D‐dopachrome tautomerase in adipose tissue inflammation and wound repair

Abstract D‐dopachrome tautomerase (D‐DT/MIF‐2) is a member of the macrophage migration inhibitory factor (MIF) cytokine superfamily, and a close structural homolog of MIF. MIF and D‐DT have been reported to be involved in obesity, but there is little known about the regulation of D‐DT in adipose tissue inflammation and wound healing. Subcutaneous adipose tissue was collected from 54 healthy donors and 28 donors with acutely inflamed wounds undergoing wound debridement. In addition, epididymal fat pads of mice were injected with lipopolysaccharide to study receptor expression and cell migration in vivo. D‐DT protein levels and mRNA expression were significantly decreased in subcutaneous adipose tissue adjacent to acutely inflamed wounds. D‐DT improved fibroblast viability and increased proliferation in vitro. While D‐DT alone did not have a significant effect on in vitro fibroblast wound healing, simultaneous addition of neutralizing MIF antibody resulted in a significant improvement of fibroblast wound healing. Interestingly, expression of the MIF and D‐DT receptor CD74 was down‐regulated while the MIF receptors CXCR2 and CXCR4 were up‐regulated primarily on macrophages indicating that the MIF‐CXCR2/4 axis may promote recruitment of inflammatory cells into adipose tissue. Our results describe a reciprocal role of D‐DT to MIF in inflamed adipose tissue, and indicate that D‐DT may be beneficial in wound repair by improving fibroblast survival and proliferation.


Introduction
Wound repair is a complex response by the host to tissue damage, and wound healing disorders contribute to significant morbidity and excess health care expenditures [1]. Acute wound healing disorders occur at areas of adipose tissue inflammation [2,3], which can lead to life-threatening complications and extended soft tissue defects requiring complex reconstruction procedures [4]. In the past several years, adipose tissue has been determined to be an important endocrine organ that orchestrates diverse responses by the release of specialized cytokines in an autocrine and paracrine manner [5]. Subcutaneous adipose tissue contributes to cutaneous wound healing through the secretion of adipokines and the differentiation of progenitor cells into keratinocytes and fibroblasts [6][7][8][9][10].
Obesity has been implicated in a number of autoimmune diseases and the increasing number of obese patients with impaired wound healing has generated significant interest in the function of adipokines during wound repair [9]. One adipokine that has been recently characterized in obesity and insulin resistance is D-dopachrome tautomerase (D-DT) [11]. D-DT was first described by Odh and colleagues as an enzyme in human melanoma, human liver and several rat organs which tautomerizes D-dopachrome, but its precise role in body physiology remained unknown for many years [12]. The description of structural and functional similarities between D-DT and the innate cytokine macrophage migration inhibitory factor (MIF) recently advanced our understanding of D-DT significantly [13,14] and led to the alternative designation of D-DT as MIF-2 [15].
We previously reported increased MIF levels in subcutaneous adipose tissue in participants with acute wound healing disorders and proposed that MIF may mediate the recruitment of inflammatory cells [4]. While MIF and D-DT share functional similarities, expression data suggest that during chronic adipose tissue inflammation and obesity, the two cytokines have opposing or complementary roles [11]. Migration inhibitory factor promotes chronic adipose tissue inflammation and its expression positively correlates with obesity and insulin resistance, whereas D-DT expression shows a negative correlation [11,16]. The goal of this study was to elucidate more precisely the function of D-DT in subcutaneous adipose tissue adjacent to acute wound healing disorders.

Adipose tissue samples
Healthy adipose tissue (HAT) was collected from healthy patients undergoing elective plastic surgery, whereas inflammatory adipose tissue (IAT) was harvested from acute wound healing disorders. Those were defined as wounds caused by external trauma, iatrogenic manipulation and surgical intervention, which were not resolved within a period of 4 weeks after trauma [2,3]. All cases showed classic signs of local inflammation (tumour, calor, rubor, dolour and functio laesa) and had negative bacterial swab samples at the time of tissue harvest. The adipose tissue was excised and sample preparation was carried out as previously described [4]. All surgeries were performed in the Department of Plastic Surgery, Hand Surgery -Burn Center of the RWTH University Hospital Aachen, Germany. The study was approved by the ethics committee of the RWTH Aachen University (EK 163/07), with written informed consent provided by patients or relatives. All experiments were performed in compliance with the Declaration of Helsinki Principles.

