bioengineering Article Collagen Fibril Density Modulates Macrophage Activation and Cellular Functions during Tissue Repair Jiranuwat Sapudom 1,† , Walaa Kamal E. Mohamed 1,2,† , Anna Garcia-Sabaté 1 , Aseel Alatoom 1 , Shaza Karaman 1 , Nikhil Mahtani 1 and Jeremy C. M. Teo 1,3, * 1 Division of Engineering, New York University Abu Dhabi, Abu Dhabi 129188, UAE; jiranuwat.sapudom@nyu.edu (J.S.); wm1081@nyu.edu (W.K.E.M.); anna.sabate@nyu.edu (A.G.-S.); aseel.alatoom@nyu.edu (A.A.); sk6349@nyu.edu (S.K.); nm3489@nyu.edu (N.M.) 2 Department of Genetics and Microbiology, Autonomous University of Barcelona, 08193 Barcelona, Spain 3 Department of Mechanical and Biomedical Engineering, New York University, New York, NY 10003, USA * Correspondence: jeremy.teo@nyu.edu; Tel.: +971-2-628-6689 † Contributed equally to this work.  Received: 4 March 2020; Accepted: 30 March 2020; Published: 31 March 2020  Abstract: Monocytes circulate in the bloodstream, extravasate into the tissue and differentiate into specific macrophage phenotypes to fulfill the immunological needs of tissues. During the tissue repair process, tissue density transits from loose to dense tissue. However, little is known on how changes in tissue density affects macrophage activation and their cellular functions. In this work, monocytic cell line THP-1 cells were embedded in three-dimensional (3D) collagen matrices with different fibril density and were then differentiated into uncommitted macrophages (MPMA ) using phorbol-12-myristate-13-acetate (PMA). MPMA macrophages were subsequently activated into pro-inflammatory macrophages (MLPS/IFNγ ) and anti-inflammatory macrophages (MIL-4/IL-13 ) using lipopolysaccharide and interferon-gamma (IFNγ), and interleukin 4 (IL-4) and IL-13, respectively. Although analysis of cell surface markers, on both gene and protein levels, was inconclusive, cytokine secretion profiles, however, demonstrated differences in macrophage phenotype. In the presence of differentiation activators, MLPS/IFNγ secreted high amounts of IL-1β and tumor necrosis factor alpha (TNFα), while M0PMA secreted similar cytokines to MIL-4/IL-13 , but low IL-8. After removing the activators and further culture for 3 days in fresh cell culture media, the secretion of IL-6 was found in high concentrations by MIL-4/IL-13 , followed by MLPS/IFNγ and MPMA . Interestingly, the secretion of cytokines is enhanced with an increase of fibril density. Through the investigation of macrophage-associated functions during tissue repair, we demonstrated that M1LPS/IFNγ has the potential to enhance monocyte infiltration into tissue, while MIL-4/IL-13 supported fibroblast differentiation into myofibroblasts via transforming growth factor beta 1 (TGF-β1) in dependence of fibril density, suggesting a M2a-like phenotype. Overall, our results suggest that collagen fibril density can modulate macrophage response to favor tissue functions. Understanding of immune response in such complex 3D microenvironments will contribute to the novel therapeutic strategies for improving tissue repair, as well as guidance of the design of immune-modulated materials. Keywords: macrophage; collagen fibril density; immunomechanobiology; monocyte infiltration; fibroblast differentiation 1. Introduction Macrophages are vital innate immune cells with remarkable plasticity that allow them to respond efficiently to pro- and anti-inflammatory signals within their microenvironment [1,2]. Their phenotype Bioengineering 2020, 7, 33; doi:10.3390/bioengineering7020033 www.mdpi.com/journal/bioengineering Bioengineering 2020, 7, 33 2 of 19 and physiology can be altered by the cytokine secretion profile of other cells in the microenvironment in order to modulate the innate and adaptive immune responses [3]. Other than participating in immunological responses, they also play an important role in tissue development, homeostasis, and tissue repair. There is increasing evidence that macrophages in distinct tissues exhibits specific phenotype, activation state and functions [4], therefore it is probable that tissues play an important role in regulating macrophage physiology. Macrophage phenotype has been traditionally