Intra-articular depletion of macrophages increases acute synovitis and alters macrophage polarity in the injured mouse knee

Abstract

Objective:

Acute synovial inflammation following joint trauma is associated with posttraumatic arthritis. Synovial macrophages have been implicated in degenerative changes. In this study, we sought to elucidate the role of intra-articular macrophages in the acute inflammatory response to fracture in the mouse knee.

Method:

A closed articular fracture was induced in two models of synovial macrophage depletion: genetically-modified MaFIA mice administered AP20187 to induce programmed macrophage apoptosis, and wild-type C57BL/6 mice administered clodronate liposomes, both via intra-articular injection. Synovial inflammation, bone morphology, and levels of F4/80+ macrophages, NOS2+ M1 macrophages, and CD206+ M2 macrophages were quantified 7 days after fracture using histology and micro-computed tomography.

Results:

Intra-articular macrophage depletion with joint injury did not reduce acute synovitis or the number of synovial macrophages 7 days after fracture in either macrophage-depleted MaFIA mice or in clodronate-treated C57BL/6 mice. In macrophage-depleted MaFIA mice, macrophage polarity shifted to a dominance of M1 macrophages and a reduction of M2 macrophages in the synovial stroma, indicating a shift in M1/M2 macrophage ratio in the joint following injury. Interestingly, MaFIA mice depleted 2 days prior to fracture demonstrated increased synovitis (P = 0.003), reduced bone mineral density (P = 0.0004), higher levels of M1 macrophages (P = 0.013), and lower levels of M2 macrophages (not statistically significant, P=0.084) compared to control-treated MaFIA mice.

Conclusion:

Our findings indicate that macrophages play a critical immunomodulatory role in the acute inflammatory response surrounding joint injury and suggest that inhibition of macrophage function can have prominent effects on joint inflammation and bone homeostasis after joint trauma.

Introduction

Posttraumatic arthritis (PTA) is a degenerative form of osteoarthritis (OA) that occurs following joint injury and is a major cause of disability nationwide. Of the nearly 27 million Americans living with OA, 12% are estimated to have a posttraumatic etiology^1–3. Despite efforts to understand the mechanisms of PTA development, there are no currently available pharmacologic agents in clinical practice that have been shown to alter its progression.

PTA develops most commonly following fracture of the articular surface of the joint, resulting in acute joint inflammation⁴. The process of inflammation encompasses a complex array of cellular changes that are components of an integrated response⁵. Although mild inflammation after injury can support healing, mounting evidence suggests that inflammation at the time of joint injury may exacerbate the degenerative changes in PTA^6–9. At 7 days after an articular fracture, significant synovitis of the joint capsule is evident histologically both in the mouse¹⁰ and in human patient biopsy samples¹¹, suggesting that trauma to the articular surface may lead to acute pathology in surrounding joint tissues through various mechanisms. C57BL/6 mice, a strain that predictably develops PTA after closed articular fracture, demonstrate increased synovitis, higher levels of inflammatory cytokines such as interleukin-1 (IL-1) and tumor necrosis factor-alpha (TNF-α) within the synovial fluid, and an influx of macrophages to the joint capsule 7 days after fracture^10,12. Conversely, MRL/MpJ mice, a “superhealer” strain that is protected from developing severe PTA, have a reduced inflammatory response following joint injury, suggesting that inflammation at the time of injury may contribute to arthritic changes^13–15.

Macrophages exhibit a spectrum of phenotypes, but can be classified generally into pro-inflammatory M1 macrophages and regulatory M2 macrophages. The pro-inflammatory response of M1 macrophages is a component of host defense, producing inflammatory cytokines and nitric oxide synthase 2 (NOS2)^16,17, whereas the tissue remodeling regulatory M2 macrophages express CD206 and other scavenger receptors, producing interleukin-10 (IL-10) and transforming growth factor beta (TGF-β)^18,19. Specifically for arthritis, inflammation, activated macrophages, and a high M1 to M2 ratio were strongly associated with joint symptoms and radiographic severity of human knee OA^20,21. Synovial macrophages have been implicated in the production of pro-inflammatory cytokines and matrix metalloproteinases (MMPs)^16,22,23, which are highly upregulated following intraarticular fracture¹³. Intra-articular inhibition of IL-1 through the use of IL-1 receptor antagonist (IL-1Ra), delivered either transiently or in a sustained manner at the time of articular fracture in mice, has been shown to attenuate the development of PTA^24,25. Furthermore, systemic depletion of macrophages has been shown to decrease joint swelling in a collagen-induced arthritis model²⁶ and reduce systemic levels of inflammatory cytokines in a hyperlipidemia model of OA²⁷; however, two separate studies of macrophage depletion in high-fat diet mouse models demonstrated an increased systemic inflammatory response^28,29. We initially sought to systemically deplete macrophages with joint fracture using clodronate liposomes; however, the immune response to the trauma of the fracture coupled with the systemic macrophage depletion resulted in a high fatality rate (data not shown). These previous studies indicate a complex role for macrophages in systemic immunoregulation. Understanding the role of acute synovial inflammation and macrophages following articular fracture could help identify novel therapeutic approaches.

