly conserved molecules, consisting of three protein chains: one ITI light chain, also referred to as bikunin, and two ITI heavy chains (ITIHs). 9n addition to the matrix-protective activity of bikunin, which represents a protease inhibitor, ITIHs interact with extracellular matrix molecules such as hyaluronan (HA) to provide ECM stability and cell-regulatory functions. 10nter-αtrypsin inhibitor heavy chain 5 (ITIH5) was identified as the fifth heavy chain member of the ITI heavy chain family in 2004. 11][14][15][16] Recently, it has been shown that the tumor suppressive function of ITIH5 may be mediated by its protein-familyspecific VIT domain (Interpro number IPR013694), and concepts have been proposed for how to use it for anticancer drug development. 17,18Still, the various potential sites of ITIH5 action in and outside the cell are poorly understood.
Gene expression profiles demonstrated abundant expression of ITIH5 in adipose tissue, and RT-PCR detected high ITIH5 mRNA levels in adipocytes. 19The same study also tested for ITIH5 protein expression in subcutaneous adipose tissue and found higher values in samples from people with obesity than from lean subjects.
However, ITIH5 was found downregulated after weight loss.This suggested a possible role for ITIH5 as a novel adipokine in regulating adipose tissue physiology, but there was no conclusive data.
To elucidate the functional role of ITIH5 in adipose tissue development, metabolism, and maintenance of homeostasis, we now measured native ITIH5 protein concentration in human adipose tissue samples and tested for correlation with the subjects' BMI as well as the concentration of pro-inflammatory mediators, which at increased levels are closely related to the development of obesity, that is, TNF-α, interleukin (IL)-6, and monocyte chemoattractant protein (MCP)-1. 20,21We isolated adipose stem cells (ASCs) and investigated proliferation, metabolic activity, and differentiation after stimulation with recombinant (r) ITIH5 protein.Finally, the anti-inflammatory activity of rITIH5 was tested on ASCs exposed to an artificial inflammatory environment.

| Human recombinant ITIH5
The rITIH5 protein (spanning 681 amino acids) used in this study reflects the secreted N-terminal part of ITIH5, including the VIT domain (Interpro number IPR013694) and the vWA domain (Interpro number SM00327), but without the intact consensus sequence required for covalently binding of hyaluronic acid.pMS-L-A-IV vector cloning of rITIH5, the in vitro synthesis based on HEK293T cells, and its functional, that is, tumor suppressive, activity have been described and confirmed previously. 17Analogously to this procedure, rITIH5-derived proteins present in the supernatant of HEK293T cells were purified by using Ni-NTA agarose beads (Qiagen, Hilden, Germany) and re-buffered in PBS via Vivaspin 6 columns (Sartorius Stedim, Göttingen, Germany) according to the manufacturer's instructions.Western blot analyses were used to confirm the purity and amount of the corresponding rITIH5 protein batches.Recombinant protein was applied at concentrations of 0.01 and 0.1 μg/mL for cell culture experiments based on previous studies, which found a significant impact on tumor cell growth when the cells were exposed to 1 μg/ mL of rITIH5. 17Because preparatory experiments indicated that ASCs were even more sensitive to rITIH5 than tumor cells, all experiments with ASCs were performed at 0.01 and 0.1 μg/mL.

| Tissue samples
Human adipose tissue samples were cut en bloc during abdominoplasty and either frozen for later determination of protein levels and for histological analysis or processed for isolation of adipose stem cells.The healthy donors (n = 25; BMI = 30.2± 4.83; age = 50.3± 10.04 years) had been informed about the utilization of their tissue and had given informed consent.The study protocol was approved by the regional ethics committee (Ethics Committee of the RWTH Aachen University Faculty of Medicine, Aachen, Germany; EK163/07), and experiments were carried out in accordance with the principles of the Declaration of Helsinki.