divided into 2 groups by simplified classification, namely classically activated macrophages (or pro-inflammatory) and alternatively activated macrophages (or anti-inflammatory). Pro-inflammatory macrophages were described as having inflammatory functions and were crucial for the resistance to pathogens and their elimination, whereas macrophages with anti-inflammatory functions were important for maintaining tissue integrity. However, this simplified nomenclature can create the illusory perception of homogeneity of macrophage subpopulations, which is not shown to exist in vivo, since both cells can phenotypically span the pro-inflammatory–anti-inflammatory spectrum to serve different functions in response to microenvironmental cues [5,6]. While soluble cytokines and cell–cell interactions are well studied, there is still a gap of knowledge on the effects of tissue characteristics as a microenvironmental signal. Tissue of the cellular microenvironment is dynamic and its density has been shown to be a prominent key player in many physiological and pathological processes, for example wound healing, scar formation and cancer development [7–10]. Furthermore, it has been demonstrated that tissue repair activities of macrophages are amplified and localized by tissue signals [11,12]. Monocytes are important in these processes; they extravasate into the targeted tissue and differentiate into specific subtype of macrophages to serve the functions that the targeted tissue needs [3]. Many studies, especially in the field of cancer biology, demonstrated that tissue density can modulate cellular invasion and metabolism [7,13,14]. However, less is known on how tissue density modulates monocyte-to-macrophage activation and in turn their phenotypes and functions. Since tissue density is a surrogate for tissue stiffness, a number of studies aimed to study how mechanical properties of materials affect macrophage response [15]. However, these have often been investigated in 2D cell culture models, which may only partially reproduce the complexity of in vivo tissue [16,17]. Although macrophages’ response has been studied in three-dimensional (3D) collagen or fibrin matrices in dependence of ligand presentation [18] and stiffness via crosslinkers [18–20], inconsistency in the results has been reported due to the difference of differentiation protocols and cell source used. Additionally, while these studies attempted to demonstrate the important role of tissue mechanical properties on macrophage response, the latter is often investigated based on a number of phenotypic specific markers without taking into account their cellular functions [21]. Despite the growing interest in this field, the direct effect of tissue density on macrophage phenotypes and functions has not been well established. Investigating how tissue density affects macrophage polarization and function allows us to understand how macrophages regulate physiological and non-homeostatic processes, especially in wounds and cancer. Furthermore, it will contribute to the development of novel therapeutic strategies for diseases, as well as guidance of the design of immuno-modulated materials. This study aimed to investigate the role of tissue/fibril density in the modulation of macrophage phenotypes and functions. Firstly, we embedded monocytic cell line THP-1 into 3D collagen matrices with different fibril density, and differentiated and activated them into specific subtypes of macrophages. The expression of phenotype-specific markers and cytokine secretion profiles were analyzed, characterizing the activation profile of macrophages in dependence of fibril density. To bridge the phenotypical characteristics of macrophages with their cellular functions, we studied their capacity in modulating monocyte infiltration and fibroblast differentiation by co-culturing with THP-1 monocytes and human primary fibroblasts, respectively. Bioengineering 2020, 7, 33 3 of 19 2. Materials and Methods 2.1. Preparation and Characterization of 3D Collagen Matrix Rat Tail type I Collagen (Advanced BioMatrix, Inc. San Diego, CA, USA), 500 mM phosphate buffer (Sigma-Aldrich, St. Louis, MO, USA) and 0.1% acetic