IL‐6 counteracts the inhibitory effect of IL‐4 on osteogenic differentiation of human adipose stem cells

Abstract Fracture repair is characterized by cytokine production and hypoxia. To better predict cytokine modulation of mesenchymal stem cell (MSC)‐aided bone healing, we investigated whether interleukin 4 (IL‐4), IL‐6, and their combination, affect osteogenic differentiation, vascular endothelial growth factor (VEGF) production, and/or mammalian target of rapamycin complex 1 (mTORC1) activation by MSCs under normoxia or hypoxia. Human adipose stem cells (hASCs) were cultured with IL‐4, IL‐6, or their combination for 3 days under normoxia (20% O 2) or hypoxia (1% O 2), followed by 11 days without cytokines under normoxia or hypoxia. Hypoxia did not alter IL‐4 or IL‐6‐modulated gene or protein expression by hASCs. IL‐4 alone decreased runt‐related transcription factor 2 (RUNX2) and collagen type 1 (COL1) gene expression, alkaline phosphatase (ALP) activity, and VEGF protein production by hASCs under normoxia and hypoxia, and decreased mineralization of hASCs under hypoxia. In contrast, IL‐6 increased mineralization of hASCs under normoxia, and enhanced RUNX2 gene expression under normoxia and hypoxia. Neither IL‐4 nor IL‐6 affected phosphorylation of the mTORC1 effector protein P70S6K. IL‐4 combined with IL‐6 diminished the inhibitory effect of IL‐4 on ALP activity, bone nodule formation, and VEGF production, and decreased RUNX2 and COL1 expression, similar to IL‐4 alone, under normoxia and hypoxia. In conclusion, IL‐4 alone, but not in combination with IL‐6, inhibits osteogenic differentiation and angiogenic stimulation potential of hASCs under normoxia and hypoxia, likely through pathways other than mTORC1. These results indicate that cytokines may differentially affect bone healing and regeneration when applied in isolation or in combination.