| Histology
Adipose tissue of n = 3 donors each with low or high BMI was cut into samples of ~5 mm side length, fixed in 4% PFA for 24 h, and then embedded in paraffin.Slices of 2 μm thickness were prepared on a microtome (Hyrax M40, Zeiss, Oberkochen, Germany), mounted on microscope glass slides, and dried overnight in an incubator (37°C).ITIH5 immunohistochemistry was performed with the Labvision Autostainer 360® (Dako, Santa Clara, CA).EnVision®Flex Target Retrieval Solution High pH 9 and the pT-Link pretreatment module (Dako, Santa Clara, CA) were used for deparaffinization, rehydration, and heat-induced epitope antigen retrieval.Sections were washed with EnVision™ Wash Buffer (20×) (Dako, Germany) and then endogenous peroxidase activity was blocked by incubation in EnVision™ FLEX Peroxidase-Blocking Reagent (Dako, Santa Clara, CA) for 5 min.Slices were then incubated with rabbit anti-ITIH5 antibody (1:200; Pineda, Berlin, Germany) for 1 h and washed with EnVision™ Wash Buffer (20×) (Dako, Santa Clara, CA) afterward.The ITIH5-antibody was detected using the horseradish peroxidase-conjugated polymer EnVision™ FLEX/HRP (Dako, Santa Clara, CA) and visualized with the EnVision™ FLEX DAB+Chromogen (Dako, Santa Clara, CA) after an incubation period of 5 min.The slides were washed again with EnVision™ Wash Buffer (20×) and counterstained in hematoxylin.After dehydration in an ascending alcohol series and xylele, the slides were cover-slipped with ROTI®Histokitt (Carl Roth, Germany).Two randomly chosen regions of interest (ROI = 1.6 × 10 5 μm 2 ) in each slide were photographed on an EVOS FL auto-imaging system (Thermo Fisher Scientific, Waltham, USA).We determined the volume of lipid vacuoles as an indicator of fat cell size using the free software ImageJ (Wayne Rasband, Institutes of Health, Bethesda, USA).

| Cell isolation
Procedures for the isolation of adipose stem cells followed the same protocol as described earlier. 22,23Adipose tissue was minced with surgical scissors, washed with PBS, and digested in 0.2% collagenase for 45 min at 37°C.The stromal cell fraction was separated by centrifugation at 400× g and filtered through a 250 nm nylon mesh (neoLab, Heidelberg, Germany).After centrifugation, cells were cultured in proliferation medium (DMEM supplemented with 10% FBS and 0.1% bFGF).Experiments were conducted with cells from P2-P5 seeded at a density of 20 000 per cm 2 , if not mentioned otherwise.

| Proliferation
Proliferation was assessed by measuring cell numbers using crystal violet (CV) staining.The procedure followed the protocol as described earlier. 24Cells were washed with PBS and fixed in isopropyl alcohol for 10 min at RT, followed by washing with PBS containing 0.05% Tween20.Then, cells were stained with 0.1% CV for 20 min.After the removal of the crystal violet solution, the plates were washed with aqua bidest .Adsorbed dye was washed out in 33% acetic acid during a 15 min incubation with gentle agitation.Seventy micro liters per sample were transferred in duplets to an optical plate, and absorbance was quantified in a microplate reader (BMG Labtech, Ortenberg, Germany) at 595 nm.

| Metabolic activity
Metabolic activity was measured by resazurin conversion as described elsewhere, with slight modifications. 25esazurin stock solution was 0.1 mg/mL PBS.After 1 h of incubation (stock:proliferation medium, 1:9) at standard conditions, 100 μL of the medium were transferred into a 96-well plate, and fluorescence was measured in duplets at wavelength of 590 nm (544 nm excitation).