acid (Sigma-Aldrich) were used to prepare collagen solution at a concentration of 1 and 3 mg/mL, as previously published [22,23], representing the tissue matrix during the inflammatory and proliferative phases, respectively. THP-1 cells were resuspended in the prepared collagen solution and were transferred onto a glutaraldehyde-coated coverslip, allowing covalent binding of collagen matrix via lysine side chain [24]. The number of THP-1 cells was 1 × 105 per 3D collagen matrix. Collagen fibrillogenesis was initiated at 37 ◦ C, 5% CO2 and 95% humidity, embedding cells uniformly in 3D space. Endotoxin levels of collagen was measured to be 0.025 EU/mL and determined that embedded THP-1 were not perturbed by the matrices to invoke an immune response (Supplementary Figure S1). Cell-free collagen matrices were analyzed to assess their topological and mechanical properties. Briefly, for topological analysis, the collagen matrices were stained with 50 µM of 5-(and-6)-carboxytetramethylrhodamine succinimidyl ester (TAMRA-SE, Sigma-Aldrich, USA) and visualized using a confocal laser scanning microscope (cLSM) (SP8; Leica, Wetzlar, Germany). The cLSM stacked images were analyzed using a home-built MATLAB script (MATLAB 2019a; MathWorks, USA), as described elsewhere [22,23]. Mechanical properties of cell-free scaffolds were analyzed non-destructively via rheological measurement using ElastoSens Bio 2 (Rheolution, Montreal, QC, Canada). The topological and mechanical characterizations were performed at least in triplicate. 2.2. Cell Culture and Macrophages Differentiation Human monocytic THP-1 cell line (ATCC, Manassas, VA, USA) were cultured in RPMI-1690 media supplemented with 10% fetal bovine serum (FBS), 1% sodium pyruvate, 0.01% of mercaptoethanol and 1% penicillin/streptomycin at 37 ◦ C, 5% CO2 and 95% humidity. Differentiating and activation protocols of THP-1-derived macrophages were adapted and modified from Genin et al. [25]. THP-1 cells embedded in 3D collagen matrices were terminally differentiated into uncommitted macrophages (MPMA ) with 300 nM phorbol 12-myristate 13-acetate (PMA; Sigma-Aldrich, Germany) in RPMI 1640 media without FBS supplement. After 6 h, differentiating media was removed. Cells were washed with phosphate-buffered saline (PBS) and rested for 24 h in RPMI 1640 without FBS supplement and PMA. Afterwards, cells were activated for 48 h into pro-inflammatory macrophages (MLPS/IFNγ ) by adding 10 pg/mL lipopolysaccharide (LPS; Sigma, USA) and 20 ng/mL interferon-gamma IFNγ (Biolegend, San Diego, CA, USA), or into anti-inflammatory macrophages (MIL-4/IL-13 ) with 20 ng/mL interleukin 4 (IL-4; Biolegend, USA) and 20 ng/mL interleukin 13 (IL-13; Biolegend, USA). 2.3. Quantitative Analysis of Macrophage Cell-Surface Markers To characterize cells regarding their surface markers, cells were stained with HLA-DRA (subunit of MHC-II) monoclonal antibody (clone: LN3) conjugated with APC-eFluor 780, CD163 monoclonal antibody (clone: eBioGHI/61 (GHI/61)) conjugated with APC, and CD206 (mannose receptor, MMR) monoclonal antibody (clone: 19.2) conjugated with PE for 30 min at standard cell culture conditions. Antibodies were diluted in cell culture medium at a ratio of 1:200. To eliminate the dead cells from the analysis, eBioscience™ Fixable Viability Dye eFluor™ 506 (dilution 1:800) was co-stained with the antibodies. All antibodies and cell viability dye were purchased from eBioscience (eBioscience, San Diego, CA, USA). Stained cells were analyzed using an Attune NxT Flow Cytometer equipped with an autosampler (Thermo Fisher Scientific, Waltham, MA, USA). Compensation settings were applied prior to running the analysis. Cell surface markers are elevated as compared to unstained samples. Unstained macrophages were depicted in Supplementary Figure S2. Experiments were performed in 4 replicates. Bioengineering 2020, 7, 33 4 of 19 2.4. Quantitative Analysis of Pro-Inflammatory Cytokines To analyze cytokines secreted by macrophages, cell culture supernatants were collected after cell activation. A bead-based multiplex immunoassay for IL-1β, IL-6, IL-8, IL-10, IL-12p70 and TNFα (Biolegend, USA) was utilized to quantify cytokines following instructions by the manufacturer. Samples were analyzed using an Attune NxT Flow Cytometer equipped with autosampler (Thermo Fisher Scientific, USA). Data analysis was undertaken using LEGENDplex™ data analysis software (Biolegend, USA). Experiments were performed in 10 replicates. 