the current strategy of choice for the treatment of large bone defects (Ho-Shui-Ling et al., 2018). Mesenchymal stem cells (MSCs) are frequently used for tissue engineering purposes, in particular adipose-derived stem cells (ASCs) represent a promising source due to their abundance, accessibility, and osteogenic differentiation potential (Farre-Guasch et al., 2018;Prins, Schulten, Ten Bruggenkate, Klein-Nulend, & Helder, 2016). However, little is known about the conditions favouring the osteogenic potential of MSCs. Mimicking some aspects of the physiological conditions of bone healing will lead to the development of therapies to aid bone repair in large defects that are currently difficult to treat.
Oxygen tension is low in or near fracture sites (Chung, Won, & Sung, 2009;Lu et al., 2013). Hypoxia has been reported to increase ASC proliferation, stemness marker expression, and chondrogenic differentiation, but to reduce adipogenic and osteogenic differentiation potential (Choi et al., 2014). It has also been reported that hypoxia promotes chondrogenic, osteogenic, and adipogenic differentiation in vitro, and that hypoxic cells show an increased bone repairing capacity in vivo (Tsai et al., 2011).
Bone injury initially causes blood vessel disruption, resulting in a hypoxic environment, followed by hematoma formation and an inflammatory response. During this inflammatory response, different proinflammatory and anti-inflammatory cytokines are released to the injury site (Loi et al., 2016;Mountziaris & Mikos, 2008). For instance, the proinflammatory cytokine interleukin-6 (IL-6) is known to be released in the first 72 hr after bone fracture and to rapidly decline thereafter (Ai-Aql, Alagl, Graves, Gerstenfeld, & Einhorn, 2008;Einhorn, Majeska, Rush, Levine, & Horowitz, 1995). IL-6 is secreted by multiple cell types, such as osteoblasts, and stimulates osteoclast formation and bone resorption, thereby playing a role in bone homeostasis (Kudo et al., 2003;Majumdar, Thiede, Haynesworth, Bruder, & Gerson, 2000). However, the exact role of IL-6 during fracture healing is still unknown, even though IL-6 knockout mouse studies indicate that lack of IL-6 delays bone healing (Yang et al., 2007).
The T helper type 2 cytokine IL-4 is also present during fracture healing (Toben et al., 2011), and is considered antiinflammatory as it inhibits the production of IL-1, tumor necrosis factor α (TNF-α), and prostaglandin E 2 by monocytes (Baumann & Gauldie, 1994). IL-4 also inhibits bone resorption (Watanabe et al., 1990), and is a chemoattractant for osteoblasts (Lind, Deleuran, Yssel, Fink-Eriksen, & Thestrup-Pedersen, 1995). However, the exact role of IL-4 during fracture healing is still unclear. We have shown earlier that treatment with IL-4 or IL-6 during 72 hr exerts opposite effects on osteogenic differentiation of MSCs, that is, IL-6 stimulates, but IL-4 inhibits osteogenic differentiation (Bastidas-Coral et al., 2016). In addition, little is known about the stimulatory or inhibitory effects of the combination of proinflammatory and anti-inflammatory cytokines in a hypoxic environment, as occurs during bone healing, on the osteogenic differentiation potential of MSCs. This represents a significant hiatus in our knowledge, considering that (a) cytokines as well as hypoxia are hallmarks of the early stages of fracture healing, (b) both cytokines and hypoxia will likely be present in any implanted tissue-engineered construct in vivo, and (c) cytokines and hypoxia possibly interact at the level of signal transduction. A previous study from our group has also demonstrated the importance of studying the effect of different cytokines in combination, showing that the combination of cytokines present in the circulation of patients with active rheumatoid arthritis might contribute to generalized bone loss by directly inhibiting osteoblast proliferation and differentiation (Pathak et al., 2014).
Under hypoxia mammalian target of rapamycin (mTOR) is inactivated, which may occur as part of the cell program to maintain energy homeostasis (Knaup et al., 2009). IL-6 is known to signal via GP130/IL-6R, activating the JAK/STAT pathway, which activates the mTOR pathway via insulin-like growth factor 1. The mammalian target of rapamycin complex 1 (mTORC1) signaling pathway is required for osteoblast proliferation and differentiation (Singha et al., 2008), and it plays an important role in the regulation of bone metabolism and skeletal development by regulating messenger RNA (mRNA) translation during preosteoblast differentiation (Bakker & Jaspers, 2015;Fitter et al., 2017).
The activation of IL-4 signaling in bone marrow MSCs (BMMSCs) favors adipogenic differentiation and prevents osteoblast differentiation in an mTORC1-dependent manner in a mutant "tight skin" mouse model of systemic sclerosis (Chen et al., 2015). Thus, hypoxia, IL-4, and IL-6 pathways may interact at the level of mTORC1. Interestingly, both the osteogenic effect of IL-6 on human BMSCs and the antiosteogenic effect of IL-4 on BMMSCs have been ascribed to mTORC1 activation (Chen et al., 2015;Deshpande et al., 2013).
The aim of this study was to determine whether IL-4, IL-6, or the combination of both cytokines affects osteogenic differentiation and the angiogenic stimulation potential of MSCs under normoxia and hypoxia. We hypothesized that IL-4 decreases, while IL-6 enhances osteogenic differentiation and vascular endothelial growth factor (VEGF) production, and that the combination of both cytokines will not enhance nor decrease osteogenic differentiation and VEGF production in MSCs, as the effect of IL-4 and IL-6 will counterbalance. The osteogenic and angiogenic effects exerted by IL-4 and/or IL-6 will be accompanied by mTORC1 activation in MSCs.

| MATERIALS AND METHODS
2.1 | Adipose tissue donors 2.2 | Isolation and culture of human adipose stem cells (hASCs) Isolation, characterization, and osteogenic differentiation capacity of hASCs has been previously reported by our group (Varma et al., 2007). For the isolation of hASCs, adipose tissue was cut into small pieces and enzymatically digested with 0.1% collage-

| Platelet lysate
Pooled platelet products from five donors were obtained from the Bloodbank Sanquin (Sanquin, Amsterdam, The Netherlands) and contained approximately 1 × 10 9 platelets/ml (Prins et al., 2009). PL was obtained by lysing the platelets through temperature shock at −80°C. Before use, PL was thawed and centrifuged at 600g for 10 min to eliminate remaining platelet fragments. The supernatant was added at 2% (v/v) to the medium.

| hASCs culture under hypoxia
HACSs were cultured under hypoxia inside a custom designed hypoxic workstation (Top Class Products and Services, Rotselaar, Belgium), where oxygen concentration was controlled via injection of N 2 as described (Nauta, Duyndam, Weijers, van Hinsbergh, & Koolwijk, 2016). Oxygen concentration inside the incubator was continuously monitored with an internal zirconia sensor, as well as by periodically external calibration with O 2 test tubes (Drager Safety, Zoetermeer, The Netherlands). To maintain the hypoxic condition of the hASCs, medium was preincubated for 3 hr under hypoxia before use. Hypoxia was defined as 1% O 2 , 5% CO 2 , and 94% N 2 .
2.6 | DNA content HASC cultured for 2 and 7 days with IL-4, IL-6, or both (IL-4 + IL-6), were washed with PBS, and lysis buffer was added. DNA content as a measure for cell number was determined using a Cyquant Cell Proliferation Assay Kit (Molecular Probes, Leiden, The Netherlands).
Absorption was read at 485 nm excitation and 528 nm emission in a microplate reader (Synergy™ HT spectrophotometer; BioTek Instruments Inc, Highland Park, Winooski, VT).

| RNA isolation and real-time reverse transcription polymerase chain reaction (RT-PCR)
RNA isolation from hASCs was performed using the RNeasy Mini Kit  (Pfaffl, Tichopad, Prgomet, & Neuvians, 2004), values were normalized to TATA-box binding protein and β-glucuronidase housekeeping genes. Real-time PCR was used to assess gene expression of KI67, runt-related transcription factor 2 (RUNX2), collagen type 1 (COL1), osteocalcin, and VEGF-165. All primers used were from Life Technologies. The primer sequences are listed in Table 1. mRNA preparations of hASCs were used as a reference and internal control in each assay.

| Alkaline phosphatase (ALP) activity
hASC cultured for 2, 4, and 7 days with IL-4 and/or IL-6 in 1% or 20% O 2 were lysed with 250 µl milli-Q water, and stored at −20°C until use. 4-Nitrophenyl phosphate disodium salt (Merck) at pH 10.3 was used as a substrate for ALP, according to the method described by Lowry (Lowry, 1955). The absorbance was read at 405 nm with a Synergy HT spectrophotometer. ALP activity was expressed as µM/ ng DNA.

| Mineralization
Matrix mineralization was analyzed by alizarin red staining after incubation of hASCs with IL-4 and/or IL-6 in 1% or 20% O 2 at Day 14 by using 2% Alizarin Red S (Sigma-Aldrich) in water at pH 4.3, as described (Fukuyo et al., 2014). Briefly, hASCs were fixed with 4% formaldehyde for 15 min and rinsed with deionized water before adding 350 µl of the alizarin red solution per well. After incubation at room temperature for 30 min, the cells were washed with deionized water. Cells that have differentiated into osteoblasts deposit mineralized matrix, which is visible as bright red nodules.
Quantification of the mineralized matrix was performed using ImageJ software 1.49 v (Wayne Rasband, National Institutes of Health, Bethesda, MD) as previously described (Shah et al., 2016).

| VEGF quantification
VEGF protein concentration was measured in the supernatant of hASCs after incubation with IL-4 and/or IL-6 in 1% or 20% O 2, at Day 7, by using a Quantikine ® Elisa kit (R&D Systems) according to the manufacturer's protocol. The absorbance was read at 450 nm with a microplate reader (Synergy HT spectrophotometer).

| Western blot
hASCs were lysed in Ripa buffer to quantify total protein concentration with a bicinchoninic acid protein assay (Pierce, Rockford, IL).