| Differentiation
For differentiation, cells were exposed to various differentiation media 23 : Adipogenic medium was high glucose medium supplemented with 2% FBS, 10 μM insulin, 0.28 μM rosiglitazone, and 0.5 mM IBMX; osteogenic medium was low glucose medium with 2% FBS, 0.8 mM pyruvate, 0.25 μM dexamethasone, 10 mM βglycerophosphate, and 200 μM ascorbate-2-phosphate; chondrogenic differentiation medium was low glucose medium with 0.8 mM pyruvate, 5 ng/mL TGF-β3, 1 μM dexamethasone, 0.17 mM ascorbate-2-phosphate, 350 mM proline, and 50 mg/mL ITS 1 Premix.The media were replaced every 2 to 3 days for 5 days during chondrogenic differentiation and for 14 days during adipogenic and osteogenic differentiation.The evaluation of trilinear differentiation was performed as described earlier. 23Briefly, adipogenic differentiation was assessed by Oil Red O staining.ASCs were PFA fixed, and stained with a 0.2% Oil Red O solution.Adsorbed dye was washed out with isopropanol.Absorbance was measured in triplets at 540 nm.For osteogenesis, cresolphthalein buffer was added to PFA-fixed cells.After adding AMP buffer, the extinction of the supernatant was quantified at 580 nm.Chondrogenic differentiation was measured by Safranin O staining.Differentiated cells were stained with 0.1% Safranin O, and the adsorbed dye was washed out with isopropanol and quantified at 540 nm.

| ITIH5 and cytokine determination by enzyme-linked immunosorbent assay (ELISA)
Frozen tissue specimens (approx. 1 g) were homogenized in 2 mL of lysis buffer (PBS containing a protease inhibitor cocktail; Roche) on ice using a tissue tearer.Homogenates were centrifuged at 2000× g for 10 min to remove large tissue particles and centrifuged at 14 000× g for 30 min.The clear supernatants of each sample were further divided into 1 mL aliquots and stored at −80°C.To measure the amount of native ITIH5 protein and inflammatory cytokines in human adipose tissue, 25 samples were analyzed using the respective ELISA kit (ELISA Duo-Sets, R&D Systems, Minneapolis, Minn; for ITIH5 quantification: BIOZOL, Eching, Germany).Extinction was measured as recommended by the manufacturer using a FLUOstar OPTIMA microplate reader (BMG LABTECH, Aylesbury, United Kingdom) at 450 nm using the reference value of 540 nm.For a proper comparison of the cytokine concentrations, we further evaluated the whole protein content of each sample using the DC Protein Assay (BioRad, Hercules, Calif.) as specified by the manufacturer.The cytokine concentration was then normalized to these results as the content of soluble factor in a picogram per total amount of protein in milligram.
In cell culture experiments, isolated ASCs were stimulated with inflammatory mediators diluted in proliferation medium (1 ng/mL IL-1β and 1 ng/mL TNF-α) and rITIH5 for 48 h.The cell supernatant of each well was analyzed in duplets for IL-6 and MCP-1 release.The concentrations chosen were taken from an earlier study on the inflammatory profile of ASCs. 23

| Analyzes and statistics
The data from all experiments were grouped in order to evaluate the results for each type of experiment and treatment.Associations between ITIH5 protein contents and the BMI of the patients, as well as inflammatory cytokines, were determined based on Pearson correlations.All data were tested for normal distribution using the Kolmogorov-Smirnov test.Normally distributed data were presented as mean values (±SEM) and were statistically analyzed by the Student's t-test for pairwise comparison against the untreated control.Non-normally distributed data were presented as box plots (median as the middle line and 25/75% as box boundaries) and were statistically analyzed using the Mann-Whitney U-test, respectively (SPSS 24, SPSS Inc., Chicago, USA).Differences associated with p ≤ .05were considered statistically significant.Figures were created using the graphics program Corel Draw X5 (Corel Corporation, Ottawa, Canada).

| ITIH5 protein in human adipose tissue
To evaluate the correlation between the native ITIH5 protein content with the BMI of the patients and the concentration of inflammatory cytokines, data were analyzed using Pearson correlation (Figure 1).Linear regression showed a plain ascend for the content of ITIH5 in adipose tissue and the patients' BMI with a significant correlation (r = .7,p < .001; Figure 1A).The same applies for the relation between ITIH5 amount and the content of TNF-α in adipose tissue (r = .5,p = .005;Figure 1B), as well as for IL-6 (r = .7,p = .001;Figure 1C) and MCP-1 (r = .7,p < .001; Figure 1D).
Analysis of histological and immunolabeling indicated that both ITIH5 expression and adipocyte size were elevated with BMI (Figure 2A-C).Morphometric analysis revealed that each individual patient presented a broad distribution of fat cell sizes.Mean adipocyte volume was larger in donors with high BMI (4459 μm 2 SEM ±310 μm 2 ) when compared to counterparts with low BMI (3400 μm 2 SEM ±164 μm 2 ; p = .003;Figure 2D).