2.5. Macrophage-Induced Monocytes Infiltration into 3D Collagen Matrices Prior to performing co-culture experiments, 1 × 106 THP-1 monocytic cells were stained with 1 µM of carboxyfluorescein succinimidyl ester (CFSE; Biolegend, USA) for 15 min at standard cell culture conditions. Afterwards, cells were washed 2 times with PBS and rested in cell culture media. The CFSE staining allowed identification of macrophages from monocytes. Macrophages were co-cultured with 1 × 105 CFSE-stained THP-1 monocytic cells for 24 h at standard cell culture conditions. As a control condition, 1 × 105 CFSE stained THP-1 were cultured onto 3D collagen matrices in the absence of macrophages. Afterwards, cells were fixed with 4% paraformaldehyde (Biolegend, USA) for 10 min and permeabilized with 0.1% Triton X100 (Sigma-Aldrich, Germany) for 10 min. Subsequently, cells were washed 2 times with PBS and stained with Hoechst 33342 (dilution 1:10,000 in PBS; Invitrogen, Carlsbad, CA, USA). To quantify monocyte infiltration into 3D collagen matrices, fluorescence stacked images of Hoechst 33342 and CFSE were obtained. Data analysis was performed using a custom-built MATLAB script (MATLAB 2019a; MathWorks, USA), as previously published [22,23]. Percentage of infiltrated monocytes and maximal distance were quantified as a function of macrophage subtypes and tissue density. Experiments were undertaken in 4 replicates at 3 random positions per sample. 2.6. Macrophage-Induced Myofibroblast Differentiation in 3D Collagen Matrices Macrophages were co-cultured with 1 × 105 human primary dermal fibroblasts (ATCC, USA) within the matrix for 3 days in standard cell culture conditions. As negative and positive controls, 105 human primary dermal fibroblasts were cultured onto 3D collagen matrices with and without the presence of 10 ng/mL of transforming growth factor beta 1 (TGF-β1; Biolegend, USA). Afterwards, RNA was extracted and gene expression of myofibroblast markers was quantified using quantitative polymerase chain reaction (qPCR, see gene expression analysis section). Furthermore, as TGF-β1 is involved in the myofibroblast differentiation, we quantified the free active TGF-β1 in the cell culture supernatant using the TGF-β1 ELISA (enzyme-linked immunosorbent assay) kit (Biolegend, USA) following the protocol by the manufacturer. All experiments were undertaken in 4 replicates. 2.7. Gene Expression Analysis Gene expression analysis for myofibroblast differentiation was performed using an established protocol, as published [26]. Briefly, TRIzol (Invitrogen, USA) was used to extract the total RNA. The RNA obtained was converted into complementary DNA (cDNA) using a high-capacity cDNA reverse transcription kit (Applied Biosystems, Foster City, CA, USA). The concentration, and the ratio of absorbance at 260 nm and 280 nm of cDNA were quantified using nanodrops (Thermo Fisher Scientific, USA) prior to performing gene expression analysis. Ribosomal protein S26 (RPS26) and beta-actin were used as a reference gene. The primers were synthesized from Bioneer (Korea). The primer sequences are listed in Supplementary Table S1. qPCR was performed using the SYBR Green PCR Master Mix (Applied Biosystems, USA). The qPCR procedure was set as follows: denaturation for 5 min at 95 ◦ C; 45 cycles of denaturation (95 ◦ C, 15 s), annealing under primer-specific conditions (30 s) and target gene-specific extension (30 s at 72 ◦ C). Fluorescence signals were measured for 20 s at 72 ◦ C. To confirm Bioengineering 2020, 7, 33 5 of 19 the specificity of the PCR products, melting curve analysis was performed at the end of each run. Experiments were undertaken in 4 replicates. Bioengineering 2.8. Statistical 2020, 7, x FOR PEER REVIEW Analysis 5 of 19 2.8. Statistical Analysis Experiments were performed at least in 4 replicates unless otherwise stated. Error bars indicate standard deviation Experiments (SD). Levels were of statistical performed significance at least in 4 replicateswere determined unless by a Mann–Whitney otherwise stated. test or Error bars indicate by one-way standard ordeviation two-way(SD). ANOVA Levelsfollowed by Tukey’s of statistical using significance wereGraphPad Prism determined by a 8Mann–Whitney (GraphPad Software, test San Diego, CA, USA). The significance level was set at < or by one-way or two-way ANOVA followed by Tukey’s using GraphPad Prism 8 (GraphPad p 0.05. Software, San Diego, CA, USA). The significance level was set at p < 0.05. 