Tissue homogenization and detection of total protein
Whole adipose tissue samples were homogenized as described earlier [17]. Total protein content of whole adipose tissue samples was measured by the DC Protein Assay (Bio-Rad Laboratories GmBH, Munich, Germany) according to the manufacturer's instructions.
Human and murine D-DT Enzyme linked immunosorbent assay (ELISA) D-dopachrome tautomerase protein levels of human adipose tissue homogenates and murine epididymal fat pads were measured by ELISA as previously reported [13].

Real-time PCR
Messenger RNA from human adipose tissue and murine epididymal fat pads was extracted by the RNeasy Mini Kit (Qiagen NV, Venlo, The Netherlands) and cDNA synthesis was performed with the QuantiTect Reverse Transcription Kit (Qiagen Inc., Valencia, CA, USA) following the manufacturer's instructions. Complementary DNA was subject to quantitative real-time (RT) PCR using the SYBR â Green Master Mix (Bio-Rad Laboratories, Inc., Hercules, CA, USA). Primers used in this study are listed in Table S1. For human samples, GAPDH was used as an internal control, while in mice b-actin served as a reference. Data were analysed with the comparative cycle time (CT) method.

CXCR2, CXCR4 and CD74 staining by flow cytometry
Adipocytes and adipose tissue macrophages (ATM) were isolated from lipopolysaccharide (LPS) or PBS-injected mice for flow cytometry by collagenase digestion as previously described [18,19]. Used antibodies are found in Table S2. ATMs were stained for CD45, CD11b and F4/80. Adipocytes were stained for CR4. Fluorescence minus one controls were included for both cell types. Samples were measured on a LSR flow cytometer (BD Bioscience, San Jose, CA, USA) and the data were analysed using FlowJo (FlowJo LLC, Ashland, OR, USA).

Isolation of human dermal fibroblasts
Human dermal fibroblasts (HDF) were isolated from healthy patients as described earlier [20] and cultured at 37°C in a humidified atmosphere at 5% CO 2 .

Measurement of cell viability by the alamarBlue â assay
Viability of HDF was measured by the alamarBlue â assay (Invitrogen, Thermo Fisher Scientific, Darmstadt, Germany) according to the manufacturer's guidelines. Human dermal fibroblasts were seeded in 96 well plates. The following day, alamarBlue â reagent was added to untreated HDF, this time-point is referred to as the baseline viability. All wells were washed and incubated with supernatants from HAT and IAT, supplemented with varying concentrations of recombinant human D-DT. After 24 hrs, another alamarBlue â measurement was performed.

Measurement of cell proliferation by the bromodeoxyuridine ELISA
Proliferation of HDF was analysed by the CytoSelect TM bromodeoxyuridine (BrdU) Cell Proliferation Elisa Kit (Cell Biolabs Inc., San Diego, CA, USA). Human dermal fibroblasts were seeded in 96 well plates. The following day, cells incubated with supernatants from HAT and IAT, supplemented with varying concentrations of recombinant human D-DT for 24 hrs. BrdU incorporation was measured according to the manufacturer's instructions.

In vitro wound healing assay
In vitro wound healing was performed by a standard in vitro wound healing assay [21]. Briefly, HDF monolayers were scratched by a 200 ll pipette tip. Medium containing 10% FCS served as a positive control, while medium containing 0.5% FCS served as a negative control. Human dermal fibroblasts were incubated with supernatants for 16 hrs, which were supplemented with different concentrations of recombinant human D-DT. 10 lg/ml Mitomycin C (Sigma Aldrich Chemie GmbH, Taufkirchen, Germany) was utilized to examine proliferation and migration-related wound healing. The MIF antibody III.D.9 was used to show MIF-dependent effects as reported earlier [22]. Photographs were acquired at 0 and 16 hrs after the scratch. The migration (addition of Mitomycin C) and migration/proliferation distance (absence of Mitomycin C) were deduced by comparing the 0 and 16 hrs photographs by the Software Photoshop CS5 (Adobe Systems, Mountain View, CA, USA).