| Statistical analysis
Values are presented as mean ± standard deviation (SD). In total three independent experiments were performed in duplicate (n T A B L E 1 List of primer sequences used for analysis of proliferation, osteogenic and angiogenic markers by hASCs by PCR
= 3) using IL-4 or IL-6 (1 and 10 ng/ml), and in triplicate (n = 3) using IL-4 in combination with IL-6 (10 ng/ml). Statistical significance was determined using analysis of variance, with application of Dunnett's multiple comparison test to compare IL-4 or IL-6 at 1 and 10 ng/ml with controls. Two-tailed paired t-test was used to compare control groups without cytokines under normoxia and hypoxia, and to compare IL-4 in combination with IL-6 with the control. A p < 0.05 was considered significant.
Statistical analysis was performed using GraphPad Prism 5.4 (GraphPad Software, San Diego, CA).

| hASC proliferation
The  (Lee et al., 2015). These findings suggest that an hypoxic environment might alter the timing of sequential gene expression of early osteogenic differentiation markers in relation to late osteogenic markers during the osteogenic differentiation process (Lee et al., 2015). This is in partial accordance with our results showing that hypoxia decreased ALP activity in hASCs cultured without cytokines by approximately 3.0-fold at Day 7 compared with normoxia (Figures 1h and 3h), but not mineral deposition by hASCs. Expression of the cytokines IL-4 and IL-6 is significantly elevated in the fracture callus (Einhorn et al., 1995). A recombination activating gene 1 knockout (RAG1 −/− ) mouse lacking the adaptive immune system, thereby having reduced IL-4 levels, does not show impaired fracture healing (Toben et al., 2011), suggesting that the presence of IL-4 is not essential for modulating MSC behavior. In contrast, another study has reported that BMMSCs from FBN1-deficient (Fbn1 +/− ) mice exhibit decreased osteogenic differentiation, and that this lineage alteration is regulated by IL4/IL4Rα-mediated activation of mTOR signaling to downregulate RUNX2 (Chen et al., 2015). This is in partial accordance with our results showing that IL-4 strongly reduced RUNX2 and In contrast to the effects of IL-4, IL-6 −/− mice show delayed callus mineralization and remodeling compared with wild-type mice, 2 weeks postfracture, indicating that IL-6 signaling plays an important role in the early stages of fracture healing (Yang et al., 2007).
Whether this is caused by a direct regulatory effect of IL-6 on MSCs, or any other cell type, or whether this can be explained by IL-6modulated effects of other signaling molecules remains to be elucidated. We showed that IL-6 enhanced mineralization of hASCs under normoxia, which is in agreement with our previous study where IL-6 also induced mineralization of hASCs under normoxic culture conditions (Bastidas-Coral et al., 2016). Therefore, IL-6 may play an important role in the osteogenic differentiation of MSCs.
During the inflammatory phase of fracture healing, cytokines are expressed simultaneously, which may cause synergistic or antagonistic effects in a hypoxic environment. Therefore, we evaluated the combined effect of IL-4 and IL-6 on proliferation, osteogenic differentiation, and angiogenic potential of hASCs. We found that IL-4 with IL-6 inhibited expression of the early osteogenic markers RUNX2 and COL1, which is similar to the effect of IL-4 alone. In addition, IL-4 combined with IL-6 enhanced mineralization of hASCs under normoxia, which is similar to the effect of IL-6 alone.
Remarkably, ALP activity and mineralization in hASCs were increased only by the combination of IL-4 with IL-6 under normoxia and hypoxia, and the inhibitory effect of IL-4 was counteracted. A study mimicking the endogenous microenvironment of muscle stem cells (MuSCs), showed that only a combination of four proinflammatory cytokines (IL-1α, IL-13, TNF-α, and interferon-γ) was able to stimulate MuSC proliferation in vivo upon muscle injury and to promote serial expansion of MuSCs in vitro (Fu et al., 2015). These findings suggest that a crosstalk between cytokine signaling and their downstream signaling pathways may occur upon tissue injury in vivo under hypoxia.
In conclusion, IL-4 reduced osteogenic differentiation of hASCs, indicating that this cytokine might inhibit bone healing and regeneration. However, the effects of IL-4 were mitigated in the presence of IL-6 ( Figure 8). This shows that for a better understanding of bone healing, for example for tissue engineering purposes, it is important to move towards more complex in vitro systems, taking into account factors such as oxygen tension and combinations of cytokines.

ACKNOWLEDGEMENTS
The authors thank Mr. Netherlands.

CONFLICT OF INTERESTS
The authors declare that there are no conflict of interests.