3. Results 3. Results We postulate that dynamic changes of tissue/collagen fibril density during physiological and Weprocesses pathological postulate that dynamic changes can modulate of tissue/collagen macrophage activation fibril density functions and cellular during physiological andThis significantly. pathological processes can modulate macrophage activation and cellular functions significantly. is especially vital during tissue repair where tissue density changes from loose to dense tissue. To This is especially vital during tissue repair where tissue density changes from loose to dense tissue. To address the underlying questions, we utilized collagen matrices at concentrations of 1 and 3 mg/mL to address the underlying questions, we utilized collagen matrices at concentrations of 1 and 3 mg/mL simulate tissue density differences. Human THP-1-derived macrophages were used for this study since to simulate tissue density differences. Human THP-1-derived macrophages were used for this study they closely resemble native monocyte-derived macrophages [15,20,25,27]. It is well accepted that since they closely resemble native monocyte-derived macrophages [15,20,25,27]. It is well accepted monocytes first infiltrate that monocytes into tissue first infiltrate and into then, tissue depending and on the on then, depending microenvironmental cues,cues, the microenvironmental theythey activate activate into specific subtype of macrophages. To mimic this situation in vitro, THP-1 cells were into specific subtype of macrophages. To mimic this situation in vitro, THP-1 cells were embedded embedded into 3D collageninto 3D collagen matrices, matrices, differentiated differentiated and activated,andasactivated, depictedasindepicted Figure in Figure 1A. 1A. 1. Differentiation FigureFigure andand 1. Differentiation activation ofofmacrophages, activation macrophages, matrix characterization matrix characterization andand morphology morphology of of THP-1THP-1 derived derived macrophages. macrophages. (A)(A) Schematicillustration Schematic illustration of of macrophage macrophagedifferentiation differentiationandand activation activation in 3D in 3D collagen collagen matrix. matrix. (B) (B) Representativeimages Representative imagesof of collagen collagen matrices matricesvisualized visualized using confocal using laserlaser confocal scanning scanning microscopy microscopy (cLSM, (cLSM, scale scale barbar = 50 = 50 µm).(C) µm). (C)Topological Topological andand mechanical mechanicalcharacterization characterizationof of 1 1 mg/mL and 3 mg/mL collagen matrices. Topological analysis was quantified at mg/mL and 3 mg/mL collagen matrices. Topological analysis was quantified at 3 random positions of3 random positions of each All each matrix. matrix. All experiments experiments were performed were performed in triplicates. in triplicates. (Data (Data are are represented represented as mean as mean ± ± standard standard deviation (SD); * significance level of p < 0.05). (D) Representative cell morphology deviation (SD); * significance level of p < 0.05). (D) Representative cell morphology of THP-1 derived of THP- 1 derived macrophages in 1 mg/mL and 3 mg/mL of 3D collagen matrices (scale bar = 20 µm). macrophages in 1 mg/mL and 3 mg/mL of 3D collagen matrices (scale bar = 20 µm). 