| rITIH5 protein reduces ASC viability
Cellular viability was assessed by crystal violet staining to evaluate ASC proliferation, and metabolic activity was investigated by resazurin conversion.After 7 days of standard culture, stimulation with the rITIH5 protein decreased cell number as well as relative metabolic activity per cell (Figure 3A).The effect on crystal violet (CV) staining was stronger when cells were exposed to a higher rITIH5 concentration (0.1 μg/mL).Pairwise comparison by the Mann-Whitney U-test found significant decreases in cell number (p ≤ .001)for 0.01 μg/mL (OD median = 0.69) and 0.1 μg/mL ITIH5 (OD median = 0.62) compared to the control (OD median = 0.79).On the other hand, both rITIH5 concentrations similarly affected ASC metabolic activity (OD median = 1.3) when compared to the untreated control (OD median = 1.4).The rITIH5 protein induced decrease in metabolic activity per cell was significant (p ≤ .05).Comparable results were found after 14 days (Figure 3B).Recombinant ITIH5 protein at 0.01 μg/mL decreased CV staining from OD median = 0.95 to OD median = 0.92, which was significant (p = .035),and 0.1 μg/mL exerted a stronger effect on cell number (OD median = 0.83; p ≤ .001).An rITIH5 concentration-dependent effect was also observed in metabolic activity.The highest resazurin conversion was detected for the control (OD median = 1.8), while it was reduced to OD median = 1.7 by 0.01 μg/mL rITIH5, and 0.1 μg/mL rITIH5 decreased metabolic activity per cell to OD median = 1.6.These differences were significant.

| rITIH5 protein inhibits ASC differentiation
An important characteristic of ASCs is their capacity for trilinear differentiation, meaning their development into adipocytes, osteoblasts, and chondrocytes, in vitro. 26In order to investigate the effect of rITIH5 protein stimulation, ASCs were differentiated into the adipogenic, osteogenic, and chondrogenic lineages (Figure 4).Oil Red O stains the lipid vacuoles of developing and mature adipocytes.After 14 d of exposure to the adipogenic differentiation medium, the rITIH5 protein reduced the Oil Red O signal in a concentration-dependent manner (Figure 4A,D).The strongest effect was found for 0.1 μg/ mL rITIH5 (OD median = 0.1), which inhibited adipogenesis at over 60% when compared to ctrl (OD median = 0.26).Pairwise comparison detected significant differences between the untreated control and both other groups (p ≤ .001).Recombinant ITIH5 protein was not effective in suppressing osteogenic differentiation of ASCs, as calcium deposition was at an equal level among all three groups (OD median = 0.14; Figure 4B).A pairwise comparison calculated no significant differences.During chondrogenic differentiation, the untreated control achieved the highest level in Safranin O staining (OD median = 0.89; Figure 4C), which detects the expression of glycosaminoglycans after 5 days of stimulation with chondrogenic differentiation medium.Recombinant ITIH5 protein decreased the staining values in a concentration-dependent manner (Figure 4E), that is, 0.1 μg/mL produced the strongest effect (OD median = 0.64) and 0.01 μg/mL was intermediate (OD median = 0.69).Pairwise comparison revealed significant differences between ctrl and rITIH5 treatment (p ≤ .001).

