rsif.royalsocietypublishing.org

Research Cite this article: Graney PL, Roohani-Esfahani S-I, Zreiqat H, Spiller KL. 2016 In vitro response of macrophages to ceramic scaffolds used for bone regeneration. J. R. Soc. Interface 13: 20160346. http://dx.doi.org/10.1098/rsif.2016.0346

Received: 3 May 2016 Accepted: 4 July 2016

Subject Category: Life Sciences – Engineering interface Subject Areas: biomaterials Keywords: macrophage phenotype, immunomodulation, bone repair, macrophage – biomaterial interactions

Author for correspondence: Kara L. Spiller e-mail: [email protected]

Electronic supplementary material is available at http://dx.doi.org/10.1098/rsif.2016.0346 or via http://rsif.royalsocietypublishing.org.

In vitro response of macrophages to ceramic scaffolds used for bone regeneration Pamela L. Graney1, Seyed-Iman Roohani-Esfahani2, Hala Zreiqat2 and Kara L. Spiller1 1 Biomaterials and Regenerative Medicine Laboratory, School of Biomedical Engineering, Science and Health Systems, Drexel University, Philadelphia, PA 19104, USA 2 Biomaterials and Tissue Engineering Research Unit, School of Aerospace, Mechanical and Mechatronic Engineering, The University of Sydney, Sydney, New South Wales 2026, Australia

KLS, 0000-0001-7798-1490 Macrophages, the primary cells of the inflammatory response, are major regulators of healing, and mediate both bone fracture healing and the inflammatory response to implanted biomaterials. However, their phenotypic contributions to biomaterial-mediated bone repair are incompletely understood. Therefore, we used gene expression and protein secretion analysis to investigate the interactions in vitro between primary human monocytederived macrophages and ceramic scaffolds that have been shown to have varying degrees of success in promoting bone regeneration in vivo. Specifically, baghdadite (Ca3ZrSi2O9) and strontium–hardystonite–gahnite (Sr–Ca2ZnSi2O7 –ZnAl2O4) scaffolds were chosen as two materials that enhanced bone regeneration in vivo in large defects under load compared with clinically used tricalcium phosphate–hydroxyapatite (TCP–HA). Principal component analysis revealed that the scaffolds differentially regulated macrophage phenotype. Temporal changes in gene expression included shifts in markers of pro-inflammatory M1, anti-inflammatory M2a and pro-remodelling M2c macrophage phenotypes. Of note, TCP–HA scaffolds promoted upregulation of many M1-related genes and downregulation of many M2a- and M2c-related genes. Effects of the scaffolds on macrophages were attributed primarily to direct cell–scaffold interactions because of only minor changes observed in transwell culture. Ultimately, elucidating macrophage–biomaterial interactions will facilitate the design of immunomodulatory biomaterials for bone repair.

1. Introduction More than 6 million bone fractures occur annually in the USA alone [1,2], and 5–10% of these fail to heal adequately owing to bone loss, failed fixation, infection and poor vascularization [1]. Bone is a unique tissue in that small fractures can self-repair, without fibrous scarring [3], but large, critical-sized bone defects remain a challenge [4]. Therefore, there is a significant need for regenerative medicine strategies that use materials to harness the intrinsic ability of bone to repair itself. Fracture healing occurs through a complex cascade of events initiated by a trauma-induced inflammatory response and involves ossification and bone remodelling [5–8]. The process of inflammation represents a crucial stage in bone repair. Inflammatory cells not only recruit osteoprogenitor cells to the fracture site, but also stimulate revascularization [6,8,9], a required process for tissue survival. Studies have shown that in the absence of a fracture haematoma, which normally initiates the inflammatory cascade, fracture healing is impaired [10,11]. However, chronic inflammation inhibits healing of bone [12] and many other tissues [13]. Thus, a precisely controlled inflammatory response is critical for successful fracture healing.

& 2016 The Author(s) Published by the Royal Society. All rights reserved.

2

study 3: grain size effects

MCSF

cts M0 y 1: stud oral effe macrophage temp

1–6 days characterize macrophage behaviour

stud dire y 2: inte ct versu ract ions s indire ct

6 days vs

day 5

1–6 days post-seeding

isolate monocytes from human blood culture in MCSF

seed M0 macrophages on scaffolds expose for 1–6 days (37ºC, 5% CO2)

isolate RNA and analyse gene expression or quantify protein secretion

Figure 1. Schematic of study design. Study 1: temporal effects. Unactivated macrophages were seeded directly on baghdadite, Sr –HT – gahnite and TCP – HA scaffolds for 6 days, and characterized in terms of gene expression at early and late times of direct cell – scaffold contact. Study 2: effects of soluble factors. Unactivated macrophages were seeded directly onto scaffolds or co-cultured using a transwell insert to discern the contributions to macrophage gene expression patterns stemming from physical scaffold properties and those resulting from ion dissolution. Study 3: grain size effects. The grain size of baghdadite scaffolds was varied and the ensuing temporal response on macrophage protein secretion was assessed at early and late stages of contact. (Online version in colour.)

Macrophages, the primary cells of the inflammatory response, are widely accepted as regulators of wound healing [14] and have been shown to play important roles in endochondral ossification [15], bone deposition and mineralization [16] and differentiation of mesenchymal progenitors [17] during fracture repair. This multifaceted behaviour stems from the fact that macrophages rapidly shift their phenotype in response to changing environmental stimuli, which is critical for successful wound healing [18]. In the normal healing process of many tissues, including bone fractures, the macrophage population shifts from predominantly pro-inflammatory (M1) to predominantly anti-inflammatory (M2) over time [19], perhaps because of sequential actions of the two populations in promoting angiogenesis [20,21]. M1 macrophages can also be activated in vitro by pro-inflammatory signals such as lipopolysaccharide and interferon-gamma. M2 macrophages can also be further described as ‘M2a’ (stimulated in vitro by interleukin (IL) 4 + IL13) or ‘M2c’ (stimulated in vitro by IL10, transforming growth factor-beta or glucocorticoids), which are believed to promote tissue deposition and remodelling, respectively, although their roles are incompletely understood [20,22]. Although microenvironmental signals have major effects on macrophage behaviour [23], it is not understood how biomaterials impact the microenvironment in a bone fracture to modulate macrophage behaviour and affect healing outcomes. Thus, there is a need to investigate how macrophages respond to biomaterials used in bone regeneration. Recently, Roohani et al. [24 –26] demonstrated that novel ceramic-based scaffolds, baghdadite (Ca3ZrSi2O9) and strontium–hardystonite –gahnite (Sr –HT– gahnite; Sr–Ca2 ZnSi2O7 –ZnAl2O4), were more successful at regenerating large bone defects under load than the clinical standard, tricalcium phosphate –hydroxyapatite (TCP– HA) scaffolds. Both baghdadite and Sr –HT–gahnite resulted in extensive new bone formation and complete bridging of critical-sized radial segmental defects in rabbits 12 weeks post-implantation, compared with only partial bridging demonstrated by TCP– HA [24–26]. In sheep segmental defect models, baghdadite scaffolds showed a significant bridging of the critical-sized defect (average 80%) with evidence of bone infiltration and

remodelling within the scaffold implant [27]. The success of these may be a result of direct effects on cells involved in bone formation, including osteoblasts [25] and osteoclasts [28,29]; the effects of these scaffolds on the behaviour of recruited macrophages are not known. This study was designed to explore the response of primary human macrophages in vitro to scaffolds that have been previously shown to exhibit varying levels of success at regenerating bone in vivo. We first evaluated the temporal changes in gene expression of unactivated primary human monocyte-derived macrophages seeded directly onto the scaffolds (figure 1, study 1). Then, to determine if the mechanism of macrophage modulation was dependent on release of soluble factors from the scaffolds, we compared the response of unactivated macrophages in direct contact or in transwell co-culture with the scaffolds for 6 days (figure 1, study 2). Finally, we completed a preliminary investigation of scaffold grain size as a potential contributing factor to macrophage activation, because topographical and mechanical cues are known to affect macrophage phenotype (figure 1, study 3). Macrophage response was evaluated in terms of gene expression for a panel of markers that were previously determined to indicate the M1, M2a and M2c phenotypes in vitro [20,22], because of their distinct roles in promoting vascularization and tissue repair, as well as those more generally associated with angiogenesis and osteogenesis.

2. Material and methods 2.1. Fabrication of scaffolds TCP– HA, baghdadite and Sr–HT– gahnite scaffolds were prepared using previously described methods [24] (see also electronic supplementary materials). To prepare baghdadite scaffolds with two different grain sizes, the fabrication method was identical to that for TCP–HA and Sr–HT–gahnite scaffolds, except a sintering temperature of 13808C for 3 h (small grain) or 12 h (large grain) was used. The porosity, compressive strength and modulus, and microstructure of the scaffolds were characterized as we have previously described [24,25].

J. R. Soc. Interface 13: 20160346

day 0

rsif.royalsocietypublishing.org

primary human monocytes

2.2. Ion dissolution studies

Primary human monocytes, isolated via negative selection from peripheral blood mononuclear cells, were purchased from the University of Pennsylvania Human Immunology Core (Philadelphia, PA). Monocytes were seeded in ultra-low attachment flasks (Corning) and cultured for 5 days in RPMI 1640 medium, supplemented with 10% heat-inactivated human serum, 1% penicillin –streptomycin and 20 ng ml21 macrophage colony stimulating factor (MCSF) for macrophage differentiation in vitro as we have previously described [20,22]. The medium was changed on day 3 and, on day 5, unactivated (M0) macrophages were gently scraped and collected.

2.4. Temporal effects of scaffolds on macrophage behaviour All scaffolds were autoclaved and equilibrated in 24-well ultra-low attachment plates in culture medium for 15 min. Unactivated M0 macrophages (1  106 cells) were seeded directly onto the scaffolds in 15 ml and allowed to attach for 1 h at 378C and 5% CO2, as described previously [31–33]. After 1 h, the culture medium was adjusted to a final volume of 1 ml, and the samples were incubated for an additional 2– 6 days in the presence of 20 ng ml21 MCSF, with medium exchange on days 2 and 3. M0 macrophages exposed to culture medium alone served as a control. On days 2 and 6, the scaffolds were transferred into 1 ml TRIzol and stored at 2208C until RNA extraction. Control cells, not exposed to scaffolds (M0), were gently scraped and collected, and centrifuged at 400g for 7 min. Cell pellets were resuspended in buffer RLT (RNeasy Micro Kit; Qiagen), and stored at 2208C until RNA extraction. Temporal studies investigating the effects of varying baghdadite grain size on macrophage activation were conducted using the same procedure; medium was collected on days 2, 3 and 6 and stored at 2808C until analysis.

2.5. Indirect macrophage –scaffold interactions To identify whether any soluble factors released from the scaffolds, such as dissolved ions, affect macrophage gene expression, scaffolds were separated from the macrophages by placing them in the apical chamber of transwell inserts (Millipore) in 24-well ultra-low attachment plates in 250 ml culture medium. Unactivated M0 macrophages (1  105 cells) were seeded in the basolateral chamber in 750 ml of culture medium. The scaffolds and cells were co-cultured in this way at 378C and 5% CO2 for 6 days, with medium exchange on days 1 and 3. On day 6, cells were lysed directly in the well in buffer RLT and stored at 2208C overnight. All samples were cultured in medium containing 20 ng ml21 MCSF.

RNA was extracted from the scaffolds or cells, DNA was inactivated with DNAse I and cDNA was prepared as we have previously described [21]. Quantitative reverse transcription polymerase chain reaction (RT-PCR) was performed using 20 ng cDNA and Fast SYBR Green Master Mix (Applied Biosystems) according to the manufacturer’s instructions. Mean quantification cycle (Cq) values were calculated from technical replicates (n ¼ 2). The expression of target genes was then normalized to the reference gene, GAPDH. Data shown represent the mean fold change + s.e.m. (n  4). Custom-designed primers (electronic supplementary material, table S1) were synthesized by Thermo Fisher Scientific.

2.7. Principal component analysis In order to visualize the global representation of gene expression data and variation between samples resulting from macrophage– scaffold interactions, principal component analysis (PCA) was implemented using MATLABw software (The MathWorks, Inc., Natick, MA). PCA is a multivariate data analysis approach that reduces the dimensionality of the data by capturing most of the variation in the dataset into new variables known as principal components [34]. Samples can then be projected onto these uncorrelated principal components in order to detect similarities and dissimilarities among samples. Prior to analysis, the z-scores were computed to standardize the data and enable comparison across the dataset. The coefficients or loadings of each gene were used to identify the genes most contributing to PCA results.

2.8. Combinatorial M1/M2 scoring To quantify the relative gene expression of macrophages exposed to TCP– HA, baghdadite or Sr– HT – gahnite scaffolds, we combined data from a panel of genes into a single score indicative of the M1/M2 character of macrophages, using an algorithm that we have recently shown to accurately predict healing of human diabetic ulcers [35]. Briefly, the M1/M2 score was defined as the ratio of the sum of the raw values of M1 gene expression (TNF, VEGF, CCR7 and IL1B) to the sum of the raw values of M2a gene expression (MRC1, PDGF and TIMP3), such that higher scores represent increased pro-inflammatory (M1-like) behaviour with respect to M2a behaviour.

2.9. Protein secretion Conditioned media from cell–scaffold interactions, stored at 2808C following collection, were thawed on ice, centrifuged briefly and analysed for the presence of protein using commercially available kits. Human tumour necrosis factor (TNF), vascular endothelial growth factor (VEGF), platelet-derived growth factor beta polypeptide (PDGFB) and bone morphogenetic protein 2 (BMP2) mini ELISA development kits were purchased from Peprotech (Rocky Hill, NJ). Human matrix metalloproteinase 7 (MMP7) and matrix metalloproteinase 9 (MMP9) Quantikine ELISA kits were purchased from R&D Systems, Inc. (Minneapolis, MN). TIMP metalloproteinase inhibitor 3 (TIMP3, MIG-5) human ELISA kits were purchased from Abcamw (Cambridge, MA). All kits were used according to the manufacturer’s instructions.

2.10. Statistical analysis Data are represented as mean + s.e.m. Prior to analysis, the fold change in gene expression data was log-transformed, and statistical analysis for all studies was performed in GRAPHPAD PRISM v. 6.0 (GRAPHPAD Software, Inc., La Jolla, CA). Data were analysed using either two-way or one-way ANOVA, as indicated, with Tukey’s post hoc analysis. A multiple t-test analysis was performed

3

J. R. Soc. Interface 13: 20160346

2.3. Monocyte isolation and macrophage culture

2.6. Quantitative analysis of gene expression using RT-PCR

rsif.royalsocietypublishing.org

A simulated body fluid (SBF) solution (pH 7.4) was prepared according to the procedure described by Kokubo & Takadama [30], and used to study the ion release profile of ceramic scaffolds. Ion release was analysed for 28 days, even though cell culture experiments were conducted for only 6 days, in an effort to thoroughly characterize the scaffolds. Cubic-shaped scaffolds (8  8  8 mm) were immersed in the SBF solution at 378C for 1, 3, 7, 14, 21 and 28 days at a solid/liquid ratio of 150 mg l21 (n ¼ 5 per time point). All scaffolds were held in plastic flasks and sealed. The concentration of ions in the solutions after soaking was tested using inductively coupled plasma atomic emission spectroscopy (Optima 3000DV; Perkin Elmer, USA).

(a)

compressive strength (MPa)

compressive modulus (MPa)

b-Ca3(PO4)2-Ca10(PO4)6(OH)2

~82

0.10

10

baghdadite (SG)

Ca3ZrSi2O9

~82

0.27b

25

baghdadite (LG)

Ca3ZrSi2O9

~82

0.38

32

Sr–HT–gahnite

Sr-Ca2ZnSi2O7-ZnAl2O4

~82

4.10a

170a

chemical formula

TCP–HA (60–40%)

4

rsif.royalsocietypublishing.org

porosity (%)

scaffold

(b) baghdadite (LG)

1 mm

1 mm

J. R. Soc. Interface 13: 20160346

baghdadite (SG)

Sr–HT–gahnite

1 mm

(c)

ion release (ppm)

Sr–HT–gahnite

baghdadite (SG)

102

102

10

10

Ca Si Sr Zn Zr P

1

1

10–1 10–1

10–2

10–2

10–3

10–3

10–4 0

7

14 day

21

28

0

7

14 day

21

28

Figure 2. Properties of ceramic-based scaffolds. (a) Chemical composition and mechanical properties of TCP – HA, small grain (SG) baghdadite, large grain (LG) baghdadite and Sr – HT – gahnite scaffolds. (b) SEM of Sr– HT – gahnite and baghdadite scaffolds of varying grain size. (c) Ion dissolution profiles of Sr – HT – gahnite scaffolds in SBF and small grain baghdadite scaffolds in phosphate-buffered saline over 28 days. aReported in [25] and breported in [24]. (Online version in colour.) on data collected from the grain size effects on macrophage behaviour and statistical significance was determined using the Holm–Sidak method. In all analyses, p , 0.01 was considered significant to minimize the potential for false positives owing to multiple testing.

Baghdadite and Sr–HT–gahnite scaffolds released comparable levels of calcium ions; however, baghdadite released high levels of silicon ions as well as phosphorus and zirconium, whereas Sr–HT–gahnite dissolution medium contained only a fraction of silicon and undetectable levels of phosphorus and zirconium.

3. Results

3.2. Temporal effects of scaffolds on macrophage gene expression

3.1. Scaffold properties A summary of the mechanical properties and ion release profiles for TCP–HA, baghdadite and Sr–HT–gahnite is provided in figure 2. All scaffolds were approximately 82% porous; however, the unique microstructural design of Sr–HT–gahnite yielded increased compressive strength and modulus, resembling that of native bone [36], whereas TCP–HA scaffolds were the weakest. Scanning electron microscopy (SEM) revealed marked differences in the microstructures of the scaffolds (figure 2b). Submicrometre crystals, owing to the presence of gahnite and rectangular grains, were dispersed throughout Sr–HT–gahnite scaffolds. Crystal formation was absent from baghdadite scaffolds. As expected, the average grain size and size distribution of small grain baghdadite (2 mm) was comparatively much smaller than large grain baghdadite (4 mm). In addition, the ion release profiles of baghdadite and Sr–HT– gahnite were characteristic of their distinct compositions.

3.2.1. Principal component analysis and combinatorial scoring Principal component 1 (PC1) retained 25.6% of the variance within the gene expression set, and clustered the data based on time (figure 3a). Interestingly, PC2 distinguished the M0 control on day 2 from all other macrophage responses, which captured 20.9% of the variance in the data. Not surprisingly, these results indicate that macrophages behave very differently when cultured in three dimensions on ceramic scaffolds compared with two-dimensional culture on ultra-low attachment plastic. The leading drivers of the variance retained within PC1 were MARCO, BGLAP, TNFSF11 and VCAN, whereas MMP9, TGFB1, TIE1 and MMP7 were major contributors to the variance of PC2 (coefficients of all genes projected onto the principal components are provided in the electronic supplementary material, figure S1). Extending the analysis to PC3 and PC4, which captured 14.3% and 9.7% of the variance, respectively,

(a)

PC2 (20.9%)

day 2, M0

day 6, all groups

3 2 1

day 2, scaffolds

rsif.royalsocietypublishing.org

4

5

M0 control, day 2 TCP–HA, day 2 baghdadite, day 2 Sr–HT–gahnite, day 2 M0 control, day 6 TCP–HA, day 6 baghdadite, day 6 Sr–HT–gahnite, day 6

5

0 –1

–3 –3

–2

–1 0 PC1 (25.6%)

1

2

3

4 (c)

4 3

M0 control baghdadite

day 6, M0 12 day 6 baghdadite

day 6 1 TCP–HA

M1/M2 score

PC4 (9.7%)

2 day 2, baghdadite

0 –1 –2 –3 –3

TCP–HA Sr–HT–gahnite

9 6

****

(b)

–4

***

–4 –5

3 0 day 2

–2

–1

0

1

2 3 PC3 (14.3%)

4

5

6

day 6

7

Figure 3. Multivariate analysis of gene expression data. (a,b) PCA of the temporal effects of baghdadite, Sr– HT – gahnite and TCP – HA scaffolds on macrophage gene expression. (a) Score plot of PC1 and 2 depicted, capturing 25.6% and 20.9% of the variance within the data, respectively. Solid line encloses all scaffoldactivated macrophages on day 2; shaded region highlights unactivated M0 control macrophages on day 2; dashed line captures all day 6 responses. (b) Score plot of PC3 and 4, capturing 14.3% and 9.7% of the variance, respectively. Solid line encloses baghdadite-activated macrophages on day 2; dashed lines enclose day 6 macrophage responses from baghdadite (red), TCP – HA (blue) and the M0 control (black). (c) M1/M2 scoring of scaffold-activated macrophages over time. Data represent mean score + s.e.m. Statistical analysis completed using two-way ANOVA with Tukey’s post hoc analysis. ***p , 0.001, ****p , 0.0001; n ¼ 4. Differences between day 2 and 6 were compared for each scaffold using a multiple t-test analysis via the Holm – Sidak method; baghdadite scaffolds were significantly different ( p , 0.01). (Online version in colour.)

revealed scaffold type as the third largest source of variation in the data (figure 3b). More specifically, PC3 separated macrophage interactions with baghdadite scaffolds from all other groups at both time points. PC3 also separated macrophage interactions with TCP–HA scaffolds from all other groups on day 6. The genes most contributing to the variance of PC3 were CD163, CCR7, VEGF and SPP1; genes driving the variance of PC4 included CCL22, ALPL and TIMP3 (electronic supplementary material, figure S1). Importantly, these results show that baghdadite scaffolds elicited different macrophage behaviour from Sr–HT–gahnite, and both scaffolds were different in their responses compared with TCP–HA scaffolds. Intriguingly, baghdadite scaffolds resulted in a significant reduction in the M1/M2a score over time (figure 3c), which may suggest that these scaffolds promote a phenotypic transition in macrophage behaviour. In contrast, TCP–HA scaffolds appeared to promote prolonged pro-inflammatory activation with an increasing M1/M2a score between days 2 and 6, which was significantly higher than all other groups. However, given that PCA and combinatorial scores do not allow thorough investigation into the contribution of each gene to the results, we then evaluated each gene individually

to more thoroughly characterize the response of macrophages to the scaffolds.

3.2.2. Temporal changes in gene expression Analysis of expression levels of individual genes revealed interesting differences between macrophages cultured on the different scaffolds. Compared with the M0 control, TCP–HA scaffolds caused upregulation of the pro-inflammatory M1 markers TNF at day 2, and VEGF and IL1B at day 6 (figure 4). CCR7 was downregulated at day 6 by macrophages on TCP–HA scaffolds. Baghdadite caused upregulation of TNF at day 2 and of IL1B at day 6, with downregulation of VEGF at day 2 relative to the M0 control. Sr–HT–gahnite caused upregulation of TNF at day 2 and downregulation of CCR7 at day 6. Differences in gene expression of M1 markers between scaffolds were also noted, with baghdadite promoting higher expression of CCR7 and TNF than TCP–HA at day 2, which then returned to baseline by day 6. TCP–HA scaffolds promoted a significant downregulation of all M2a markers at day 6 (figure 5), which was likely to be the major contributing factor to the increasing M1/M2a score

J. R. Soc. Interface 13: 20160346

–2

M1 markers

6 TCP–HA 1

CCR7

10–1

IL1B

10–1

**

#

10–2

10–3

#

#

#

10–3 day 2

day 6

day 2 1

TNF

1 10–1

** #

10–2

**

**

10–2

10–4

fold change over GAPDH

Sr–HT–gahnite

J. R. Soc. Interface 13: 20160346

fold change over GAPDH

1

baghdadite

#

day 6 VEGF

10–1 #

**** #

**** **** #

10–2

10–3

10–3

10–4 day 2

day 2

day 6

day 6

Figure 4. Gene expression of macrophages exposed to baghdadite, Sr–HT–gahnite and TCP–HA scaffolds over 6 days, based on markers indicative of the M1 phenotype. Data represent mean fold change over GAPDH + s.e.m. Statistical analysis was performed on log-transformed data using two-way ANOVA with Tukey’s post hoc analysis. Hash symbol denotes significance ( p , 0.01) relative to the M0 control. **p , 0.01, ***p , 0.001, ****p , 0.0001; n ¼ 4. (Online version in colour.)

M2a markers M0 control

fold change over GAPDH

10

TCP–HA

Sr–HT–gahnite

1

CCL22

MRC1

10–1 1

10–1

**** #

**** ****

**** #

# #

#

10–2 10–3 #

#

10–4 10–5

10–2 day 2

day 2

day 6 1

PDGFB

1 fold change over GAPDH

baghdadite

10–1

10–1

10–2

** ***

10–3

**** **** **** #

10–2

#

10–4

day 6 TIMP3

**** **** #

#

day 2

day 6

10–3 day 2

day 6

rsif.royalsocietypublishing.org

M0 control

Figure 5. Gene expression of macrophages exposed to baghdadite, Sr–HT–gahnite and TCP–HA scaffolds over 6 days, based on markers indicative of the M2a phenotype. Data represent mean fold change over GAPDH + s.e.m. Statistical analysis was performed on log-transformed data using two-way ANOVA with Tukey’s post hoc analysis. Hash symbol denotes significance ( p , 0.01) relative to the M0 control. **p , 0.01, ***p , 0.001, ****p , 0.0001; n ¼ 4. (Online version in colour.)

M0 control CD163

1

baghdadite 1

MARCO

1

Sr–HT–gahnite MMP7

**** **** #

**

*** *** #

10–1

10–1

#

10–2

#

10–2

10–3

10–3 day 6

day 2

** **** ***

1

#

day 6 SPP1

day 2

day 6 VCAN

1

** **** **

10–1 1

***

#

#

10–2 #

10–1

10–3

10–1 day 2

day 6

day 2

day 6

day 2

day 6

Figure 6. Gene expression of macrophages exposed to baghdadite, Sr – HT – gahnite and TCP – HA scaffolds over 6 days, based on markers indicative of the M2c phenotype. Data represent mean fold change over GAPDH + s.e.m. Statistical analysis was performed on log-transformed data using two-way ANOVA with Tukey’s post hoc analysis. Hash symbol denotes significance ( p , 0.01) relative to the M0 control. **p , 0.01, ***p , 0.001, ****p , 0.0001; n ¼ 4. (Online version in colour.)

over time. All scaffolds promoted downregulation of the M2a marker CCL22 compared with the M0 control at both days 2 and 6, with the exception of Sr–HT–gahnite scaffolds at day 2. Gene expression of M2a markers PDGFBB and TIMP3 was unaffected by baghdadite scaffolds at either time point, but MRC1 was downregulated at day 6 relative to the M0 control. Sr–HT–gahnite scaffolds promoted downregulation of CCL22, MRC1 and TIMP3 at day 6. With respect to differences between scaffolds, macrophages cultured on TCP–HA scaffolds expressed lower levels of three of the four M2a markers than macrophages on baghdadite or Sr–HT–gahnite scaffolds at both time points. In addition to promoting downregulation of many M2a markers, TCP–HA scaffolds also promoted downregulation of several M2c markers at day 6 (CD163, MMP7, MMP9 and SPP1) relative to the M0 control (figure 6). In contrast, baghdadite scaffolds promoted upregulation of two M2c markers at day 6 (CD163 and VCAN), as well as downregulation of MARCO at day 2 and MMP7 at day 6, although to a lesser extent than TCP–HA. Sr–HT–gahnite scaffolds promoted downregulation of MMP7 and MMP9 at day 6, also to a lesser extent than TCP–HA scaffolds. In general, macrophages cultured on baghdadite scaffolds expressed higher levels of the M2c genes than macrophages on the other scaffolds, especially at the later time point. Taken together, these results suggest that TCP –HA scaffolds promote the highest levels of M1-related gene expression, whereas baghdadite scaffolds promote a shift towards M2c-related gene expression, especially at the later time point. On the other hand, the only genes that showed different trends in terms of direction of expression for baghdadite and TCP– HA, relative to the M0 control, were CCR7 and CD163, and hybrid activation states that were not distinctly M1 or M2 were observed for all scaffolds.

7

Further extending our analysis to genes more generally related to bone repair revealed only minor changes in angiogenic markers (TGFB1 and TIE1), and did not provide any indication of osteoblast- or osteoclast-like behaviour (ALPL, BGLAP, RUNX2 and TNFSF11) induced by TCP–HA, baghdadite or Sr–HT–gahnite scaffolds (electronic supplementary material, figure S2).

3.3. Effects of soluble factors on macrophage response 3.3.1. Principal component analysis Baghdadite, Sr–HT–gahnite and TCP–HA scaffolds have different mechanical properties, microstructure and ion dissolution profiles. As an initial step towards investigating why macrophages behaved differently on each scaffold, we tested the effects of released soluble factors on macrophage activation by culturing the macrophages and scaffolds in separate chambers of a transwell culture system. Again, we used PCA to visualize the gene expression data, enabling us to identify patterns and emphasize variation among the scaffolds. Interestingly, PC1, PC2 and PC3 failed to separate macrophages cultured in the transwell system with any of the scaffolds (figure 7a,b). These three principal components captured 88.1% of the variation in the data, suggesting that soluble signals released from the scaffolds do not account for differences in gene expression of macrophages.

3.3.2. Gene expression Individual gene analysis revealed minor changes in gene expression of CCL22, IL1B and MMP7 by macrophages cultured in the transwell system with the scaffolds (figure 8 and electronic supplementary material, figure S3), in agreement with the principal component analysis results.

J. R. Soc. Interface 13: 20160346

10

MMP9

10

#

10–1

#

10–2

day 2 fold change over GAPDH

TCP–HA

rsif.royalsocietypublishing.org

fold change over GAPDH

M2c markers

6 5

PC2 (20.9%)

4

direct, baghdadite transwell, all scaffolds

3 2 1 M0

0

–3

–2

–1

0

1

2

3

4

5

PC3 (9.2%)

(b) 3.0 direct, TCP–HA 2.5 transwell, all scaffolds 2.0 1.5 1.0 direct, gahnite 0.5 M0 direct, 0 baghdadite –0.5 –1.0 –1.5 –2.0 –4 –3 –2 –1 0 1 2 3 4 5 PC1 (58.0%)

Figure 7. PCA of direct and indirect (transwell) interactions on macrophage gene expression. (a) Score plot of PC1 and 2 depicted, capturing 58.0% and 20.9% of the variance within the data, respectively. (b) Score plot of PC1 and 3, which captured 9.2% of the variance. Shaded regions highlight direct effects of baghdadite (red), TCP – HA (blue), Sr – HT – gahnite (green) and unactivated M0 control (black). Solid line indicates clustering of all transwell effects. (Online version in colour.)

3.4. Grain size effects on macrophage response Considering released factors had only minor effects on macrophage gene expression, we next evaluated if changing topography would affect macrophage behaviour by varying the grain size within baghdadite scaffolds. We also used protein secretion analysis instead of gene expression analysis to confirm that macrophages secrete the protein products of these genes at appreciable levels. While all proteins evaluated were secreted at the expected levels based on previous reports [20,21], varying baghdadite grain size between 2 and 4 mm had only a modest effect on macrophage behaviour (electronic supplementary material, figure S4). For example, macrophages cultured on larger grain baghdadite scaffolds secreted slightly higher levels of TNF at day 2 ( p , 0.05).

4. Discussion In this study, we aimed to explore in vitro the response of macrophages to bone substitutes that have been used with varying degrees of success to repair critical-sized bone defects in vivo. Overall, our findings suggest that different scaffold chemistries differentially influence macrophage activation in vitro, which we have also shown for materials used in chronic wound care [33]. Changing activation states of macrophages in response to implanted materials would be expected to have profound effects on bone formation given the importance of macrophages for bone repair [19]. While all scaffolds promoted upregulation of certain M1related genes, TCP– HA caused downregulation of M2a- and M2c-related genes at both early and late time points, whereas

baghdadite scaffolds promoted higher levels of expression of M2c-related genes, especially at the later time point. While Sr –HT–gahnite scaffolds did not have major effects on macrophage behaviour relative to M0 controls, they did promote higher levels of expression of several M2a- and M2c-related genes relative to TCP –HA scaffolds. However, it is important to note that the macrophage phenotypes observed in this work in response to baghdadite, Sr–HT – gahnite and TCP–HA scaffolds were hybrid activation states that were not distinctly M1 or M2. Moreover, the fact that many genes were regulated in the same direction and to similar extents suggests that more studies are needed to confirm if these changes correspond to functional differences. Nevertheless, the potential of these scaffolds to differentially regulate macrophage behaviour may have important implications for harnessing the natural healing ability of bone. Promoting proper vascularization following injury is a major challenge of tissue engineering strategies for fracture healing. Considering that the inflammatory response plays an important role in stimulating angiogenesis and healing, and previous work has demonstrated that sequential M1 and M2 macrophage activation is required to support these processes [19,20], scaffolds that promote early M1- and late M2-like activation would also be expected to promote bone repair. Indeed, baghdadite and Sr–HT–gahnite scaffolds have been previously shown to be more successful in promoting bone regeneration in a critical-sized bone defect model in rabbits [24,25]. Interestingly, baghdadite scaffolds promoted upregulation of many genes related to the M2c phenotype compared with the M0 control. While the role of M2c macrophages in tissue regeneration is still poorly understood, recent studies

J. R. Soc. Interface 13: 20160346

–2 –4

direct, TCP–HA

direct, gahnite

–1

8

rsif.royalsocietypublishing.org

(a)

M0 control, direct TCP–HA, direct baghdadite, direct Sr–HT–gahnite, direct M0 control, transwell TCP–HA, transwell baghdadite, transwell Sr–HT–gahnite, transwell

fold change over GAPDH fold change over GAPDH

10–1

#

CD163 #

#

10–1

10–1 # #

direct

transwell

CD80

1

10–2

#

direct

transwell

MRC1

1

10–1

10–1

10–2

**

**

direct

transwell

VCAN #

# #

10–2 10–2

10–3 direct

transwell

IL1B

1

10–2

1

**** ***

10–1

10–3

10–1

10–2

** **

10–3 direct

10–3 direct

transwell

10–2 10–3

transwell TIMP3

1

10–1

10–4

# #

10–3

10–4

1

** #

10–2

10–4

CCL22

1

M2c markers #

** *** #

direct

#

transwell

10–1

**** **

10–2

#

10–3

transwell

MMP7

1

#

direct

10–4

direct

# #

transwell

VEGF

1

M0 control 10–1

TCP–HA **

baghdadite

#

10–2

Sr–HT–gahnite

10–3 direct

transwell

Figure 8. Effects of soluble factors on macrophage gene expression. Data represent mean fold change over GAPDH + s.e.m. Statistical analysis was performed on log-transformed data using one-way ANOVA with Tukey’s post hoc analysis. Hash symbol denotes significance ( p , 0.01) relative to the M0 control. **p , 0.01, ***p , 0.001, ****p , 0.0001; n  4. (Online version in colour.) suggest that they are involved in tissue remodelling [22], which plays an important role in bone repair. The fracturehealing cascade involves the formation of a cartilaginous fracture callus, mineralization and resorption of the callus, and production of woven bone as the fracture site is revascularized, which is later remodelled through a coordinated process of bone resorption and formation to restore the native tissue. Several studies have shown that the absence of the tissue-remodelling enzyme, MMP9, following injury delays endochondral ossification and alters healing [37,38]. Similarly, the absence of osteopontin, encoded by SPP1 and believed to facilitate the uptake of mineralized matrix [39], causes altered tissue remodelling in mice and, consequently, reduced biomechanical properties [40]. Regulation of macrophage phenotype observed in this study was attributed predominantly to direct interactions between macrophages and the scaffolds, as only modest effects of scaffold-released soluble factors on macrophage activation were

observed. Macrophage activation may be caused by a number of factors, including ion dissolution, surface topography and mechanical properties. The immunomodulatory effects of metal ions such as titanium, zinc, zirconium and strontium are now widely appreciated, resulting in a paradigm shift away from inert bone substitutes [41]. In one study, conditioning macrophages in b-tricalcium phosphate extracts increased the osteogenic differentiation of mesenchymal stem cells, supporting the role of macrophages in biomaterial-induced osteogenesis [42]. Although the mechanisms behind metal ion-induced bone regeneration are not well understood (reviewed in [43]), the presence of ions probably affects macrophages. Zirconia (in baghdadite) has been shown to be pro-inflammatory [44], whereas strontium and zinc (in Sr–HT–gahnite) have been shown to affect bone remodelling [45,46] and promote anti-inflammatory activity [47]. Future studies are focused on determining whether the presence of these ions contributed to the differential activation of macrophages observed in this work.

9

J. R. Soc. Interface 13: 20160346

fold change over GAPDH

CCR7

1

M2a markers

rsif.royalsocietypublishing.org

fold change over GAPDH

M1 markers

5. Conclusion In this work, we investigated the effects of pro-regenerative scaffolds on macrophage behaviour. Overall, our findings suggest that macrophage activation is influenced by scaffold properties. To this end, we have shown that baghdadite and Sr–HT–gahnite scaffolds regulate macrophage phenotype differently from TCP–HA scaffolds, which caused upregulation of inflammatory (M1) genes and downregulation of M2 genes (both M2a and M2c), though hybrid activation states were observed. While additional work is needed to confirm these findings on a functional level, the ability of baghdadite and Sr–HT–gahnite scaffolds to modulate macrophage behaviour would be expected to have important implications for bone regeneration. In addition, we have demonstrated that the ability of these scaffolds to regulate macrophage behaviour is attributed to a combination of direct and indirect cell– scaffold interactions, with direct interactions having the dominant effects. Ultimately, improved understanding of the interactions between scaffolds that regenerate bone in vivo and cells of the inflammatory response will aid in the design of biomaterials to facilitate bone repair and tissue regeneration. Authors’ contributions. P.L.G. designed and conducted experiments and prepared the manuscript; S.-I.R.-E. synthesized and characterized the scaffolds, and helped draft the manuscript; H.Z. and K.L.S. conceived of and coordinated the study and revised the manuscript; K.L.S. assisted in data interpretation and provided critical revisions of the manuscript. All authors gave final approval for publication.

Competing interests. Sr –HT– gahnite is the subject of a global licensing agreement with Allegra Orthopaedics.

Funding. The authors gratefully acknowledge the Australian NHMRC, ARC and the Rebecca Cooper Medical Foundation, as well as the United States NSF-CNIC support of this work (award no. 1425737). P.L.G. is also the recipient of a Koerner Family Fellowship.

Acknowledgements. The authors acknowledge Dr Chandrasekaran Nagaswami and Dr Sylvain Le Marchand for their technical assistance.

References 1.

2.

Cheung C. 2005 The future of bone healing. Clin. Podiatric Med. Surg. 22, 631–641. (doi:10.1016/j. cpm.2005.07.002) Wu AC, Raggatt LJ, Alexander KA, Pettit AR. 2013 Unraveling macrophage contributions to bone repair. BoneKey Rep. 2, 373. (doi:10.1038/bonekey.2013.107)

3.

4.

Carano RA, Filvaroff EH. 2003 Angiogenesis and bone repair. Drug Discov. Today 8, 980–989. (doi:10.1016/S1359-6446(03)02866-6) Hausman MR, Rinker BD. 2004 Intractable wounds and infections: the role of impaired vascularity and advanced surgical methods for treatment.

5.

Am. J. Surg. 187(Suppl.1), S44– S55. (doi:10.1016/ S0002-9610(03)00304-0) Karnes JM, Daffner SD, Watkins CM. 2015 Multiple roles of tumor necrosis factor-alpha in fracture healing. Bone 78, 87 –93. (doi:10.1016/j.bone. 2015.05.001)

10

J. R. Soc. Interface 13: 20160346

to be an excellent indicator of cell identity and physiological state [55], especially for discerning macrophage activation [56–58], additional work is needed to confirm phenotypic changes in macrophages on a functional level. Finally, we investigated only the effects of these scaffolds on macrophages, but other immune cells, including dendritic cells and resident tissue macrophages, as well as cells specific to bone repair, would be expected to interact with macrophages and have major effects on bone regeneration. In future studies, we plan to explore the interactions between macrophages, activated by bone substitutes, and other cells involved in the fracturehealing cascade. Despite these limitations, this study suggests that scaffold properties differentially activate macrophages, and that, with further investigation, it may be possible to proactively modulate macrophage behaviour using scaffold design.

rsif.royalsocietypublishing.org

Macrophage activation may also be affected by surface topography. Previously, it was shown that macrophages accumulated on rough but not smooth titanium surfaces, though both led to bone formation in a rat subcutaneous model, suggesting differences in the mechanism of healing [48]. Other studies have also demonstrated the role of microstructure and surface topography in promoting anti-inflammatory behaviour in human macrophages [49 –51]. In this work, we did find some effects of changing grain size on the activation of macrophages on baghdadite scaffolds, though the effects were modest. However, these findings are limited to the small region of grain sizes tested in this study and additional work is necessary to draw conclusions regarding the influence of grain size on macrophage –baghdadite interactions. It is also possible that grain size plays a role in macrophage responses to other materials not investigated in this study. In future work, the effects of surface topography on M1, M2a and M2c activation will be explored in greater detail. Lastly, macrophage activation could also be caused by mechanical properties. To date, the effect of scaffold stiffness on macrophage phenotype has been relatively underexplored and remains unclear. Irwin et al. [52] attempted to shed light on the ambiguous relationship between inflammatory cells and substrate modulus using the THP-1 cell line, a human monocytic cell line that can be differentiated into macrophages. Although cells attached preferentially to stiffer substrates, secretion of pro-inflammatory cytokines was variable. In contrast, another study found that stiffness of poly(ethylene glycol) (PEG) substrates did not affect in vitro attachment of RAW 264.7 macrophages; however, stiffer substrates promoted pro-inflammatory gene expression in primary murine macrophages when stimulated with lipopolysaccharide [53]. More recently, Guo et al. [54] showed that the modulus of poly(ester urethane) scaffolds implanted subcutaneously in rats can be tailored to modulate macrophage phenotype. Given the potential for the ceramic scaffolds investigated in this study to modulate macrophage behaviour, future studies to specifically isolate substrate modulus would be interesting. There were several limitations to this study. Because of the complexity of these scaffolds, we were not able to isolate any one variable that contributes to modulation of macrophage behaviour. Our goal was to explore the response of macrophages to different scaffolds in order to identify the most promising areas for more detailed investigation in future studies. Thus, subsequent work will focus on carefully defining how changing scaffold ion content, topography and mechanical properties affects macrophage activation, which could have major effects on the design of scaffolds for bone regeneration. Another major limitation is that we evaluated only a small subset of the thousands of genes involved in bone regeneration. Moreover, although gene expression has been shown

6.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

muscle. Tissue Eng. A 16, 115– 125. (doi:10.1089/ ten.TEA.2009.0130) Quinlan E, Partap S, Azevedo MM, Jell G, Stevens MM, O’Brien FJ. 2015 Hypoxia-mimicking bioactive glass/collagen glycosaminoglycan composite scaffolds to enhance angiogenesis and bone repair. Biomaterials 52, 358 –366. (doi:10.1016/j. biomaterials.2015.02.006) Witherel CE, Graney PL, Freytes DO, Weingarten MS, Spiller KL. 2016 Response of human macrophages to wound matrices in vitro. Wound Repair Regen. 24, 515 –524. (doi:10.1111/wrr.12423) Ringner M. 2008 What is principal component analysis? Nat. Biotechnol. 26, 303 –304. (doi:10. 1038/nbt0308-303) Nassiri S, Zakeri I, Weingarten MS, Spiller KL. 2015 Relative expression of proinflammatory and antiinflammatory genes reveals differences between healing and nonhealing human chronic diabetic foot ulcers. J. Invest. Dermatol. 135, 1700–1703. (doi:10.1038/jid.2015.30) Roohani-Esfahani SI, Chen Y, Shi J, Zreiqat H. 2013 Fabrication and characterization of a new, strong, bioactive ceramic scaffold for bone regeneration. Mater. Lett. 107, 378–381. (doi:10.1016/j.matlet. 2013.06.046) Colnot C, Thompson Z, Miclau T, Werb Z, Helms JA. 2003 Altered fracture repair in the absence of MMP9. Development 130, 4123–4133. (doi:10. 1242/dev.00559) McDonald MM, Morse A, Mikulec K, Peacock L, Baldock PA, Kostenuik PJ, Little DG. 2013 Matrix metalloproteinase-driven endochondral fracture union proceeds independently of osteoclast activity. J. Bone Miner. Res. 28, 1550–1560. (doi:10.1002/ jbmr.1889) Pedraza CE, Nikolcheva LG, Kaartinen MT, Barralet JE, McKee MD. 2008 Osteopontin functions as an opsonin and facilitates phagocytosis by macrophages of hydroxyapatite-coated microspheres: implications for bone wound healing. Bone 43, 708–716. (doi:10. 1016/j.bone.2008.06.010) Duvall CL, Taylor WR, Weiss D, Wojtowicz AM, Guldberg RE. 2007 Impaired angiogenesis, early callus formation, and late stage remodeling in fracture healing of osteopontin-deficient mice. J. Bone Miner. Res. 22, 286 –297. (doi:10.1359/ jbmr.061103) Mohammadi H, Hafezi M, Nezafati N, Heasarki S, Nadernezhad A, Ghazanfari SMH, Sepantafar M. 2014 Bioinorganics in bioactive calcium silicate ceramics for bone tissue repair: bioactivity and biological properties. J. Ceram. Sci. Tech. 5, 1–12. Chen Z, Wu C, Gu W, Klein T, Crawford R, Xiao Y. 2014 Osteogenic differentiation of bone marrow MSCs by b-tricalcium phosphate stimulating macrophages via BMP2 signalling pathway. Biomaterials 35, 1507–1518. (doi:10.1016/j. biomaterials.2013.11.014) Lakhkar NJ, Lee IH, Kim HW, Salih V, Wall IB, Knowles JC. 2013 Bone formation controlled by biologically relevant inorganic ions: role and controlled delivery from phosphate-based glasses.

11

J. R. Soc. Interface 13: 20160346

8.

19. Schlundt C et al. 2015 Macrophages in bone fracture healing: their essential role in endochondral ossification. Bone. (doi:10.1016/j.bone.2015.10.019) 20. Spiller KL, Anfang RR, Spiller KJ, Ng J, Nakazawa KR, Daulton JW, Vunjak-Novakovic G. 2014 The role of macrophage phenotype in vascularization of tissue engineering scaffolds. Biomaterials 35, 4477–4488. (doi:10.1016/j.biomaterials.2014.02.012) 21. Spiller KL, Nassiri S, Witherel CE, Anfang RR, Ng J, Nakazawa KR, Yu T, Vunjak-Novakovic G. 2015 Sequential delivery of immunomodulatory cytokines to facilitate the M1-to-M2 transition of macrophages and enhance vascularization of bone scaffolds. Biomaterials 37, 194– 207. (doi:10.1016/j. biomaterials.2014.10.017) 22. Spiller KL, Nassiri S, Raman P, Rajagopalan R, Sarmandy M, Weingarten MS. 2015 Discovery of a novel M2c macrophage gene expression signature indicates a major role in human wound healing. Wound Repair Regen. 23, A40. 23. Lavin Y, Winter D, Blecher-Gonen R, David E, KerenShaul H, Merad M, Jung S, Amit I. 2014 Tissueresident macrophage enhancer landscapes are shaped by the local microenvironment. Cell 159, 1312 –1326. (doi:10.1016/j.cell.2014.11.018) 24. Roohani-Esfahani SI, Dunstan CR, Davies B, Pearce S, Williams R, Zreiqat H. 2012 Repairing a critical-sized bone defect with highly porous modified and unmodified baghdadite scaffolds. Acta Biomater. 8, 4162 –4172. (doi:10.1016/j.actbio.2012.07.036) 25. Roohani-Esfahani SI, Dunstan CR, Li JJ, Lu Z, Davies B, Pearce S, Field J, Williams R, Zreiqat H. 2013 Unique microstructural design of ceramic scaffolds for bone regeneration under load. Acta Biomater. 9, 7014 –7024. (doi:10.1016/j.actbio.2013.02.039) 26. Roohani-Esfahani SI, Newman P, Zreiqat H. 2016 Design and fabrication of 3D printed scaffolds with a mechanical strength comparable to cortical bone to repair large bone defects. Sci. Rep. 6, 19468. (doi:10.1038/srep19468) 27. Li JJ, Roohani-Esfahani SI, Dunstan CR, Quach T, Steck R, Saifzadeh S, Pivonka P, Zreiqat H. 2016 Efficacy of novel synthetic bone substitutes in the reconstruction of large segmental bone defects in sheep tibiae. Biomed. Mater. 11, 015016. (doi:10. 1088/1748-6041/11/1/015016) 28. Lu Z, Wang G, Roohani-Esfahani I, Dunstan CR, Zreiqat H. 2014 Baghdadite ceramics modulate the cross talk between human adipose stem cells and osteoblasts for bone regeneration. Tissue Eng. A 20, 992 –1002. (doi:10.1089/ten.TEA.2013.0470) 29. Ramaswamy Y, Wu C, Van Hummel A, Combes V, Grau G, Zreiqat H. 2008 The responses of osteoblasts, osteoclasts and endothelial cells to zirconium modified calcium-silicate-based ceramic. Biomaterials 29, 4392–4402. (doi:10.1016/j. biomaterials.2008.08.006) 30. Kokubo T, Takadama H. 2006 How useful is SBF in predicting in vivo bone bioactivity? Biomaterials 27, 2907–2915. (doi:10.1016/j.biomaterials.2006.01.017) 31. Lesman A, Habib M, Caspi O, Gepstein A, Arbel G, Levenberg S, Gepstein L. 2010 Transplantation of a tissue-engineered human vascularized cardiac

rsif.royalsocietypublishing.org

7.

Mountziaris PM, Spicer PP, Kasper FK, Mikos AG. 2011 Harnessing and modulating inflammation in strategies for bone regeneration. Tissue Eng. B, Rev. 17, 393–402. (doi:10.1089/ten.TEB. 2011.0182) Santo VE, Gomes ME, Mano JF, Reis RL. 2013 Controlled release strategies for bone, cartilage, and osteochondral engineering—part I: recapitulation of native tissue healing and variables for the design of delivery systems. Tissue Eng. B, Rev. 19, 308–326. (doi:10.1089/ten.TEB.2012.0138) Schindeler A, McDonald MM, Bokko P, Little DG. 2008 Bone remodeling during fracture repair: the cellular picture. Semin. Cell Dev. Biol. 19, 459–466. (doi:10.1016/j.semcdb.2008.07.004) Kon T, Cho TJ, Aizawa T, Yamazaki M, Nooh N, Graves D, Gerstenfeld LC, Einhorn TA. 2001 Expression of osteoprotegerin, receptor activator of NF-kB ligand (osteoprotegerin ligand) and related proinflammatory cytokines during fracture healing. J. Bone Miner. Res. 16, 1004 –1014. (doi:10.1359/ jbmr.2001.16.6.1004) Grundnes O, Reikeras O. 1993 The importance of the hematoma for fracture healing in rats. Acta Orthop. Scand. 64, 340– 342. (doi:10.3109/1745367 9308993640) Park SH, Silva M, Bahk WJ, McKellop H, Lieberman JR. 2002 Effect of repeated irrigation and debridement on fracture healing in an animal model. J. Orthop. Res. 20, 1197 –1204. (doi:10. 1016/S0736-0266(02)00072-4) Schmidt-Bleek K, Schell H, Schulz N, Hoff P, Perka C, Buttgereit F, Volk HD, Lienau J, Duda GN. 2012 Inflammatory phase of bone healing initiates the regenerative healing cascade. Cell Tissue Res. 347, 567 – 573. (doi:10.1007/s00441-011-1205-7) Sindrilaru A et al. 2011 An unrestrained proinflammatory M1 macrophage population induced by iron impairs wound healing in humans and mice. J. Clin. Invest. 121, 985–997. (doi:10. 1172/JCI44490) Murray PJ, Wynn TA. 2011 Protective and pathogenic functions of macrophage subsets. Nat. Rev. Immunol. 11, 723– 737. (doi:10.1038/nri3073) Raggatt LJ et al. 2014 Fracture healing via periosteal callus formation requires macrophages for both initiation and progression of early endochondral ossification. Am. J. Pathol. 184, 3192–3204. (doi:10.1016/j.ajpath.2014.08.017) Alexander KA et al. 2011 Osteal macrophages promote in vivo intramembranous bone healing in a mouse tibial injury model. J. Bone Miner. Res. 26, 1517–1532. (doi:10.1002/jbmr.354) Vi L, Baht GS, Whetstone H, Ng A, Wei Q, Poon R, Mylvaganam S, Grynpas M, Alman BA. 2015 Macrophages promote osteoblastic differentiation in vivo: implications in fracture repair and bone homeostasis. J. Bone Miner. Res. 30, 1090 –1102. (doi:10.1002/jbmr.2422) Brown BN, Sicari BM, Badylak SF. 2014 Rethinking regenerative medicine: a macrophage-centered approach. Front. Immunol. 5, 510. (doi:10.3389/ fimmu.2014.00510)

45.

47.

48.

54. Guo R, Merkel AR, Sterling JA, Davidson JM, Guelcher SA. 2015 Substrate modulus of 3D-printed scaffolds regulates the regenerative response in subcutaneous implants through the macrophage phenotype and Wnt signaling. Biomaterials 73, 85 –95. (doi:10.1016/j.bio materials.2015.09.005) 55. Duan H et al. 2015 Transcriptome analyses reveal molecular mechanisms underlying functional recovery after spinal cord injury. Proc. Natl Acad. Sci. USA 112, 13 360–13 365. (doi:10.1073/pnas. 1510176112) 56. Becker M, De Bastiani MA, Parisi MM, Guma FT, Markoski MM, Castro MA, Kaplan MH, Barbe-Tuana FM, Klamt F. 2015 Integrated transcriptomics establish macrophage polarization signatures and have potential applications for clinical health and disease. Sci. Rep. 5, 13351. (doi:10.1038/srep13351) 57. Martinez FO, Gordon S, Locati M, Mantovani A. 2006 Transcriptional profiling of the human monocyte-to-macrophage differentiation and polarization: new molecules and patterns of gene expression. J. Immunol. 177, 7303–7311. (doi:10. 4049/jimmunol.177.10.7303) 58. Xue J et al. 2014 Transcriptome-based network analysis reveals a spectrum model of human macrophage activation. Immunity 40, 274–288. (doi:10.1016/j.immuni.2014.01.006)

12

J. R. Soc. Interface 13: 20160346

46.

49. Bartneck M, Schulte VA, Paul NE, Diez M, Lensen MC, Zwadlo-Klarwasser G. 2010 Induction of specific macrophage subtypes by defined micro-patterned structures. Acta Biomater. 6, 3864–3872. (doi:10. 1016/j.actbio.2010.04.025) 50. Brodbeck WG, Nakayama Y, Matsuda T, Colton E, Ziats NP, Anderson JM. 2002 Biomaterial surface chemistry dictates adherent monocyte/macrophage cytokine expression in vitro. Cytokine 18, 311– 319. (doi:10.1006/cyto.2002.1048) 51. Paul NE, Skazik C, Harwardt M, Bartneck M, Denecke B, Klee D, Salber J, Zwadlo-Klarwasser G. 2008 Topographical control of human macrophages by a regularly microstructured polyvinylidene fluoride surface. Biomaterials 29, 4056 – 4064. (doi:10.1016/j.biomaterials.2008. 07.010) 52. Irwin EF, Saha K, Rosenbluth M, Gamble LJ, Castner DG, Healy KE. 2008 Modulus-dependent macrophage adhesion and behavior. J. Biomater. Sci. Polym. Ed. 19, 1363– 1382. (doi:10.1163/ 156856208786052407) 53. Blakney AK, Swartzlander MD, Bryant SJ. 2012 The effects of substrate stiffness on the in vitro activation of macrophages and in vivo host response to poly(ethylene glycol)-based hydrogels. J. Biomed. Mater. Res. A 100, 1375–1386. (doi:10.1002/jbm.a.34104)

rsif.royalsocietypublishing.org

44.

Adv. Drug Deliv. Rev. 65, 405–420. (doi:10.1016/j. addr.2012.05.015) Obando-Pereda GA, Fischer L, Stach-Machado DR. 2014 Titanium and zirconia particle-induced proinflammatory gene expression in cultured macrophages and osteolysis, inflammatory hyperalgesia and edema in vivo. Life Sci. 97, 96 –106. (doi:10.1016/j.lfs.2013.11.008) Gentleman E, Fredholm YC, Jell G, Lotfibakhshaiesh N, O’Donnell MD, Hill RG, Stevens MM. 2010 The effects of strontium-substituted bioactive glasses on osteoblasts and osteoclasts in vitro. Biomaterials 31, 3949– 3956. (doi:10.1016/j. biomaterials.2010.01.121) Zhang W et al. 2013 The synergistic effect of hierarchical micro/nano-topography and bioactive ions for enhanced osseointegration. Biomaterials 34, 3184– 3195. (doi:10.1016/j.biomaterials.2013.01.008) Bao B, Prasad AS, Beck FW, Godmere M. 2003 Zinc modulates mRNA levels of cytokines. Am. J. Physiol. Endocrinol. Metab. 285, E1095–E1102. (doi:10. 1152/ajpendo.00545.2002) Chehroudi B, Ghrebi S, Murakami H, Waterfield JD, Owen G, Brunette DM. 2010 Bone formation on rough, but not polished, subcutaneously implanted Ti surfaces is preceded by macrophage accumulation. J. Biomed. Mater. Res. A 93, 724–737. (doi:10.1002/jbm.a.32587)

In vitro response of macrophages to ceramic scaffolds used for bone regeneration.

Macrophages, the primary cells of the inflammatory response, are major regulators of healing, and mediate both bone fracture healing and the inflammat...
1MB Sizes 2 Downloads 6 Views