3.1. Characterization of Cell-Free 3D Collagen Matrices Prior to performing the cell experiment, we characterized our reconstituted 3D collagen matrices in terms of topological and mechanical properties in a cell-free condition. As shown in Figure 1B, it Bioengineering 2020, 7, 33 6 of 19 3.1. Characterization of Cell-Free 3D Collagen Matrices Prior to performing the cell experiment, we characterized our reconstituted 3D collagen matrices in terms of topological and mechanical properties in a cell-free condition. As shown in Figure 1B, it could be observed that fibril density is increased with collagen monomer concentration. The mean pore size was quantified using an image-based toolbox and measured to be 11.07 ± 0.62 µm and 3.74 ± 0.75 µm for 1 mg/mL and 3 mg/mL collagen concentrations, respectively (Figure 1C). The pore size of our 3D collagen matrices are in the range as it is found between 3 to 11 µm in the interstitial tissue [28]. Collagen fibril size was 669 ± 36 nm, similar for both matrices, hence pore size alone is the variable factor in our collagen system here. Elastic modulus of the collagen matrices was inspected using non-destructive rheological measurement, and the mean elastic modulus was measured to be 41.58 ± 9.18 Pa and 211.57 ± 6.79 Pa for 1 mg/mL and 3 mg/mL collagen concentrations, respectively (Figure 1C). The results demonstrated that the elastic modulus increased with higher fibril density, as published elsewhere [22,28]. The thickness of collagen matrices were in the range of 300–400 µm, allowing ample diffusion of nutrients and mediators into the matrices [29]. THP-1 cells were embedded into the well-defined 3D collagen matrices, differentiated and activated into specific subtypes of macrophages, MPMA , MLPS/IFNγ and MIL-4/IL-13 . Cell embedding did not affect the fibrillation and topological properties of the collagen matrix [29,30]. All macrophage subtypes appeared to exhibit a rounded morphology and no matrix-dependent cell morphological differences could be observed (Figure 1D). As is known, pro- (M1) and anti-inflammatory (M2) macrophages could be distinguished by their morphological appearance on 2D cell culture models, but not in 3D fibrillar collagen matrices [18,31]. To demonstrate how specific subtype of macrophages behave in different collagen fibril density, we studied the expression of cell-surface markers and associated genes, cytokines secretion and cell functionality. 3.2. Dense Matrix Enhanced Expression of MHC-II, CD163 and CD206 As shown in Figure 1D and mentioned above, macrophage subtypes could not be distinguished by their morphology when cultured within 3D collagen matrices. To distinguish macrophage subtypes, we quantified our cells with widely accepted macrophage-specific markers, namely MHC-II (through HLA-DR at the protein level, and HLA-DRA at the gene level), CD163 and CD206, using flow cytometry and qPCR after macrophage activation. While upregulation of MHC-II is commonly used as an indicator for increased inflammatory activities via pro-inflammatory antigen-presenting cells, CD163 and CD206 are reserved for the opposite, anti-inflammatory cells [3]. As shown in Figure 2A, MLPS/IFNγ significantly expressed higher MHC-II, a major histocompatibility complex class II, at both protein (Figure 2A(ii)) and gene expression level (Figure 2A(iii)), compared to MPMA and MIL-4/IL-13 . At the gene level, we found MHC-II tends to be expressed higher in denser tissue, this is significantly pronounced for MLPS/IFNγ . This fibril density dependency is not seen at the protein level. A recent report suggested that MHC-II expression is regulated via exogenous mechanical signals by activation of the p38 MAPK pathway [32] which might explain the differential gene expression of MHC-II in our case. fibrous form at the microscopic level, as reported [36]. This suggests that CD206 expression is sensitive to the biophysical changes of cellular microenvironments. Besides, CD206 plays a role in the cell attachment to collagen via fibronectin type II domain of the CD206 receptor [37] as well as in collagen degradation [38], which might explain the high expression of CD206 in denser collagen matrices. Bioengineering 2020,Furthermore, 7, 33 CD206 is reported to be expressed non-specifically in macrophage7 of 19 phenotypes [39]. 2. Quantitative FigureFigure 2. Quantitativeanalysis analysisof ofTHP-1 THP-1derived derived macrophage surfacemarkers. macrophage surface markers. (A) (A) MHC-II, MHC-II, (B) (B) CD163 and (C) CD206 were analyzed at both the protein level and gene expression. CD163 and (C) CD206 were analyzed at both the protein level and gene expression. All subfigures, (i) All subfigures, depictsthe (i) depicts therepresentative representative histogram histogram plotplot of stained cells ascells of stained a function as a of fluorescence function signal intensity. of fluorescence signal The respective intensity. quantitative The respective analysis analysis quantitative of geometric mean of themean of geometric fluorescence intensity is shown of the fluorescence in is intensity shown subfigure (ii) All in subfigure (ii) experiments All experiments werewere performed at least performed in triplicate. at least GeneGene in triplicate. expression was was expression quantitatively quantitatively analyzed analyzed usingusing quantitative quantitative polymerase polymerase chainchain reaction reaction (qPCR) (qPCR) andand thethe results results areare shown shown in(iii) in subfigures subfigures Data are(iii) Data are represented represented as mean ± as SD;mean ± SD; * significance * significance level of p level of The < 0.05. p < 0.05. The #, characters §, % represent the significance level of p < 0.05 when compared to MPMA , MLPS/IFNγ , and MIL-4/IL-13 macrophages, respectively. CD163, a scavenger receptor, was uniformly expressed at the cell surface for all our macrophage subtypes (Figure 2B(ii)). However, significant differences in CD163 gene expression could be observed in MIL-4/IL-13 and it also appeared to be tissue density dependent (Figure 2B(iii)), whereby a denser matrix stiffness promotes enhanced expression of CD163 in MIL-4/IL-13 . Previous studies investigating the expression of CD163 on THP-1-derived macrophages have been conflicting. Tedesco et al. reported that CD163 is undetectable, although Genin et al. revealed no significant difference in CD163 expression between THP-1-derived pro- and anti-inflammatory macrophages [25,33,34], and finally Zhu et al. demonstrated that approximately 20% of THP-1-derived anti-inflammatory macrophages expressed Bioengineering 2020, 7, 33 8 of 19 CD163 [27]. All these studies were performed on 2D cell culture plastic. Furthermore, CD163 was found to be hardly presented at the cell surface of peripheral blood mononuclear cell (PBMC)-derived macrophages both in 3D and 2D culture models [30]. It has also been reported that CD163 is expressed in macrophages activated by IL-10 rather than IL-4 in vitro [35]. Next, we analyzed CD206 (Figure 2C), a mannose receptor. No differences in CD206 on the cell surface could be observed in all our macrophage subtypes (Figure 2C(ii)), as previously reported in THP-1-derived pro- and anti-inflammatory macrophages [33]. However, we found upregulation of CD206 gene expression in dense collagen matrix in all macrophages. The expression level of CD206 was reported to be correlated to collagen morphology, because collagen can exist in the globular or fibrous form at the microscopic level, as reported [36]. This suggests that CD206 expression is sensitive to the biophysical changes of cellular microenvironments. Besides, CD206 plays a role in the cell attachment to collagen via fibronectin type II domain of the CD206 receptor [37] as well as in collagen degradation [38], which might explain the high expression of CD206 in denser collagen matrices. Furthermore, CD206 is reported to be expressed non-specifically in macrophage phenotypes [39]. Overall, we found that MHC-II was expressed higher in MLPS/IFNγ macrophages in dense matrix at the gene level but was similar at the protein level, and higher at both the gene and protein level when compared to other macrophage phenotypes. CD163 was expressed on all macrophages, but it found significantly higher in MIL-4/IL-13 at the gene level, and significantly higher in dense matrix. CD206 appeared to be present on all macrophages, but higher in MLPS/IFNγ and MIL-4/IL-13 at gene level in dense matrix. Our data suggest the importance of fibril density in regulation of macrophage activation. However, the analysis of surface markers alone is inconclusive for distinguishing between the THP-1-derived macrophage subtypes. It has to be mentioned that in our study, macrophages were differentiated and activated in 3D collagen matrices, thus the expression pattern of surface markers might differ from cells differentiated and activated on 2D surfaces. 3.3. Dense Matrix Enhances Secretion of Cytokines To further characterize macrophage activation state along with the cell-surface markers, we investigated the secretion of cytokines by our THP-1 derived macrophage subtypes. We measured the secretion level of IL-1β, IL-6, IL-8, IL-10, IL-12p70 and TNFα, during two different biologically relevant scenarios, in the presence of activators and 3 days after the removal of activators in the cell culture media (Figure 3). Removal of activators not only serve to allow us to observe changes in cytokine secretion profiles, moreover, it is important for our co-culture experiments to be void of exogenous activators other than biochemical signals from our macrophage systems. The microenvironment with and void of activators is also physiologically relevant to tissue repair, recapitulating the overlapping inflammatory phase and repair phase, respectively. As shown in Figure 3, MLPS/IFNγ secreted higher levels of IL-1β and TNFα when compared to MPMA and MIL-4/IL-13 . Except for IL-8, which is undetectable in MPMA , the secretion pattern of released cytokines from MPMA was similar to MIL-4/IL-13 . The secretion levels of IL-10 and IL-12p70 were relatively low in all macrophages with no differences between subtypes, contrary to the secretion levels of the same cytokines from THP-1-derived macrophages on 2D surfaces [15] and 3D freeze-dried and crosslinked collagen scaffolds [20]. We quantified gene expression of IL-10 and found very low expression. The inconsistency of the results might arise from the activation protocol, cell culture dimensionality, scaffold preparation, as well as differences in pore size and elasticity range. substrate) compared to untreated collagen matrices [18]. Enhanced stiffness via increment of fibril density or tuning of collagen fibril bending stiffness using crosslinkers could very well impact different downstream signaling. The exact mechanism of how secretion of pro-inflammatory cytokines is enhanced with an increase of fibril density is still unknown. At this stage we could not provide mechanistic insights into how fibril density regulates cytokine release from macrophages.9 of 19 Bioengineering 2020, 7, 33 However, we aim to address this underlying question in future studies. Figure 3. Cytokine secretion by THP-1-derived macrophages in the presence of activators and 3 days Figure 3. Cytokine secretion by THP-1-derived macrophages in the presence of activators and 3 days after removal of activators. Cytokine secretion (A) IL-1β, (B) IL-6, (C) IL-8, (D) IL-10, (E) IL-12p70 and after removal of activators. Cytokine secretion (A) IL-1β, (B) IL-6, (C) IL-8, (D) IL-10, (E) IL-12p70 and (F) TNFα. (F) TNFα. experiments The The experiments werewereperformed performed at at least inin least 4 replicates. 4 replicates.Data Dataare are represented mean±± SD; as mean represented as * significance SD; * significance < 0.05. level of plevel of p The < 0.05.characters #, §, % The characters #, represent the significance §, % represent level the significance p

20 µm µm cells and (C) maximum migration distance of CFSE-stained THP-1 cells. Cells found >20 beneath beneath the Coll I matrix surface were counted as migrated cells. The maximum migration distance was was the Coll I matrix surface were counted as migrated cells. The maximum migration distance defineddefined as the asdistance the distance crossed crossed byby 10%10%ofofallallcells. cells. (D) (D) Quantitative Quantitative analysis analysisof of MCP-1 MCP-1secretion by by secretion macrophages in a monoculture condition after 1 day of culture using enzyme-linked macrophages in a monoculture condition after 1 day of culture using enzyme-linked immunosorbent immunosorbent assay assay (ELISA, (ELISA, n =Data n = 4). 4). Data areare represented represented asasmean mean±± SD; SD; ** significance significancelevel ofof level p