| DISCUSSION
Anveden et al. first described abundant ITIH5 protein expression in adipose tissue, which was successfully substantiated in the present study. 19Furthermore, the present work demonstrated that ITIH5 expression correlates with the patient's BMI and the state of tissue inflammation.Stimulation of ASCs with rITIH5 resulted in downregulation of ASC proliferation, differentiation, and also a reduction of secretion of pro-inflammatory cytokines.Surprisingly, these new results suggest that ITIH5 does not contribute to obesity but acts as a negative feedback regulator of ASC biology to maintain adipose tissue homeostasis.
As a major component of the metabolic syndrome in obesity, adipose tissue becomes dysfunctional and expands improperly to store excess energy. 27Therein, fat mass expansion is the result of excessive lipid accumulation, which occurs by volume enlargement of adipocytes (hypertrophy) and/or by increasing adipocyte numbers (hyperplasia).2][33] As we measured the positive correlation between ITIH5 levels in adipose tissue with the patients' BMI, we confirmed that adipose tissue with a high ITIH5 level is characterized by hypertrophic adipocytes.However, it is unlikely that ITIH5 accounted for adipocyte hyperplasia in these tissues because staining of lipid vacuoles in developing adipocytes exposed to rITIH5 in vitro indicated that the protein inhibited lipogenesis.Transferred to mature adipocytes, this activity would not lead to hypertrophy but would have opposite effect.As mature adipocytes are post-mitotic, hyperplasia represents an increase in de novo adipocyte formation, that is, adipogenesis of precursor cells.Adipose tissue contains ASCs and preadipocytes (ASCs after adipogenic commitment) that can differentiate into mature adipocytes. 34hus, dissection of the molecular processes that regulate proliferation and adipogenic differentiation of precursors and identification of mechanisms to modulate adipocyte function are crucial for understanding the maintenance of healthy adipose tissue as well as for developing new therapeutic strategies to prevent and treat this complex disease.
In the past two decades, advances in obesity research have led to the recognition that adipose tissue is an active endocrine organ that secretes more than 600 bioactive factors termed adipokines. 35These soluble factors play a critical role in various biological and physiological functions, for example, regulation of appetite and satiety control, insulin sensitivity, energy expenditure, inflammation, and homeostasis. 36Furthermore, adipokines contribute in an autocrine and paracrine way to the modulation of adipogenesis. 37Especially the proinflammatory adipokines, for example, TNF-α and IL-6, inhibit adipogenesis via interaction with the adipogenesis master regulators C/EBPα and PPAR-γ. 38,39Besides that, the chemoattractant protein chemerin is increasingly expressed in the adipose tissue of mice fed a high-fat diet, while it serves as a negative regulator of adipogenesis as it impairs PPAR-γ expression and adipogenic differentiation. 40WNT1-induciblesignaling pathway protein 2 (WISP2) is an adipokine that is increasingly released in obesity and regulates precursor cell adipogenic commitment and PPAR-γ induction by inhibiting BMP-4. 41,42Gremlin-1 has been identified recently as a novel adipokine that is released by mature adipocytes and markedly increased in the adipose tissue of hypertrophic obesity. 43In knockdown approaches, Gremlin-1 has been shown to actively inhibit adipogenesis in ASCs by antagonizing bone morphogenetic protein (BMP)-2 and -4. 446][47] BMPs are extracellular multifunctional signaling cytokines belonging to the large transforming growth factor (TGF)-β super family. 48,49TIH5 has been reported to block tumor growth and metastasis by modulating the TGF-β signaling cascade via the co-receptor endoglin. 12Furthermore, it has been shown that ITIH5 can stimulate the differentiation of human fibroblasts into myofibroblasts in association with TGF-β and EGFR signaling. 50In more detail, it was found that the TSG-6-driven transfer of ITIH5 to hyaluronic acid was a key for the phenotypic activation of fibroblasts, allowing the formation of CD44-EGFR complexes and CD44 relocation within the cell membrane.Thus, it is possible that ITIH5 triggered downregulation of TGF-β superfamily targets in ASCs, for example, BMP-2 and -4, which consequently inhibited adipogenic differentiation.2][53][54][55] One could speculate that the ITIH5 -TGF-β interaction is strong in ASCs but mainly directed toward soft-tissue development, which could account for the observed inhibitory effects on adipogenic and chondrogenic but not osteogenic differentiation.
While some adipokines have been identified to regulate (adipogenic) differentiation of ASCs, to our best knowledge, there is no study dealing with an effect on ASC cell number.][58] ASC proliferation is primarily directed by various growth  59 Interestingly, members of the TGF-β superfamily also induce stem cell proliferation.TGF-β1 as well as BMP-3 promote proliferation through the βcatenin and activin signaling pathways. 60,61e therefore hypothesize that ITIH5 shifts TGF-β activity so that ASC proliferation is inhibited.
3][64][65][66] This chronic inflammatory response is a key driver of the metabolic syndrome, and it increases proportionally with the expansion of adipose tissue.There are many adipokines with pro-inflammatory activity, for example, leptin, visfatin, resistin, and fatty acid binding protein 4 (FABP4), whereas a smaller number of adipokines are produced that can decrease inflammation and have protective effects on metabolic dysfunction, such as omentin, chemerin, vaspin, and adiponectin. 67However, the anti-inflammatory activity of these molecules is mainly described by downregulation of NF-κB and upregulation of PPAR-γ, thus shifting macrophage phenotypes into alternatively activated M2 cells. 68In a comparative in vitro study, we found that stimulated ASCs secrete higher amounts of IL-6 and -8 than M1 macrophages, which underlines that the ASC secretome might be strongly involved in driving local tissue inflammation and influencing immune cell behavior. 69In the present study, we found that ITIH5 inhibited the release of inflammatory cytokines from ASCs exposed to an artificial inflammatory environment, mimicked by the presence of TNF-α and IL-1β.Thus, in the setting of adipose tissue development, ITIH5 secretion by mature adipocytes could exert regulating effects on metabolic dysfunction by controlling the inflammatory responses of ASCs in a paracrine way.