In vivo macrophage migration model
In vivo macrophage migration into epididymal fat pads of mice was performed as published earlier [4,23]. Ten-week-old male C57BL/6 wildtype (WT) mice purchased from Charles River Laboratory (Wilmington, DE, USA) were used. All mice were fed with standard chow and water Technologies, Carlsbad, CA, USA). Next, labelled PMs were injected retro-orbitally. Mice were killed after 48 hrs and epididymal fat tissue was harvested, and the stromal vascular fraction (SVF) was isolated. The isolated ATM were stained with Cd11b AlexaFluor700 (eBioscience, San Diego, CA, USA), F4/80 eFluor450 (eBioscience) and Cd45 APC (Biolegend, San Diego, CA, USA) as reported earlier [4].

Statistical analysis
For all statistical analysis and diagrams, the software GraphPad Prism â (GraphPad Software, Inc., La Jolla, CA, USA) was used. Data were expressed as mean AE S.E.M. After testing for normal distribution using the Shapiro-Wilk-W-test, statistical differences between groups were calculated by one-way ANOVA in case of three or more variables and t-test in case of two variables. Post hoc testing was performed using the Bonferroni-test. A P-value of <0.05 was considered as significant.

Decreased D-DT protein levels and mRNA expression in adipose tissue from wound healing disorders
To study the role of D-DT in the context of wound healing, D-DT protein levels from HAT and IAT were measured by ELISA. Inflammatory adipose tissue showed significantly reduced D-DT levels when compared to HAT (Fig. 1A). Simultaneously, D-DT mRNA expression was analyzed by quantitative real-time PCR in order to determine if attenuated D-DT protein levels were related to a down-regulation of gene expression. Similar to the D-DT protein levels, mRNA expression of D-DT was decreased in IAT when compared with HAT (Fig. 1B). Immunohistochemical staining also indicated a lower expression of D-DT protein in mature adipocytes in IAT (Fig. 1C). Positive and negative controls can be found in Figure S1.

D-DT levels do not correlate with the donor's body mass index
Given the reported positive correlation of MIF protein levels and the BMI of HAT donor's [4] and elevated MIF plasma levels in obese patients [16], we also examined the correlation of D-DT levels with the BMI of HAT donor patients and observed that unlike MIF [4], D-DT did not correlate with the BMI of the donor in HAT (Spearman r = 0.1664; Fig. S2).

D-DT increases the viability and proliferation of HDF
To evaluate the role of D-DT in wound healing, we measured the effect of D-DT on the viability and proliferation of HDFs. Human dermal fibroblasts incubated with supernatants from IAT showed significantly decreased cell viability compared to HDF treated with supernatants from HAT. However, once HDF were cultured with IAT supernatants and treated with D-DT, the cell viability increased significantly in a dose-dependent manner ( Fig. 2A). We observed that D-DT also induced HDF proliferation in a dose-dependent manner (Fig. 2B).
Increased wound healing in in vitro scratch assay of HDF monolayers by D-DT and anti-MIF We further used a standardized in vitro scratch assay that evaluates the proliferation and migration of HDF from the wound edges. Mitomycin C was optionally added to eliminate the proliferation of HDF and solely focus on the migration. Within the range of 50-200 ng/ml recombinant D-DT, a trend towards increased wound healing was detected, but a statistically significant dose-dependent effect was not observed (Fig. 3A). When mitomycin C was added, the addition of D-DT did not lead to an improved healing of HDF monolayers (Fig. 3C). However, when neutralizing anti-MIF antibodies and recombinant  D-DT were added simultaneously, the proliferation of HDF increased significantly (Fig. 3B). These results suggest that D-DT may have an opposing effect on wound repair which becomes apparent when MIF is blocked, and that the process is primarily based on HDF proliferation rather than migration (Fig. 3D).

Differential expression of MIF and D-DT receptors in human and murine adipose tissue
MIF and D-DT both bind and signal through the CD74/CD44 receptor complex [15]. The chemokine receptors CXCR2 and CXCR4 also have been described as receptors for MIF [24]. The mRNA expression of CXCR2 and CXCR4 was up-regulated in human IAT (Fig. 4A and B) and in mice a similar up-regulation in gene expression after LPS injection was observed for Cxcr2 in epididymal fat pads (Fig. 5A). However, the mRNA levels for Cxcr4 were not altered (Fig. 5B) in mouse adipose tissue. Interestingly, expression of CD74 was down-regulated in both human IAT (Fig. 4C) as well as in mouse adipose tissue (Fig. 5C). Immunohistochemical staining of CXCR2, CXCR4 and CD74 confirmed the results of the qPCR analysis and indicated an upregulation of CXCR2 and CXCR4 in IAT, whereas CD74 was downregulated ( Fig. 4D-I). Finally, flow cytometry analysis of murine adipose tissue revealed that the increase in CXCR2 and CXCR4, and the decrease in CD74 primarily occurred in macrophages, whereas receptor expression in adipocytes did not differ (Fig. 5D-I).