In this study, we investigated the effect of a transient synovial depletion of macrophages on the acute inflammatory response following joint injury using two different methods of macrophage depletion. Macrophage Fas-induced Apoptosis (MaFIA) mice are a transgenic mouse strain with Fas-suicide gene on a macrophage promoter that allows for the specific depletion of macrophages using the small molecule AP20187³⁰. By injecting AP20187 directly into the joint space, we locally depleted macrophages in the knee joint surrounding a closed articular fracture. Alternatively, we locally depleted macrophages in C57BL/6 mice with joint injury using an intra-articular injection of clodronate encapsulated in liposomes^31,32. We hypothesized that acute transient intra-articular depletion of macrophages following joint injury would reduce synovitis and alter the inflammatory response. On the contrary, however, we found that macrophage depletion induced intensive synovitis 7 days after fracture, indicating macrophages play a critical role in modulating the acute inflammatory response to joint injury.

Methods

Animals

All animal procedures were performed in accordance with an IACUC-approved protocol at Duke University. Male MaFIA transgenic mice generated on the C57BL/6 background (n = 27) and male C57BL/6 mice (n = 27) were purchased from The Jackson Laboratories. MaFIA mice have mutant human FK-506 binding protein 1A, 12 kDa (FKBP12)-Fas suicide construct and enhanced green fluorescent protein (eGFP) placed under the mouse colony stimulating factor 1 receptor promoter (Csf1r), which allows for inducible and reversible apoptosis of macrophages and dendritic cells in the presence of the dimerization drug AP20187 and fluorescent visualization of Csf1r cells³⁰. The AP20187 molecule preferentially binds to mutant human FKBP on Csf1r cells and initiates apoptosis through the Fas-suicide pathway.

Articular fracture and macrophage depletion

MaFIA mice and C57BL/6 mice received a closed articular fracture of the left tibial plateau at 16 weeks of age, as previously described^13,14,33. Briefly, animals were anesthetized (iso-flurane, concentration 2%) and placed in a custom cradle with the left hindlimb in neutral position (90° flexion). Using a precision material testing system (ELF 3200, TA Instruments, Newcastle, DE), a custom indenter applied a 10N compressive pre-load to the anterior aspect of the left proximal tibial plateau, followed by compression applied at a rate of 20N/s until fracture. The maximum displacement of the indenter was limited to 2.7 mm to create moderate fractures of the tibial plateau. The right hind limb of each mouse served as an uninjured contralateral control.

Under sterile conditions, MaFIA mice received an intra-articular injection of either carrier control solution (n = 3 per time point) or 60 μg of AP20187 (6 μL injection of 10 mg/ml concentration), consisting of 4% ethanol,10% polyethylene glycol 400, and 1.7% Tween 20 in water³⁴ to locally deplete macrophages (n = 6 per time point). In another set of experiments, C57BL/6 mice received an intra-articular injection of either liposomes containing PBS (n = 3 per time point) or 30 μg of clodronate encapsulated in liposomes (6 μL injection of 5 mg/ml concentration)³⁵ to locally deplete macrophages (n = 6 per time point). Mice received intraarticular injections either 2 days prior to fracture (Day −2, n = 9 per genotype), immediately following fracture (Day 0, n = 9 per genotype), or 2 days following fracture (Day −2, n = 9 per geno-type). Mice were sacrificed 7 days post-fracture at 17 weeks of age; no fixation or surgical intervention were implemented (Fig. 1).

Fig 1. Experimental timeline.

Mice received a closed articular fracture of the left knee and a single intra-articular injection of AP20187, clodronate, or carrier control administered either 2 days prior to fracture (Day −2, n = 9), immediately following fracture (Day 0, n = 9), or 2 days following fracture (Day 2, n = 9). Mice were sacrificed 7 days following fracture.

MicroCT analysis

All hind limbs from MaFIA mice and C57BL/6 mice were placed in 10% neutral-buffered formalin for 48 h and then scanned by microCT (Skyscan 1176, Bruker, Madison, Wisconsin). Morphometric bone parameters were determined in the distal femoral condyles, proximal tibial plateau immediately distal to subchondral bone, and metaphyseal region of tibia beginning at the fibular attachment, as previously described¹². Parameters reported in the femoral condyles were trabecular bone fraction (bone volume/total volume) and bone density (mg/cm³). Parameters reported in the tibial plateau and metaphyseal regions include bone volume (mm³) and bone density (mg/cm³).

Histology

Following micro-CT analysis, limbs were decalcified in 1.35N hydrochloric acid solution (Cal-Ex Decalcification Solution, Fisher Scientific) for 5 days. The limbs were then dehydrated in ethanol, infiltrated with xylene, and paraffin-embedded in the coronal orientation. Serial sections (8 μm) in the mid-joint region of both hind limbs from MaFIA mice and clodronate-treated C57BL/6 mice were stained with standard hematoxylin and eosin (H&E) to visualize inflammatory changes in the synovial structure and cellularity. Digital images of the synovial insertion of the lateral femur, medial femur, lateral tibia, and medial tibia were obtained and evaluated separately using a modified form of an established synovitis score for changes in synovial lining thickness and cellular density in the synovial stroma (maximum site score 6) by three blinded graders^10,36. The lining and stroma scores were summed for all four regions to generate a total synovitis score for each knee joint³⁷.

Consecutive joint sections (8 mm) were also used for immune-histochemistry (IHC) to identify macrophages (F4/80, Serotec MCA497G), M1 macrophages (iNOS, Abcam ab3523, reported as NOS2⁺ in text) and M2 macrophages (CD206, Abcam ab64693) present within the lining and stroma of the synovial capsule of the fractured limb (n = 3 from carrier control groups; n = 3 from macrophage depleted groups; subset of joints selected with synovitis scores closest to the mean and demonstrate cellularity), as previously reported^28,38–40. The F4/80⁺, NOS2⁺, and CD206+ cells were quantified for the lining and stroma separately by graders from digital images.

For immunohistochemistry of F4/80 macrophages, sections from MaFIA mice and clodronate-treated C57BL/6 mice were pretreated with 0.01% proteinase K (Sigmae–Aldrich) for 5 min at 37°C for antigen retrieval. Endogenous peroxidase was quenched with 3% H₂O₂ in methanol for 30 min. Sections were incubated with a rat anti-mouse monoclonal antibody against a surface marker of activated macrophages (F4/80, Serotec MCA497G) at a 1:150 dilution at room temperature for 1 h.

For immunohistochemistry of NOS2 M1 macrophages, sections from MaFIA mice and clodronate-treated C57BL/6 mice were pretreated with 0.005% proteinase K (Sigmae–Aldrich) for 5 min at 37°C for antigen retrieval. Endogenous peroxidase was quenched with 3% H₂O₂ in methanol for 1 h. Sections were incubated with a rabbit anti-mouse polyclonal antibody against NOS2, an enzymatic marker of activated classical M1 macrophages (iNOS, Abcam ab3523) at a 1:300 dilution at 4°C overnight.

For immunohistochemistry of CD206 M2 macrophages, sections from MaFIA mice and clodronate-treated C57BL/6 mice were pretreated with Citrate Buffer, pH 6.0 (H-3300, Vector laboratories, Burlingame, CA) for 15 min at 90°C for antigen retrieval. Endogenous peroxidase was quenched with 3% H₂O₂ in methanol for 30 min. Sections were incubated with a rabbit anti-mouse polyclonal antibody against the mouse Mannose Receptor surface marker of activated M2 macrophages (CD206 Mouse anti-Human, Abcam ab64693) at a 1:500 dilution at room temperature for 1 h.

For all immunohistochemistry, negative controls received blocking serum instead of the primary antibody. Following incubation, appropriate species-specific polymer detection kits were utilized (ImmPRESS anti-rat kit MP-7440; ImmPRESS anti-rabbit kit MP-7401, Vector laboratories), followed by DAB substrate (SK-4100, Vector laboratories) for chromogenic detection and hematoxylin counter stain (H-3494, Vector laboratories). Digital images of the joint tissue were obtained (BX53 with DP73 camera, cellSens software, Olympus, Center Valley, PA) for stained and negative control sections.

Statistical analysis

A priori sample size calculations using previous data^13,28 determined that a sample size of n = 3 calculated for 2-tailed independent samples t-test was needed to detect a 25% difference in IHC assessment of macrophages (Supplemental S1). As our prior experience with systemic macrophage depletion (data not shown) resulted in adverse events and death in some animals, the sample size of the macrophage depleted groups were increased to n = 6. We report no adverse events or loss of animals with local intraarticular depletion. To examine statistical difference in synovitis and bone morphology, a multifactorial analysis of variance (ANOVA) was used with depletion or carrier treated as categorical variable and fractured or contralateral control limb treated as a repeated measure with treatment time points analyzed independently. To examine statistical differences in immunohistochemical assessment of macrophages a Student’s t-test was used which compared treatment group of depleted or carrier in fractured limbs with treatment time points and region (lining or synovial stroma) within the joint analyzed independently. In all cases, statistical significant is reported at the 95% confidence level and residual diagnostics presented in Supplemental S2 (STATISTICA v.7, StatSoft, Inc.).

Results

Macrophage depletion with joint injury does not reduce synovitis

Intra-articular depletion of macrophages did not reduce joint synovitis at any time point for either MaFIA mice or clodronate-treated C57BL/6 mice [Fig. 2(A) and (B)]. Surprisingly, pre-depletion of macrophages 2 days prior to injury (Day–2) in MaFIA mice significantly increased synovitis in the injured limb when compared to a carrier control-treated injured limb (Fig. 2(A), P = 0.003; estimate of between-group difference 9.44 with 95% CI 1.77, 17.11), and with the clodronate pre-depletion, synovitis scores in the fractured limbs were lower but not statistically significant in comparison to fractured limbs with carrier control (Fig. 2(B), P = 0.071; estimate of between-group difference −4.17 with 95% CI − 9.14, 0.81). H&E staining of synovial tissue from MaFIA and clodronate-treated C57BL/6 mice treated 2 days after fracture (Day 2) shows increased cellularity and thickness of the synovial lining layer in fractured joints when compared to contralateral control joints [Fig. 2(C) and (D)].

Fig. 2. Effect of intra-articular macrophage (MΦ) depletion on synovitis in the mouse knee at 7 days after fracture.

(A) Total joint synovitis score of the knee in contralateral control and fractured limbs for MaFIA treatment groups. *Significantly increased synovitis of the knee in macrophage depleted fractured limbs with AP20187 compared to fractured limbs receiving carrier control (*P < 0.05) with treatment at Day −2. #Significantly increased synovitis of the knee in fractured limbs compared to respective contralateral control limbs (#P < 0.05). (B) Total joint synovitis score of the knee in contralateral control and fractured limbs for Clodronate treatment groups. #Significantly increased synovitis of the knee in fractured limbs compared to respective contralateral control limbs (#P < 0.05). (C–D) Representative histological sections stained with H&E demonstrating increased cellularity in the synovial joint capsule with fracture in both carrier control and macrophage depleted groups treated at Day 2 for (C) MaFIA and (D) Clodronate groups with yellow arrows indicating cellularity in synovial lining with increase in fractured limbs, and S indicating synovial stroma with orange arrows indicating increased cellularity, T for tibia and M for meniscus in contralateral control and fractured limbs, scale bar = 50 μm.

Pre-injury macrophage depletion in MaFIA mice decreases bone mineral density

Compared with carrier control-treated injured limbs, macrophage depletion in MaFIA mice 2 days prior to fracture (Day −2, P = 0.002; estimate of between-group difference −0.37 with 95% CI 0.61, −0.14) significantly decreased Bone Mineral Density (BMD) in the metaphyseal region of the tibia [Fig. 3(A) and (B)]. This reduction in BMD, illustrated in microCT images [Fig. 3(B)], coincided with the increased synovitis observed in Day–2 macrophage-depleted MaFIA mice. BMD was not altered in MaFIA mice depleted on Day 0 or Day 2 [Fig. 3(A) and (B)]. Similarly, there were no significant differences in BMD in the metaphyseal region between control and macrophage-depleted C57BL/6 mice with clodronate liposome treatment [Fig. 3(C) and (D)]. Other statistically significant changes in bone morphology are detailed in Table I with the majority of changes being a reduction in bone volume or fraction and BMD in fractured limbs compared to contralateral control limbs, and in select experimental conditions, a reduction in bone volume or fraction with macrophage depletion in both fractured and control limbs.

Fig. 3. Bone Morphology in tibial metaphysis of macrophage (MΦ) depleted and carrier control mice.

A) Bone Mineral Density (BMD) of control and fractured limbs for all MaFIA treatment groups with *P < 0.05. (B) MicroCT axial cross-sectional images in MaFIA fractured experimental joints. (C) Bone Mineral Density (BMD) of control and fractured limbs for all clodronate treatment groups. (D) MicroCT axial cross-sectional images in clodronate fractured experimental joints. Scale bar is 1 mm.

Table I. Bone morphology.

Measures of experimental groups at 7 days post-fracture for contralateral control and fractured experimental limbs as assessed by microCT for the femoral condyles, tibial plateau, and tibial metaphysis (mean ± 95% confidence intervals)

Macrophage Depletion at Day −2
	Carrier Control		MaFIA MO Depleted		ANOVA(P value)
Location and Measurement	nonfx limb	fx limb	nonfx limb	fx limb	Group	Limb	Group*Limb
Femoral Condyles
Cancellous Bone Fraction (BV/TV)	77.4 (72.2–82.7)	70.7 (62.2–79.3)	70.5 (66.7–74.2)	66.9 (62.3–71.5)	0.034	0.045	0.519
Bone Mineral Density (mg/mm3)	1.54(1.43–1.64)	1.40 (1.22–1.57)	1.39(1.31–1.64)	1.32 (1.22–1.41)	0.032	0.046	0.525
Tibiai Plateau
Bone Volume (mm3)	70.4 (66.1–74.8)	65.1 (53.4–76.8)	65.4 (63.3–67.5)	64.4 (61.1–67.8)	0.158	0.019	0.076
Bone Mineral Density (mg/mm3)	1.38 (1.28–1.47)	1.26(1.02–1.51)	1.27 (1.23–1.32)	1.25 (1.18–1.32)	0.159	0.018	0.074
Tibial Metaphysis
Bone Volume (mm3)	197(138–256)	267 (176–357)	175 (142–209)*	154(113–195)*	0.004	0.226	0.040
Bone Mineral Density (mg/mm3)	0.681 (0.548–0.814)	0.855 (0.725 – 0.986)+	0.597 (0.513–0.680)	0.482 (0.353 – 0.611)+	0.002	0.520	0.014
	Carrier Control		Coldronate MO Depleted		ANOVA(P value)
Location and Measurement	nonfx limb	fx limb	nonfx limb	fx limb	Group	Limb	Group*Limb
Femoral Condyles
Cancellous Bone Fraction (BV/TV)	76.6 (73.0–80.3)	73.4 (67.8–79.0)	75.6 (73.2–77.9)	73.3 (70.1–76.4)	0.728	0.068	0.745
Bone Mineral Density (mg/mm3)	1.54 (1.47–1.62)	1.47 (1.35–1.60)	1.52 (1.47–1.57)	1.47 (1.40–1.54)	0.722	0.068	0.749
Tibial Plateau
Bone Volume (mm3)	63.4 (58.1–68.8)	65.0 (59.8–70.2)	61.0 (57.3–64.7)	61.3 (57.6–65.1)	0.044	0.637	0.762
Bone Mineral Density (mg/mm3)	1.22 (1.11–1.33)	1.25 (1.17–1.33)	1.17 (1.10–1.24)	1.18 (1.10–1.26)	0.041	0.556	0.808
Tibial Metaphysis
Bone Volume (mm3)	182 (−201 –565)	245 (157–333)	259 (229–289)	286 (238–334)	0.068	0.287	0.658
Bone Mineral Density (mg/mm3)	0.750 (0.405–1.10)	0.762 (0.495–1.03)	0.800 (0.722–0.878)	0.891 (0.775–1.01)	0.057	0.434	0.542
Macrophage Depletion at Day 0
	Carrier Control		MaFIA MO Depleted		ANOVA(P value)
Location and Measurement	nonfx limb	fx limb	nonfx limb	fx limb	Group	Limb	Group*Limb
Femoral Condyles
Cancellous Bone Fraction (BV/TV)	74.8 (69.0–80.7)	66.3 (61.9–70.8)	67.9 (63.3–72.6)	62.4 (57.0–67.8)	0.092	0.002	0.453
Bone Mineral Density (mg/mm3)	1.48 (1.36–1.60)	1.31 (1.21–1.40)	1.34(1.24–1.44)	1.22 (1.11–1.34)	0.092	0.002	0.443
Tibial Plateau
Bone Volume (mm3)	67.4 (57.6–77.1)	62.5 (54.9–70.0)	63.3 (62.3–64.3)	62.7 (59.7–65.7)	0.269	0.022	0.053
Bone Mineral Density (mg/mm3)	1.31 (1.10–1.52)	1.21 (1.05–1.37)	1.23 (1.20–1.25)	1.21 (1.15–1.28)	0.266	0.022	0.053
Tibial Metaphysis
Bone Volume (mm3)	152 (117–187)	214 (108–320)	158 (133–182)	172 (115–228)	0.408	0.073	0.229
Bone Mineral Density (mg/mm3)	0.568 (0.442–0.693)	0.684 (0.446–0.921)	0.566 (0.487–0.645)	0.591 (0.409–0.772)	0.523	0.181	0.366
	Carrier Control		Coldronate MO Depleted		ANOVA(P value)
Location and Measurement	nonfx limb	fx limb	nonfx limb	fx limb	Group	Limb	Group*Limb
Femoral Condyles
Cancellous Bone Fraction (BV/TV)	77.2 (73.5–80.8)	70.5 (64.4–76.6)	78.8 (75.1–82.5)	74.8 (70.4–79.3)	0.288	0.001	0.287
Bone Mineral Density (mg/mm3)	1.53 (1.45–1.61)	1.39(1.26–1.52)	1.56(1.49–1.64)	1.48 (1.39–1.58)	0.288	0.001	0.288
Tibial Plateau
Bone Volume (mm3)	71.9 (67.8–76.0)	68.9 (54.0–83.8)	71.8 (68.2–75.4)	70.2 (67.5–72.8)	0.807	0.059	0.531
Bone Mineral Density (mg/mm3)	1.41 (1.32–1.50)	1.35 (1.03–1.67)	1.41 (1.33–1.48)	1.37 (1.31–1.43)	0.826	0.064	0.568
Tibial Metaphysis
Bone Volume (mm3)	257 (165–348)	261 (111–411)	241 (208–273)	244 (206–282)	0.508	0.798	0.978
Bone Mineral Density (mg/mm3)	0.825 (0.615–1.04)	0.828 (0.612–1.04)	0.809 (0.676–0.941)	0.791 (0.670–0.912)	0.686	0.882	0.828
Macrophage Depletion at Day 2
	Carrier Control		MaFIA MO Depleted		ANOVA(P value)
Location and Measurement	nonfx limb	fx limb	nonfx limb	fx limb	Group	Limb	Group*Limb
Femoral Condyles
Cancellous Bone Fraction (BV/TV)	72.1 (66.6–77.5)	68.2 (62.1–74.2)	73.0 (70.1–76.0)	67.0 (62.5–71.5)	0.965	0.014	0.566
Bone Mineral Density (mg/mm3)	1.43 (1.31–1.54)	1.34 (1.22–1.47)	1.44 (1.38–1.51)	1.32 (1.22–1.41)	0.941	0.014	0.555
Tibiai Plateau
Bone Volume (mm3)	67.2 (63.6–70.7)	64.4 (58.2–70.6)	67.3 (64.4–70.3)	64.3 (59.0–69.5)	0.989	0.137	0.929
Bone Mineral Density (mg/mm3)	1.31 (1.23–1.38)	1.25 (1.12–1.38)	1.31 (1.25–1.37)	1.25 (1.13–1.36)	0.984	0.139	0.924
Tibiai Metaphysis
Bone Volume (mm3)	186 (121–251)	218 (148–288)	197 (170–223)	223 (202–243)	0.602	0.020	0.770
Bone Mineral Density (mg/mm3)	0.682 (0.594–0.770)	0.726 (0.658–0.793)	0.679 (0.614–0.744)	0.719 (0.651–0.786)	0.825	0.248	0.948
	Carrier Control		Coldronate MO Depleted		ANOVA(P value)
Location and Measurement	nonfx limb	fx limb	nonfx limb	fx limb	Group	Limb	Group*Limb
Femoral Condyles
Cancellous Bone Fraction (BV/TV)	76.4 (70.0–82.9)	73.1 (67.4–78.8)	80.0 (76.2–83.7)	73.5 (68.9–78.1)	0.470	0.013	0.384
Bone Mineral Density (mg/mm3)	1.52 (1.38–1.65)	1.44 (1.32–1.56)	1.59(1.51–1.67)	1.46(1.36–1.55)	0.445	0.013	0.405
Tibial Plateau
Bone Volume (mm3)	72.6 (68.2–77.0)	70.9 (61.0–80.8)	72.3 (69.8–74.8)	71.1 (67.7–74.4)	0.971	0.361	0.881
Bone Mineral Density (mg/mm3)	1.42 (1.33–1.52)	1.39 (1.18–1.60)	1.42 (1.36–1.47)	1.39(1.32–1.46)	0.954	0.369	0.891
Tibial Metaphysis
Bone Volume (mm3)	246 (218–274)*	305 (276–334)*+	254 (237–271)	256 (240–273)+	0.031	0.004	0.006
Bone Mineral Density (mg/mm3)	0.771 (0.632–0.911)	0.884 (0.612–1.16)	0.852 (0.799–0.905)	0.832 (0.769–0.895)	0.732	0.092	0.026

MQ = macrophage, nonfx = non-fractured, fx = fractured. Values represent Mean ± SD. Tukey post-hoc testing for Group*limb effect: P < 0.05 for

^*fx vs nonfx limbs;

⁺depleted fx vs control fx limbs.

Transient local macrophage depletion alters synovial macrophage polarity

Surprisingly, transient local macrophage depletion within the acute window of joint injury did not reduce the number of synovial F4/80+ cells, a marker for activated murine macrophages, at 7 days post-fracture in either MaFIA mice or clodronate-treated C57BL/6 mice [Fig. 4(A) and (B)]. In fact, macrophage-depleted MaFIA mice depleted on Day 0 (P = 0.04; estimate of between-group difference 72.4 with 95% CI 4.1, 140.8) and Day 2 (P = 0.03; estimate of between-group difference 117.0 with 95% CI 19.8, 214.2) had significantly higher numbers of F4/80⁺ macrophages in the synovial stroma compared with the carrier control [Fig. 4(A)], with an increase of F4/80⁺ macrophages in the synovial stroma in MaFIA mice depleted on Day–2 that was not statistically significant (P = 0.08; estimate of between-group difference 78.4 with 95% CI - 15.3, 172.2). The synovial lining of macrophage-depleted MaFIA mice showed a small reduction was F4/80⁺ cells at Day –2 and Day 0, but this reduction was not significant at 7 days post-fracture (Day –2, P = 0.23; estimate of between-group difference −6.3 with 95% CI - 18.7, 6.1 and Day 0, P.= 0.42; estimate of between-group difference −10.4 with 95% CI - 42.5, 21.6). C57BL/6 mice pre-depleted with clodronate liposomes at Day–2 also demonstrated higher levels of F4/80⁺ cells in the synovial stroma 7 days post-fracture [Fig. 4(B)], but the results were also not statistically significant (P = 0.07; estimate of between-group difference 31.2 with 95% CI - 4.5, 67.0). Differences in F4/80⁺ cells within the synovial tissue in Day 2 MaFIA and clodronate-treated C57BL/6 mice can be visualized in digital immunohistochemistry images [Fig. 4(C)].

Fig. 4. Effect of intra-articular macrophage (MΦ) depletion on levels of F4/80⁺ macrophages.

(A) Total number of F4/80⁺ macrophages in lining and stroma of fractured limbs for carrier control and experimental macrophage depleted MaFIA treatment groups. *Significant increase of F4/80⁺ macrophages in experimental treatment groups compared to carrier control groups (*P < 0.05). Experimental depleted groups showed increased F4/80⁺ macrophages within the stroma compared to carrier control groups, but the differences were not statistically significant (#P < 0.08). (B) Total number of F4/80⁺ macrophages in lining and stroma of fractured limbs for carrier control and experimental macrophage depleted clodronate treatment groups. With depletion on Day −2, the experimental treatment group demonstrated increased F4/80⁺ macrophages in the stroma compared to the carrier control, but the differences were also not statistically significant (#P < 0.08). (C) Medial femur of histological sections stained with an antibody against the F4/80 surface marker of activated macrophages from the control and experimental treatment groups for MaFIA and clodronate mice receiving local injection 2 days after fracture (Day 2). Scale bar is 50 mm; brown = positive IHC stain, blue = hematoxylin counterstain.

Interestingly, the observed increase in F4/80⁺ macrophage cells in the synovial stroma of macrophage-depleted MaFIA mice coincided with a significant increase in the number of inflammatory M1 macrophages, as indicated by NOS2⁺ cells on immunohistochemistry (Day –2, P = 0.013, estimate of between-group difference 5.44 with 95% CI 1.89, 9.00; Day 0, P = 0.09, estimate of between-group difference 3.33 with 95% CI - 0.83, 7.50; Day 2, P = 0.038, estimate of between-group difference 3.67 with 95% CI 0.30, 7.03) [Fig. 5(A) and (C)]. Additionally, this increased number of F4/80⁺ macrophages and NOS2⁺ M1 macrophages in macrophage-depleted MaFIA mice corresponded to lower, although not statistically different levels of CD206⁺ cells, a marker for regulatory M2 macrophages, in mice depleted on Day 2 (P = 0.086, estimate of between-group difference −15.2 with 95% CI - 33.7, 3.4) and Day –2 (P – 0.05, estimate of between-group difference −50.17 with 95% CI - 100.64, 0.30) [Fig. 6(A) and (C)]. These findings indicate a dominance of M1 macrophages relative to M2 macrophages following macrophage depletion and joint injury in MaFIA mice. Macrophage pre-depletion using clodronate in C57BL/6 mice (Day –2) resulted in an increase of F4/80⁺ cells in the synovial stroma that was not statistically significant (P = 0.07; estimate of between-group difference 31.2 with 95% CI - 4.5, 67.0), but clodronate depletion showed no statistically significant differences in NOS2⁺ cells or CD206⁺ cells in macrophage-depleted C57BL/6 mice in the lining or stroma at any time point [Figs. 5(B) and 6(B)]. Differences in Day 2 depleted MaFIA and clodronate-treated C57BL/6 mice of NOS2⁺ cells [Fig. 5(C)] and CD206⁺ cells [Fig. 6(C)] within the synovial tissue can be visualized in digital images of immunohistochemistry. In reviewing the overall changes as quantified from all regions in the joint, intra-articular macrophage depletion regardless of method preferentially depleted CD206⁺ M2 macrophages in the articular fracture injury model at 7 days post-fracture (P = 0.03; estimate of between-group difference −24.5 with 95% CI - 46.5, −2.6) with the pre-depletion time at Day –2 having the lowest level of CD206+ M2 macrophages (P = 0.003; estimate of between-group difference 46.4 with 95% CI 13.0, 79.8), as demonstrated in Table II.

Fig. 5. Effect of intra-articular macrophage (MΦ) depletion on levels of NOS2⁺ inflammatory macrophages.

(A) Total scores for the number of NOS2⁺ macrophages in lining and stroma of fractured limbs for carrier control and experimental macrophage depleted MaFIA treatment groups. *Significant increase of NOS2⁺ macrophages in experimental treatment groups compared to carrier control groups (*P < 0.05). (B) Total scores for number of NOS2⁺ macrophages in lining and stroma of fractured limbs for carrier control and experimental macrophage depleted clodronate treatment groups. (C) Medial femur of histological sections stained with an antibody against the NOS2 marker of activated macrophages from the fractured knee of carrier control and experimental treatment groups for MaFIA and clodronate mice receiving local injection 2 days after fracture (Day 2). Scale bar is 50 mm; brown = positive IHC stain, blue = hematoxylin counterstain.

Fig. 6. Effect of intra-articular macrophage (MΦ) depletion on levels of CD206⁺ regulatory macrophages.

(A) Total number of CD206⁺ macrophages in lining and stroma of fractured limbs for carrier control and experimental macrophage depleted MaFIA treatment groups. With depletion on Day −2 and Day 2, experimental treatment limbs showed a decrease in CD206⁺ macrophages in the stroma compared to carrier control, but the differences were not statistically significant (#P < 0.09). (B) Total number of CD206⁺ macrophages in lining and stroma of fractured limbs for carrier control and experimental macrophage depleted clodronate treatment groups. (C) Medial femur of histological sections stained with an antibody against the CD206 surface marker of activated macrophages from the control and experimental treatment groups for MaFIA and clodronate mice receiving local injection 2 days after fracture (Day 2). Scale bar is 50 mm; brown = positive IHC stain, blue = hematoxylin counterstain.

Table II.

Quantification of CD206⁺ M2 macrophages from immunohistochemistry.

Depletion Days Since Depletion		Control*	Depleted
Day.−2**	9	25.4 (4.83–46.1)	12.6(5.77–19.5)
Day 0	7	66.8 (14.7–119)	49.4 (1.16–97.7)
Day 2	5	87.1 (55.5–119)	43.8(24.8–62.7)
ANOVA:	depletion time P = 0.003	depletion group P = 0.03	time • group P = 0.47

Mean (95% confidence intervals).

ANOVA:

^*P=0.03

^**P=0.003 depletion time: Day−2<Day 0 (P = 0.004) and Day 2 (P = 0.02) Tukey’s post-hoc

Discussion

Transient, resolving inflammation plays a critical role in wound repair; however, severe, non-resolving inflammation associated with joint injury has been shown to contribute to joint degeneration and arthritic changes^6–9. Particularly, infiltration of synovial macrophages, the major cell sources responsible for elevating levels of inflammatory cytokines, including IL-1 and TNF-α, within the synovial fluid, are observed in C57BL/6 mice following joint fracture^10,12. The inflammation-associated joint tissue catabolism, however, is diminished in MRL/MpJ “superhealer” mice following articular fracture, and subsequently prevents PTA development^13,14. Furthermore, targeted inhibition of IL-1 using intra-articular delivery of interleukin 1 receptor antagonist (IL-1Ra) reduces joint inflammation and prevents the onset of PTA^24,25. Thus, we hypothesized that a transient, intra-articular depletion of macrophages would reduce joint synovitis during the acute phase of joint injury. However, we found that acute local macrophage depletion with articular fracture, using both the transgenic MaFIA mouse model or using clodronate liposomes in C57BL/6 mice, did not reduce synovitis or levels of macrophages in the synovium 7 days after fracture. In macrophage-depleted MaFIA mice, intraarticular macrophage depletion paradoxically altered the polarity of macrophages to a dominance of M1 inflammatory macrophages within the joint synovium. Our data suggest that the synovial macrophage population plays a pivotal immunomodulatory role in the acute inflammatory response to joint injury.

Macrophages are key modulators of bone homeostasis by regulating osteoblasts through osteal macrophages (osteomacs), cells critical for bone remodeling and mineralization^34,41. Synovial macrophages, specifically, have been shown to mediate osteophyte formation⁴². In a study of parathyroid hormone (PTH)-induced bone anabolism, systemic macrophage depletion in MaFIA mice resulted in a reduction of cortical and trabecular bone mass that was not responsive to PTH⁴³. However, systemic depletion of macrophages in the same model using clodronate liposomes had the opposite effect⁴³, suggesting that these methods may target different macrophage subpopulations. While we did not observe a significant increase in BMD when synovial macrophages were depleted locally using clodronate, we did observe a significant decrease in BMD in pre-depleted MaFIA mice (Day–2). This decrease in bone density accompanied by a significant increase in synovitis at 7 days following fracture is similar to early rheumatoid arthritis pathology of periarticular bone loss in association with active synovitis and inflammation^44,45. Findings from this study further strengthen the evidence that macrophages play a critical role in bone homeostasis, particularly within the region of a fracture.

Surprisingly, local macrophage depletion using either method did not reduce acute joint synovitis following articular fracture at any time point evaluated. A limitation of the study was that macrophage depletion was not evaluated at 1–2 days after administration. Assessment of the synovium at 7 days post-fracture was selected based on previously reported results that significant synovitis of the joint capsule with infiltration of macrophages is evident histologically both in the mouse^10,13 and in human patient biopsy samples¹¹ at this time point. The finding that neither depletion method reduced synovitis with articular fracture may be due to a preferential depletion of anti-inflammatory cells⁴⁶. Furthermore, when MaFIA mice were depleted 2 days prior to fracture (Day –2), joint synovitis significantly increased with joint injury. Consistent with these findings, macrophage depletion with joint injury also did not reduce the number of F4/80⁺ macrophages present within the synovium 7 days post-fracture. While others have successfully reduced the synovial macrophage population using systemic administration of AP20187 to MaFIA mice²⁸ or intraarticular injection of clodronate liposomes^22,42, our model involving joint trauma via articular fracture prevented this effect. In macrophage-depleted MaFIA mice on Day 0 and Day 2, we observed a significant increase of F4/80⁺ macrophages within the deep layers of the synovial stroma, suggesting either a compensatory influx or proliferation of macrophages. In diet-induced obese MaFIA mice, systemic macrophage depletion similarly induced systemic inflammation and did not attenuate the severity of OA²⁸. Systemic macrophage depletion using clodronate liposomes in diet-induced obese C57BL/6 mice likewise resulted in increased systemic inflammation in the form of neutrophilia and altered circulating cytokines²⁹. Also a limitation of this study is that additional cell types within the synovium were not identified. However, we can speculate that similar to previous findings²⁸, neutrophils may increase as a compensatory mechanism. Future investigations may look at immune cell phenotyping via polychromatic flow cytometry to better characterize the cell subsets involved following articular fracture.

Additionally, in MaFIA mice, systemic depletion of macrophages in a skeletal muscle injury model has been shown to cause a compensatory increase, or rebound, of macrophages, while systemic clodronate-treated mice did not have the same effect⁴⁷,⁴⁸.We have previously shown that the acute 7-day period following articular fracture is associated with greater synovitis, which may contribute to long-term dysregulation of chondrocyte function and lead to degenerative changes in the articular cartilage characteristic of PTA¹⁰. Our goal in this study was to modulate the acute inflammatory response with a transient depletion of macrophages. However, we found no reduction in the number of macrophages and in some cases an increased number, suggesting a possible rebound effect in response to a transient depletion of macrophage in the setting of an intra-articular fracture. Fate-mapping approaches used in a recent study determined that monocyte-derived macrophages recruited to the synovium actively contribute to joint inflammation, but a local population of synovial epithelial-like macrophages provided a protective barrier⁴⁸. This new insight suggests that intra-articular depletion may preferentially deplete this anti-inflammatory population of macrophages.

Interestingly, our macrophage depletion method resulted in a change in the polarity of macrophages to a dominance of the M1 classically-activated lineage in macrophage-depleted MaFIA mice with a trend toward lower levels of M2 regulatory macrophages within the synovial stroma. One possible explanation of this shift in macrophage polarity is that an influx of monocytes, from locations such as the bone marrow, may have further differentiated into M1 macrophages in response to fracture. Alternatively, native synovial M2 macrophages may have switched their phenotype to M1 macrophages following injury. Dissecting the precise mechanisms by which the transient intra-articular macrophage depletion significantly increased M1, but not M2, synovial macrophages in MaFIA mice warrants further investigation.

A similar shift in macrophage polarity in AP20187-treated MaFIA mice has been reported in a glioma model, resulting in both an influx of M1 macrophages and a reduction in the proliferation of M2 macrophages⁴⁹. In a model of systemic macrophage depletion in MaFIA mice with a destabilized medial meniscal injury, synovial macrophages were successfully reduced at day 0; this depletion predominantly affected the M2 macrophage population at 9 weeks post-depletion²⁸. The M1/M2 ratio is critical for healthy bone homeostasis; enhancing the M2 macrophage population can improve bone healing outcomes⁵⁰. The dominant M1 macrophage population in our macrophage-depleted MaFIA mice with decreased BMD suggests that M1 macrophages may be reciprocally detrimental for bone homeostasis, and further reinforces the importance of the ratio of M1 to M2 macrophages.

Although we observed an increase in F4/80⁺ macrophages in the clodronate-treated C57BL/6 mice, a shift in macrophage polarity was not observed. Additionally, no differences in bone morphology were observed in the clodronate-treated C57BL/6 mice. These differences may be the result of inherent characteristics of the macrophage depletion methods. Mechanistically, administration of AP20187, a small molecule inhibitor, targets all CSF⁺ cells in MaFIA mice, while clodronate liposomes induce apoptosis in phagocytic cells. As a result, AP20187 depletes all monocytic cells in MaFIA mice, while clodronate primarily targets mature macrophage cells. This may explain the stable M1/M2 polarity in the clodronate-treated mice due to the lack of effect on the monocytic lineage itself. Clodronate also has anti-inflammatory effects independent of its effects on macrophages by inhibiting pro-inflammatory cytokines, making it useful in preventing synovial inflammation in models of rheumatoid arthritis and ankylosing spondylitis⁵¹. This may account for the attenuated inflammatory phenotype of the clodronate-treated mice when compared to the MaFIA depleted mice. In addition to its effect on inflammation, clodronate inhibits bone resorption by reducing osteoclastic activity⁵², providing a potential explanation for the variation in bone morphological alterations between MaFIA and C57BL/6 mice. Furthermore, clodronate liposomes are much larger than the AP20187 molecule, which leads to difficulty in penetrating the synovium and vasculature⁵³, making their effect less potent.

Our findings suggest that while prolonged inflammation has degenerative effects on cartilage, macrophages are critical immunomodulators in the setting of joint injury²⁸. This study provides new insight on the regulatory role of macrophages in joint synovitis and bone homeostasis after joint fracture.