| CONCLUSION
The expression of ITIH5 correlates significantly with the subject's BMI and the inflammatory state of adipose tissue.By inhibiting both ASC proliferation and adipogenic differentiation, ITIH5 appears to regulate adipose hyperplasia and hypertrophy.Furthermore, by decreasing the release of inflammatory cytokines from ASCs, ITIH5 could represent a relevant player in the maintenance of adipose tissue homeostasis by protecting it from a chronic inflammatory state.Thus, we argue that ITIH5 itself, or the molecular pathway it mediates in human ASCs, may be an interesting new target for counteracting the development of obesity.The foregoing notwithstanding, the downstream pathways that mediate the effects of ITIH5 in ASCs are unclear and require further investigation.

F I G U R E 1
Relationship between the content of native ITIH5 protein in adipose tissue (n = 25) and the donors' BMI (A) as well as the concentration of TNF-α (B), IL-6 (C), and MCP-1 (D) in the tissue samples.The concentration of each factor is indicated as pg/mg of the whole protein amount.The correlation was analyzed using Pearson correlation (r).p-Values ≤.05 were considered statistically significant.

F I G U R E 2
Representative photographs of adipocytes from a donor (A) with a low BMI (=22 kg/m 2 ) and (B) with a high BMI (=36 kg/m 2 ).ITIH5 was detected by immunolabeling using DAB (black arrows).Staining artifacts might have occurred from oil films from truncated adipocytes (asterisks).The volume of adipocytes was determined by measuring the volume of lipid vacuoles.The relationship between cell volume and the donors' BMI is depicted in (C).The cell size of 3 donors with low BMI (n = 247 cells) and high BMI (n = 227 cells) is presented in (D).The pairwise comparison was performed using the Student's t-test, **p ≤ .01.

F I G U R E 4
Effect of recombinant ITIH5 protein exposure on trilinear differentiation of ASCs.Adipogenesis was measured after 14 days by (A) Oil Red O staining.Osteogenesis was analyzed by (B) cresolphthalein staining for quantification of the extracellular matrix calcification after a 14 -day incubation period.After 5 days of stimulation, chondrogenesis was quantified by (C) Safranin O staining.Photographs indicate differences in the number and size of (D) stained lipid droplets in adipocytes, and (E) proteoglycan formation in stained chondrocytes (bar = 50 μm).The number of experiments and donors was as follows: n = 60 (4 donors).Statistical analysis was performed with the Mann-Witney U-test: ***p ≤ .001versus ctrl.| 11 of 13 RUHL et al. factors families, for example, fibroblast growth factors and platelet-derived growth factors.