Decreased D-DT levels in mouse adipose tissue after LPS stimulation
Mouse D-DT levels were measured in the epididymal fat pads one day after injection of LPS into epididymal fat pads. D-DT levels decreased in the LPS challenged group when compared to the PBS-injected control group (Fig. 5A), recapitulating our observations in human HAT and IAT.

D-DT exhibits decreased macrophage recruitment into mouse adipose tissue after LPS injection
To further elucidate the functional role of D-DT in IAT, we performed an in vivo cell tracking experiment. Labelled PMs were injected retro-orbitally into mice after induction of a local inflammatory response in epididiymal fat pads by a combination injection of LPS/MIF/D-DT. When compared to the control group (PBS injection), adipose tissue from LPS treated mice showed increased PM invasion. Injection of LPS and MIF led to an increased PM accumulation in the epididymal fat pads that was significantly higher than in the LPS treated group. An increase in PM content was also observed when D-DT was co-injected with LPS; however, this effect was lower compared to MIF and did not reach statistical significance (Fig. 6B).

Discussion
In this study, we observed a decrease in D-DT expression at both the protein and mRNA levels in IAT when compared to HAT. The first study to describe D-DT as an adipokine was provided by Iwata et al., who showed that D-DT was mainly expressed in adipocytes and not SVF cells [11]. Furthermore, the study showed that D-DT mRNA was expressed during the course of adipogenic differentiation only in differentiated adipocytes. Interestingly, in this study the mRNA expression of D-DT in adipocytes correlated negatively with the donor's BMI. Our study could not confirm a negative correlation of D-DT levels with donor's BMI; however, this may be attributed to the fact that our study focused on the D-DT levels in entire subcutaneous adipose tissue rather than only adipocyte-specific D-DT. These observations stand in contrast to MIF, which shows a positive correlation with BMI and obesity, and has been shown to promote adipose tissue inflammation [16,[25][26][27]. In our study, immunohistochemical staining revealed that in healthy tissue D-DT expression was overall diminished in the adipocytes of IAT when compared to controls. A recent study showed that MIF is expressed in both adipocytes as well as SVF cells and that within inflamed adipose tissue MIF mRNA expression and protein levels were significantly increased [4]. These findings point to potentially opposite functions of D-DT and MIF in adipose tissue and wound repair.
Many studies have described wound healing to be impaired in obese patients [9]. Favourable effects improving wound repair involve keratinocyte and fibroblast migration, proliferation, differentiation as well as keratinocyte and fibroblast interactions [28]. We showed that supernatants from IAT had a harmful effect on the viability of HDF. But when co-stimulated with D-DT, the harmful effect of IAT supernatants was partially reversed and the cell viability and proliferation increased in a D-DT dose-dependent manner. These observations indicate that soluble factors from IAT have detrimental effects on HDF cell viability and proliferation and that the reduction of D-DT in IAT may play a role. D-DT alone did not show significant effects on HDF in the scratch assay. The simultaneous addition of MIF antibody, which was shown to have beneficial effects on wound healing by impacting on fibroblast cell migration [4], and recombinant D-DT, however, resulted in a significant increase in HDF proliferation further suggesting that MIF and D-DT may have opposing effects to fibroblast wound repair. Although the exact mechanisms by which D-DT exerts its effects are unknown, the activation of kinases such as the extracellularsignal regulated kinases 1/2 may be a possibility [13].
Tissue resident macrophages can either arise from hematopoietic or from embryonic progenitor cells and they fulfill tissue-specific and niche-specific functions [29]. Previous studies report that during obesity most adipose tissue resident macrophages originate from infiltrating blood monocytes and some studies have suggested that the migration and extension of macrophages in adipose tissue leads to increased local inflammation, which in turn could impact wound healing processes [30][31][32]. MIF and D-DT have been described to inhibit the chemotaxis of peripheral monocytes [13]. When LPS was injected into epididymal fat pads, D-DT protein levels were decreased, similar to decreased D-DT levels in IAT. When murine epididymal fat pads were challenged with LPS plus MIF or D-DT, only LPS plus MIF led to a significant increase in invaded PM when compared to the LPS treated group. Our data suggest that MIF may be the more dominant factor in macrophage recruitment than D-DT.   Presently, little is known about the expression of MIF/D-DT receptors in adipose tissue and their association with adipose tissue inflammation. Both MIF as well as D-DT bind to the cell surface receptor CD74, but only MIF has been described to signal through the noncognate receptors CXCR2 and CXCR4 [15]. Our study showed an increased expression of CXCR2 and CXCR4 in IAT, elevated levels of CXCR2 in challenged mouse adipose tissue, while the expression of CD74 was down-regulated in adipose tissue inflammation in both humans and mice. Both CXCR2-mediated signalling, as well as CXCR4-mediated signalling have been implicated with adipose macrophage and monocyte recruitment and the CD74/CD44 complex has been shown to promote cell survival and pro-inflammatory responses [15,24,33]. MIF expression was shown to be increased in human IAT, and promoted macrophage infiltration to site of inflammation [4], correlating with our data of increased CXCR2/CXCR4 expression in human tissue. By flow cytometry we also showed that the up-regulation of CXCR2 and CXCR4 is mainly found on macrophages, whereas receptor expression on adipocytes was unaltered. Recent studies have described the involvement of macrophage recruitment in acutely inflamed adipose tissue, which occurred in combination with decreased expression of intercellular adhesion molecule-1 and CD44, two relevant adhesion molecules inducing monocyte infiltration [4,27]. As CD74 needs CD44 for its beneficial functions (e.g. cell proliferation), the observed decrease of CD44 expression is an indication that CD74 also is not involved in this context; this correlates well with the decrease in CD74 expression found in our study. Together with the in vivo experiments, our results suggest that macrophage recruitment into IAT is primarily promoted by MIF and CXCR2/4 rather than through the D-DT-CD74 axis.
We have to acknowledge some limitations of the present work which is an initial study to examine a role for D-DT during acute adipose tissue inflammation and wound repair. First, the studies examining the effect of D-DT on HDF were performed in an in vitro setting in an artificial environment and have to be interpreted with caution. While the tissue expression of D-DT is low, high concentrations of recombinant D-DT were necessary to see effects in the in vitro viability, proliferation and wound healing assays. Definite conclusions cannot be made until additional in vivo studies are performed. Second, the LPS injection model used in our in vivo experiment is a simplified model that does not mimic the more complex wound healing disorders of the human samples. We also acknowledge that LPS was injected into epididymal fat pads and not into subcutaneous adipose tissue which may be a more appropriate tissue type.
Clearly, it would be interesting to further investigate the specific function of D-DT in wound repair in more extended studies, for example, using an in vivo wound model, preferably including an adipose tissue conditional knockout of D-dt. Furthermore, in our study we only have focused on whole adipose tissue without differentiation of the respective cell fractions. Therefore, it may be interesting to investigate changes in MIF, D-DT, and receptor expression after separation into adipose tissue (e.g. after collagenase digestion) in future studies.
In conclusion, our study indicates that D-DT may play a distinct role from MIF in adipose tissue inflammation during wound healing disorders. In contrast to MIF, D-DT protein and mRNA levels were decreased in IAT, which may be detrimental for wound healing since supplementation with recombinant D-DT led to significantly increased HDF viability and proliferation. Furthermore, our data suggest that in contrast to D-DT, MIF is the more potent chemokine that facilitates the recruitment of macrophages to the inflammatory site by its receptors CXCR2/4. A schematic illustration of our findings and the reciprocal roles of MIF and D-DT in adipose tissue is demonstrated in Figure 7. However, additional in vivo studies are required especially to draw a final conclusion regarding the effect of D-DT on fibroblasts as the in vitro studies show several major limitations.

Conflict of interest
The authors confirm that there are no conflicts of interest.

Supporting information
Additional Supporting Information may be found online in the supporting information tab for this article: