The Implant-Induced Foreign Body Response Is Limited by CD13-Dependent Regulation of Ubiquitination of Fusogenic Proteins

Key Points

• CD13 is a negative regulator of macrophage fusion and giant cell formation.

• Implant-induced FBGC formation and implant site fibrosis are increased in CD13^KO mice.

• CD13 controls fusogen expression and actin protrusion to regulate fusion.

Visual Abstract

Abstract

Implanted medical devices, from artificial heart valves and arthroscopic joints to implantable sensors, often induce a foreign body response (FBR), a form of chronic inflammation resulting from the inflammatory reaction to a persistent foreign stimulus. The FBR is characterized by a subset of multinucleated giant cells (MGCs) formed by macrophage fusion, the foreign body giant cells (FBGCs), accompanied by inflammatory cytokines, matrix deposition, and eventually deleterious fibrotic implant encapsulation. Despite efforts to improve biocompatibility, implant-induced FBR persists, compromising the utility of devices and making efforts to control the FBR imperative for long-term function. Controlling macrophage fusion in FBGC formation presents a logical target to prevent implant failure, but the actual contribution of FBGCs to FBR-induced damage is controversial. CD13 is a molecular scaffold, and in vitro induction of CD13^KO bone marrow progenitors generates many more MGCs than the wild type, suggesting that CD13 regulates macrophage fusion. In the mesh implant model of FBR, CD13^KO mice produced significantly more peri-implant FBGCs with enhanced TGF-β expression and increased collagen deposition versus the wild type. Prior to fusion, increased protrusion and microprotrusion formation accompanies hyperfusion in the absence of CD13. Expression of fusogenic proteins driving cell–cell fusion was aberrantly sustained at high levels in CD13^KO MGCs, which we show is due to a novel CD13 function, to our knowledge, regulating ubiquitin/proteasomal protein degradation. We propose CD13 as a physiologic brake limiting aberrant macrophage fusion and the FBR, and it may be a novel therapeutic target to improve the success of implanted medical devices. Furthermore, our data directly implicate FBGCs in the detrimental fibrosis that characterizes the FBR.

Introduction

In mammals, both homotypic and heterotypic cell–cell fusion are crucial to multiple physiological and pathological processes (1). The mechanisms and proteins controlling cell–cell fusion in different cell types are remarkably diverse across the various systems in their structure, function, and utilization of auxiliary pathways. However, some of the molecules responsible for merging individual membranes appear to be shared and thus participate in fusion of more than one cell type. Several of these key shared proteins have been identified, but our mechanistic understanding of precisely how these may promote cell–cell fusion is still largely theoretical and somewhat controversial, suggesting that additional proteins that drive fusion are yet to be discovered (2, 3). Clearly, identifying novel membrane-active fusion proteins and dissecting their underlying mechanisms will provide the clues to inform and direct further investigations into shared and unique mechanisms of membrane remodeling, membrane merger, and cell–cell fusion.

Multinucleated giant cells (MGCs) are specialized cells of the immune system that result from the fusion of mononuclear myeloid cells to form a single, large, multinucleated cell with a shared cytoplasm. MGC formation is typified by osteoclast fusion during bone remodeling, but the immune response to therapeutic implantation of foreign materials such as medical devices, mesh supports, arthroscopic implants, and biosensors also provokes a robust MGC response, the foreign body response (FBR) (4–7). Initially, circulating proteins adsorb to the surface of the implant, eliciting acute and chronic inflammatory responses mediated by cytokine release from infiltrating immune cells (6). Macrophages at the implant site fuse to form a type of specialized MGC called “foreign body giant cells” (FBGCs), the hallmark cells of the FBR. This is followed by fibrin matrix deposition, granulation tissue, and fibrous capsule formation around the implanted biomaterial (6), leading to loss of host–implant communication and implant failure. Although FBGCs have been shown to be consistently present at the site of implant, the relative contribution of FGBCs in FBR and more specifically, in fibrosis, is poorly understood (8, 9). However, numerous investigations have linked FBGCs to the degradation of foreign particles through secretion of reactive oxygen species (ROS), matrix metalloproteinases (MMPs), and acids (10–14) and are thought to be a major source of cytokines and chemokines (15, 16), ROS (10, 11), and angiogenic mediators. To date, investigations to improve implant failure have focused primarily on improving the biocompatibility of medical devices or preventing the encapsulation of biomaterial, but they have been unsuccessful.

FBGC formation is a multistep process that initially requires differentiation of precursors into macrophages, which fuse to form active MGCs in response to cytokines IL-4 and IL-13. Membrane fusion involves various membrane-associated processes, such as membrane assembly (17–19), clustering of membrane molecules (20), endocytosis (21), adhesion (22, 23), and cytoskeletal protrusion and rearrangement (24), leading to the characteristic MGC. Independent investigations have implicated specific molecules as fusion promoters, or fusogens, such as dendritic cell–specific transmembrane protein (DC-STAMP) (25), osteoclast stimulatory transmembrane protein (26), dynamin (21), tetraspanins (18, 27–29), ATP6v0d2 (30), syncytin-1 (31), annexins (32), and S100 (33). However, how and if these proteins are part of a larger network has not been elucidated (3, 34), thus warranting further investigation to integrate these into a unifying mechanism.

CD13 is a multifunctional transmembrane aminopeptidase that we and others have shown to be involved in cell migration, actin cytoskeletal organization, cell–cell and cell–extracellular matrix adhesion, receptor-mediated endocytosis, and recycling (35–40). Recently, we have demonstrated, to our knowledge, a novel role of CD13 in osteoclastogenesis at the level of cell–cell fusion (41) where CD13-knockout (CD13^KO) mice exhibited a low bone mass phenotype with increased osteoclast numbers per bone surface area, consistent with enhanced giant cell–mediated osteolysis and bone destruction. In the present study, we extend these observations and demonstrate that in vitro induction of fusion in CD13-deficient myeloid progenitors generated from bone marrow (BM) or thioglycolate (TG)-elicited peritoneal macrophages results in hyperfusion to generate MGCs that are considerably larger in size and contain many more nuclei than those from their wild-type (WT) counterparts, suggesting that CD13 also regulates MGC formation at the stage of fusion. In an established murine model of FBR, the s.c. implantation of polyethylene mesh, used in hernia repair, in CD13^KO mice produces significantly more FBGCs over time with a concomitant increase in levels of the serum cytokines IL-1β and TNF-α. Furthermore, collagen deposition and intracellular TGF-β expression levels in FBGCs were significantly enhanced in the peri-implant region in the CD13^KO compared with WT mice, suggesting that FBGCs directly contribute to the FBR. Mechanistically, the expression levels of the key fusion proteins, DC-STAMP, the tetraspanins CD9 and CD81 were sustained at high levels in CD13^KO compared with WT MGCs postfusion by a ubiquitin/proteasomal degradation pathway with loss of association of fusogens with the scaffold protein, Cullin-4A, a component of the CUL4-RING E3 ubiquitin ligase (CRL4 E3 ligase) complex involved in protein turnover (42, 43). In addition, actin rearrangements are critical for the formation and stability of the fusion complex along with the several proteins, including syncytins, that are involved in the merging of the cells. Indeed, transmission electron microscopy (TEM) analysis of myeloid cells undergoing fusion indicated that CD13^KO macrophages clearly produced more actin protrusions and microprojections than WT prior to fusion as early as day 2 (d2) postfusion. Taken together, to our knowledge, our study provides three novel insights into the mechanism of FBGC formation and the evolution of the FBR: (1) FBGCs not only identify the FBR but also play an active role in promoting FBR-dependent fibrosis; (2) CD13 regulates the cytoskeletal changes necessary for protrusion-dependent FBGC fusion; and (3) CD13 controls protein turnover by promoting the proper assembly of ubiquitination complexes. Taken together, CD13 acts as a brake to regulate physiologic macrophage fusion in two distinct lineages, osteoclasts and FBGCs, and may be a target for therapeutic intervention at early stages to control pathologic consequences of aberrant cell–cell fusion.

Materials and Methods

Mice

Global young (8–10 wk) WT and CD13^KO (C57BL/6J) male and female mice were generated and housed at the Gene Targeting and Transgenic Facility at the University of Connecticut School of Medicine. All experimental and control mice were littermates. All procedures were performed in accordance with the guidelines and regulations approved by the University of Connecticut School of Medicine Institutional Animal Care and Use Committee. The University of Connecticut School of Medicine is fully accredited by the Association for Assessment and Accreditation of Laboratory Animal Care International. Its Public Health Service assurance number is A3471-01 (D16-00295), and its U.S. Department of Agriculture registration number is 16-R-0025. Euthanasia by CO₂ followed by cervical dislocation was performed and is an accepted method consistent with American Veterinary Medical Association Guidelines for the Euthanasia of Animals to minimize the pain or discomfort in animals.

Flow cytometry

BM progenitors were stained with the Abs (Supplemental Table I) for phenotypic analysis and sorting. Flow cytometry of live cells derived from WT and CD13^KO BM were performed with a BD FACSAria cytometer to obtain (CD3/B220/Nk1.1)⁻ CD11b^lo/− CD115^hi Ly6G⁺ progenitor cells. Data were analyzed using BD FACSDiva version 9.0 and FlowJo version 9.9 software. For measurement of surface fusogens, anti-CD9 Ab-FITC (BioLegend; clone MZ3) and anti–DC-STAMP Ab (Creative Diagnostics; clone 2B3) were used.

Isolation and culture of macrophage progenitors from bone marrow

BM cells were obtained by isolating the femur and tibia from WT and CD13^KO mice of both sexes. BM cells were harvested from femur and tibia bones with 10 ml 1× PBS supplemented with 2% FBS. Cells were treated with RBC lysis buffer and filtered through a 40-μm cell strainer (Corning; 431750). (CD3/B220/Nk1.1)⁻ CD11b^lo/− CD115^hi Ly6G⁺ progenitor cells were flow sorted and seeded at a density of 20,000–25,000 cells per well on a 96-well dish at d0 in DMEM (Life Technologies; 11965092) with 10% FBS and 1% penicillin-streptomycin. Cells were treated with 20 ng/ml M-CSF for 2 d for macrophage differentiation and proliferation, followed by treatment with 30 ng/ml each of Il-4 (R&D Systems; 404ML010) and IL-13 (R&D Systems; ML010) cytokines in complete media and incubated at 37°C with 5% CO₂ for an additional 3–5 d, supplemented with fresh cytokines every 48 h to allow fusion to proceed.

Generation and culture of thioglycolate-elicited peritoneal macrophages

Both WT and CD13^KO mice of both sexes were i.p. injected with 1 ml 4% Brewer TG solution (BD; 211716). After 48 h, mice were euthanized via CO₂ and cervical dislocation and injected with 10 ml cold PBS into the peritoneal cavity to collect peritoneal macrophages. Peritoneal lavage cells suspended in PBS were extracted and treated with RBC lysis buffer for 5 min, centrifuged at 1200 rpm for 5 min, and resuspended in α-MEM (Life Technologies; 12571063) media supplemented with 10% FBS (Life Technologies; 16250078) and 1% penicillin-streptomycin (Life Technologies; 15140122). Total live cells were counted manually with a Bright Line hemacytometer with Trypan Blue dye. A total of 50,000–75,000 cells per well were seeded on a 24-well dish at d0. Fusion-inducing cytokines IL-4 and IL-13 were added at 60 ng/ml each in complete media at d3 and incubated at 37°C with 5% CO₂ for an additional 3–5 d.

In vivo implant analysis

WT and CD13^KO mice of both sexes were anesthetized with isoflurane. Prior to implantation, mice were injected s.c. with 8 mg/kg bupivacaine (Hospira NDC; 0409-1163-18) at the site of implantation to numb the area. A 0.5-inch × 0.5-inch sterile polypropylene mesh implant (Ethicon; PMXS) was inserted s.c. in the flank region of the mice at d0, followed by peritoneal injection with 30 µl meloxicam (Covetrus NDC; 11695-6936-1) to reduce pain. Implants were harvested at d1, d7, and d14 following euthanasia with CO₂ and cervical dislocation. Harvested implants were fixed in 4% formalin overnight and embedded in paraffin. Formalin-fixed, paraffin-embedded (FFPE) implant sections were deparaffinized and subjected to immunofluorescence (IF) or light microscopy as indicated.

Preparation of optical glass for thioglycolate-elicited macrophage fusion

An optical coverglass was acid cleaned with 12 N HCl, washed with deionized water, rinsed with ethanol, and allowed to dry overnight. The acid-cleaned coverglass was treated with 1 mg/ml paraffin wax (Sigma-Aldrich) solubilized in toluene for a few minutes, then excess solution was removed, washed with sterile water, allowed to dry, and sterilized under UV light.

Immunofluorescence assay

TG-elicited peritoneal macrophages (TG-Macs) grown on coverslips were washed twice with 1× PBS and fixed with 4% paraformaldehyde for 20 min. Cells were then permeabilized for 5 min with 0.1% Triton-X and blocked for 1 h with 5% BSA. Primary Abs (Supplemental Table I) were added in 1% normal serum in 5% BSA buffer and incubated at 4°C overnight. Coverslips were washed twice with 1× PBS, and Alexa Fluor secondary Abs were added in addition to DAPI (nuclear stain) in 1% BSA and incubated at 4°C overnight. Coverslips were mounted onto microscope slides with Prolong Gold Antifade mounting media (Molecular Probes; P10144).

For tissue sections, postdeparaffinization, implant sections were subjected to Ag retrieval with 10 mM sodium citrate and microwaved for 1 min 25 s. Slides were blocked with 5% serum in 5% BSA buffer for 1 h at room temperature (RT). Slides were washed twice with 1× PBS, and primary Abs were added with 1% normal serum in 5% BSA buffer and incubated at 4°C overnight. Alexa Fluor secondary Abs and DAPI were added in 1% BSA and incubated for 1 h at RT. All samples were imaged at excitation wavelengths of 488 nm (Alexa Fluor 488), 543 nm (Alexa Fluor 594), and 405 nm (DAPI). Samples were imaged by an inverted fluorescence microscope and analyzed by using Zeiss ZEN 2.0 Pro blue edition software.

Giemsa staining

BM progenitors or TG-Macs grown on plastic were fixed with 100% methanol for 10 min, followed by Giemsa staining (Harleco; R03055), and diluted 1:20 in dH₂O for 25 min. Cells were then washed twice with dH₂O and allowed to air dry. Cells were imaged via light microscopy using a Zeiss fluorescence inverted microscope and analyzed by using Zeiss ZEN 2.0 Pro blue edition software. The average protrusion length of fusing cells was measured with ImageJ software (https://imagej.nih.gov/ij/).

Measurement of fusion index

Random fields with similar cell densities of Giemsa-stained or IF images were considered for evaluation of the fusion index and quantified by three individuals in a blinded manner as follows:

$$\%\mathrm{\hspace{0.17em}Fusion\hspace{0.17em}index\hspace{0.17em}}=\left[\mathrm{Total\hspace{0.17em}number\hspace{0.17em}of\hspace{0.17em}nuclei}\mathrm{\hspace{0.17em}in\hspace{0.17em}fused\hspace{0.17em}cells\hspace{0.17em}}\left(>\text{3}\right)/\mathrm{Total\hspace{0.17em}number\hspace{0.17em}of\hspace{0.17em}nuclei\hspace{0.17em}in}\mathrm{\hspace{0.17em}fused\hspace{0.17em}and\hspace{0.17em}non}-\mathrm{fused\hspace{0.17em}cells}\right]\text{\hspace{0.17em}×\hspace{0.17em}1}00$$

Masson trichrome staining

After deparaffinization, in vivo implant tissue was stained with Masson trichrome. Masson trichrome staining was performed at the Research Histology Core Labs at UConn Health. The intensity of fibrosis by Masson trichrome was calculated using ImageJ software (https://imagej.nih.gov/ij/).

Electron microscopy

Flow-sorted mouse BM cells were cultured in 96-well dishes as described above. To make cells amenable to processing for electron microscopy, pieces of Aclar plastic (Electron Microscopy Sciences [EMS]; 50425-10) were placed in the bottom of the 96-well dishes, allowing cells to grow and fuse on the removable substrate.

After fusion, the cells grown on Aclar were fixed in 2.5% glutaraldehyde (EMS) in 0.1 M sodium cacodylate buffer (EMS) on ice for 15–30 min. Cells were rinsed in 0.1 M cacodylate buffer five times, then postfixed in 1% osmium tetroxide (EMS) and 0.8% potassium ferricyanide in 0.1 M cacodylate buffer for 15–30 min. Cells were then rinsed with Milli-Q filtered water and stained with 1% aqueous uranyl acetate (EMS) for 15–30 min. Samples were rinsed five times with water, then dehydrated in a graded series of ethanol (50%, 75%, 95%, 100%,) for 5–10 min each. Using an aluminum dish, the Aclar pieces with cells were further dehydrated in two changes of propylene oxide (EMS) for 2–3 min each. Next, the cells were infiltrated in Polybed 812 epoxy resin with BDMA in graded steps with propylene oxide. The cells on Aclar were flat embedded between larger pieces of Aclar plastic with spacers made by layers of parafilm.

TEM image acquisition/analysis

Seventy-nanometer sections were collected on copper mesh EM grids (EMS). Fusing WT and CD13^KO monocytes/macrophages that were within 2 μm of each other were imaged at 4000× on a Hitachi H-7650 at 80 kV. Cell protrusions and microprojections (original magnification ×4000) between two neighboring monocytes at d2 postfusion that were at least 1 μm apart were counted by three individuals in a blinded manner.

Quantitative RT-PCR

Total RNA was extracted using TRIzol reagent (Invitrogen) according to the manufacturer’s instruction and as described before (41). The relative transcript level was normalized to the GAPDH level. GenBank primer sequences (http://pga.mgh.harvard.edu/primerbank/) were used to determine fusogen primer sequences. PCR primer sequences used were as follows: DC-STAMP, 5′-GGGGACTTATGTGTTTCCACG-3′ (forward) and 5′-ACAAAGCAACAGACTCCCAAAT-3′ (reverse); CD9, 5′-ATGCCGGTCAAAGGAGGTAG-3′ (forward) and 5′-GCCATAGTCCAATAGCAAGCA-3′ (reverse); CD81, 5′-CAGATCGCCAAGGATGTGAAG-3′ (forward) and 5′-GCCACAACAGTTGAGCGTCT-3′ (reverse); GAPDH, 5′-GGATTTGGTCGTATTGGG-3′ (forward), 5′-GGAAGATGGTGATGGGATT-3′ (reverse). Transcript data were analyzed using CFX Manager version 3.1 (https://www.bio-rad.com/en-us/sku/1845000-cfx-manager-software?ID=1845000) (Bio-Rad Laboratories).

Capture ELISA-based pulse-chase experiment

To assess reexpression of fusogens in cells undergoing fusion, WT or CD13^KO TG-Macs at d5 were surface labeled with 0.2 mg/ml NHS-S-Biotin (Pierce) at 4°C for 30 min. Labeled cells were washed in cold PBS, and internalization was allowed to proceed at 37°C for 30 min, at which point the remaining surface biotin was removed by treatment with 20 mM MesNa in 50 mM Tris-HCl (pH 8.6) and 100 mM NaCl for 15 min at 4°C. MesNa was quenched by lodoacetamide (20 mM) treatment. Cells were then chased in label-free complete medium for 30 min at 37°C. A second round of MesNa was administered to remove surface biotin from recycled fusogens to measure the relative amount of internalized fusogen in the cell lysate by capture ELISA. Cells were lysed with a protease inhibitor mixture in radioimmunoprecipitation assay lysis buffer. To quantify the amount of remaining biotinylated fusogens present in the cell lysate, 96-well microtiter plates were coated with 5 μg/ml anti-CD9 (MyBiosource; MBS7113510) or anti–DC-STAMP (Creative Diagnostics; clone 2B3) in 0.05 M sodium carbonate (pH 9.6) at 4°C. Plates were blocked in PBS containing 5% BSA at RT for 1 h. Captured fusogens were measured by overnight incubation of cell lysate at 4°C. Plates were washed in PBS/Tween-20 and incubated with streptavidin-HRP for 1 h at 4°C, and biotinylated CD9 or DC-STAMP was measured by addition of a chromogenic agent, ortho-phenylenediamine. Fusogen surface reexpression was calculated on the basis of the assumption that percentage internalization + percentage surface reexpression = 100%; therefore percentage fusogen reexpression = 100% − percentage fusogens internalized after chase. To assess fusogen reexpression excluding new protein synthesis, the cells were treated with 100 µg/ml cycloheximide (CHX) for 6 h prior to internalization. To further assess the role of fusogen recycling, the cells were treated with 100 µg/ml CHX and 5–10 µM MG132 for 6 h prior to internalization.

Chemiluminescence-based assay of polyubiquitination of target fusogens

WT and CD13^KO peritoneal macrophages under the fusogenic condition were treated with MG132 at 2.5 µg, 5.0 µg, and 10 µg for 3 h, followed by lysis with radioimmunoprecipitation assay lysis buffer. Cell lysate (20 µg) was added to a proteolysis targeting chimeras (PROTAC) assay plate (LifeSensors; PA950) and incubated at RT for 2 h. Wells were washed, and anti–DC-STAMP (Creative Diagnostics; clone 2B3) or CD9 (MyBiosource; MBS7113510) Ab was added in 1× blocking agent and incubated at RT for 1 h followed by addition of HRP-conjugated secondary Ab and incubated at RT for 1 h. Detection reagents (supplied in the kit) were added, and the plates were subsequently read by a microplate reader according to the manufacturer’s instruction.

Statistical analysis

Statistical analysis was performed using an unpaired, two-tailed Student t test using GraphPad Prism software, and results are representative of the mean ± SD or ± SEM as indicated. Differences at p ≤ 0.05 were considered significant.

Results

CD13 negatively regulates myeloid cell fusion

Previous studies from our laboratory and other investigations have established that CD13 is an essential mechanistic component of cellular functions required for cell–cell fusion (35, 40). More recently, we have demonstrated that compared with WT animals, CD13^KO mice had markedly reduced bone density with increased numbers of osteoclasts per bone surface (41). In the published work, mice of both genotypes showed equivalent osteoclast progenitor populations, bone formation, and mineral apposition rates, suggesting a functional defect in CD13^KO osteoclasts. Indeed, cytokine-mediated osteoclastogenesis in vitro in CD13^KO myeloid progenitors generated larger osteoclasts with considerably more nuclei and significantly higher bone resorptive function than WT, indicating that CD13 controls osteoclastogenesis at the fusion stage. Extending these observations, we investigated whether CD13 also impacts fusion of a distinct lineage of fusogenic myeloid cells that derive from different progenitors, the MGCs (44). We differentiated flow-sorted WT and CD13^KO BM-derived progenitors (CD3⁻, B220⁻NK1.1⁻, CD11b^lo/−, CD115^hi, Ly6G⁺) into macrophages in the presence of M-CSF for 3 d, followed by incubation in IL-4 + IL-13 for an additional 3–5 d to promote MGC differentiation and fusion (Fig. 1A, 1D, Supplemental Fig. 1A). Giemsa-stained cells in areas with comparable cell densities were analyzed in a blinded manner, and the extent of MGC fusion at d3 postinduction was represented as the fusion index ([total number of nuclei in fused cells per field/total number of nuclei per field] × 100]. We found significant increases (2-fold) in MGC formation in CD13^KO cells compared with WT as early as d3 (Fig. 1D), suggesting that the presence of CD13 limits or restricts the fusion process. However, the total number of nuclei/dish was equivalent between genotypes, confirming that this accelerated fusion in CD13^KO cultures is not due to enhanced proliferation rates (Supplemental Fig. 1B). To circumvent the technical obstacle that BM-derived progenitor cells fuse poorly on the glass substrates required for high-resolution optical immunohistochemical analysis, we adapted an alternative method of treating optical glass that enabled macrophage fusion (45). WT and CD13^KO TG-Macs seeded on plastic for Giemsa staining (Fig 1B, 1E) or treated optical glass for IF staining (Fig 1C, 1F) readily fused in the presence of MGC cytokines IL-4 + IL-13. This recapitulated our results in BM-derived progenitors as identified by brightfield and positive immunostaining for Mac-3, the mouse macrophage differentiation Ag, where the fusion index was significantly higher in the absence of CD13 (plastic, 2-fold; optical glass, 2.5-fold). Taking these data together, we conclude that CD13 acts as a physiologic regulator to control the rate of myeloid cell fusion.

FIGURE 1.

Lack of CD13 leads to exaggerated MGC formation in vitro. Flow-sorted live (CD3/B220/Nk1.1)⁻ CD11b^lo/− CD115^hi Ly6G⁺ BM progenitor cells from WT and CD13^KO mice at d3 after M-CSF + IL-4 + IL-13 treatment were analyzed by Giemsa staining (A) (cytoplasm, blue; nuclei, pink/purple) and TG peritoneal macrophages by Giemsa (B) and IF (C) staining. (A) BM-derived progenitor cells differentiated in the presence of M-CSF for 3 d followed by IL-4 + IL-13 for an additional 3 d. (B) TG-elicited WT and CD13^KO peritoneal macrophages at d3 after IL-4 + IL-13 treatment on plastic. (C) IF analysis of TG-elicited WT and CD13^KO peritoneal macrophages at d5 after IL-4 + IL-13 treatment on treated optical glass surface with Mac-3⁺ (red) and DAPI (blue, nuclear stain). Original magnification ×20. (D–F) Percentage fusion index of WT and CD13^KO BM derived MGC at d3 (D), TG-elicited MGC at d3 on plastic (E), and TG-elicited MGC at d5 on treated optical glass (F). Cells with more than three nuclei indicate hypernucleated giant cells. Data represents ±SD of three independent experiments. n = 6/genotype. Scale bar, 50 μm. *p < 0.05, **p < 0.01.

CD13 promotes FBGC formation and fibrosis in vivo

Synthetic mesh implants are frequently used in hernia and other soft tissue repair and often induce macrophage fusion to form FBGCs that are the histologic hallmark of the FBR. Although their histological presence is characteristic of the FBR, it is unclear whether FBGCs directly contribute to the debilitating fibrosis of the FBR or are merely bystanders (8, 9, 14, 15, 46), because abundant cytokines and chemokines, proteases, acids, and free radicals are also present at the implant site (6, 47–49). To determine if CD13 contributes to FBGC formation in vivo and in turn whether FBGCs themselves influence the FBR, we used an in vivo mesh implant model where a small sterile polyethylene mesh implant was inserted s.c. into the back of 8–10-wk-old WT and CD13^KO mice of both sexes. Implants were harvested at d1, d7, and d14, followed by evaluation of structural and inflammatory parameters in FFPE tissue implants. Immunohistochemical analysis indicated that the number of Mac-3⁺ FBGCs, identified as cytoplasmic immunoreactivity to activated mouse macrophage differentiation (Mac-3) Ag with more than three nuclei indicated by white arrows, was significantly increased (>2-fold) in implants isolated from CD13^KO mice compared with WT counterparts at d14 (Fig. 2A–2C), supporting our in vitro MGC fusion data, with a concomitant increase in the circulating levels of the cytokines IL-1β (1.8-fold) and TNF-α (2.2-fold) at d14 compared with d1 after implant (Fig. 2D, 2E), indicating systemic inflammation. Of note, at d7 postimplant, neither the number of Mac-3⁺ FBGCs nor levels of serum cytokines were significantly different between the genotypes. Furthermore, strong CD13 membrane expression was seen in the WT FBGCs (Fig. 2F, 2G) and MGC (Supplemental Fig. 1D–1G) and at the cell–cell contacts, consistent with a functional contribution of CD13 in FBGCs and FBR.

FIGURE 2.

CD13^KO mice show amplified FBGC formation in an in vivo mesh implant model. (A and B) Mac-3⁺ (red) FBGCs with more than three nuclei surrounding the mesh (indicated as I) in WT (A) or CD13^KO (B) mice at d1, d7, and d14 postimplant in a polyethylene mesh implant model. (C) Quantification of Mac-3⁺ FBGCs per field over time indicates enhanced FBGCs in the absence of CD13 at d14 after implant. (D and E) Implants in CD13^KO mice induced elevated serum cytokine levels (IL-1β and TNF-α) compared with WT mice at d14 after implant. (F and G) CD13 (SL13 mAb, red) is highly expressed in WT FBGC membrane (F) but not in CD13^KO FBGCs (G). Data represent ±SD of three independent experiments. n = 5/genotype. Scale bar, 50 μm. *p < 0.05.

Collagen accumulation at the implant site is indicative of tissue fibrosis and advanced-stage FBR (49, 50). ImageJ analysis of Masson’s trichrome–stained implants showed a 1.89-fold increase in peri-implant collagen deposition in CD13^KO compared with WT at d14 postimplantation (Fig. 3A–3C), consistent with the notion that heightened numbers of FBGCs promote the evolution to fibrosis. At the earlier 7-d time point, fibrosis was mild and not significantly different between genotypes. In agreement with enhanced fibrosis, IF analysis of TGF-β and quantification by ImageJ analysis in implants showed that the intensity of profibrotic TGF-β expression in the FBGCs themselves was 1.84-fold higher at CD13^KO implant sites than in the WT counterpart (Fig. 3D, 3E), although, importantly, the number of F480⁺ macrophages remained equivalent between genotypes (Fig. 3F, 3G). We conclude that the CD13-dependent increase in the number of mesh-induced peri-implant FBGCs exacerbates fibrosis and implicates a direct impact of FBGC numbers on FBR progression.

FIGURE 3.

(A and B) Fibrosis is accelerated in CD13^KO in vivo mesh implant model. Masson’s trichrome staining indicates enhanced peri-implant (implant; I) collagen deposition (blue, fibrosis) in CD13^KO mice (B) compared with WT (A) at d14 but not at d1 or d7 after implant. Quantification of the trichrome-stained collagen content (C) per field over time by Fiji analysis. (D and E) Increased TGF-β expression in CD13^KO FBGCs (double white asterisks) at the site of the implant compared with WT as indicated by the intensity of TGF-β in FBGCs/field analyzed by Fiji software. (F and G) IF analysis of implants indicated equivalent numbers of F480⁺ macrophages (red, indicated by white arrows) in the peri-implant region in both genotypes. Original magnification ×20. Scale bar, 50 μm. Data represent ±SD of three independent experiments. n = 3/genotype. *p < 0.05.

CD13 controls MGC fusion by regulating actin protrusion and endocytic vesicle formation

In addition to the fusogenic proteins, cytoskeletal organization and rearrangements are also crucial for MGC formation. Studies have shown that actin filaments are involved in the formation of fusion pores, which allow the exchange of cytoplasmic contents necessary for the cells to merge and form a single cell (12, 24, 51–54). In addition, the processes of receptor endocytosis/recycling and the formation of membrane protrusions are key mechanisms controlling cell–cell fusion (3, 21, 45, 51, 52, 55), as demonstrated by the abundant actin⁺ cell extensions formed in MGCs in response to IL-4 induction (4, 56) and abrogation of macrophage fusion by actin depolymerization (57, 58). Previously, we and others have shown that CD13’s intracellular domain localizes and tethers signaling molecules to cytoskeletal proteins at the membrane during receptor recycling, suggesting that CD13 may directly impact cytoskeletal structures such as actin protrusions (35, 37, 40, 59). Giemsa staining of BM progenitors at the early stages of MGC fusion induction showed that CD13^KO macrophages produced significantly more (2.8-fold; Fig. 4A–4C) and longer (1.7-fold; Fig. 4D) actin protrusions than WT cells as early as d2 postinduction, linking CD13 expression to the formation of actin-based protrusions, as we had previously demonstrated in endothelial cells (60). Interestingly, mononuclear cells appear to attach to MGCs along the length of the protrusions more frequently in CD13^KO than in WT cells (Fig. 4E, 4F), as indicated by phalloidin staining, which may provide clues to why fusion accelerates under conditions supporting actin protrusions. To further establish that these CD13-dependent actin protrusions are responsible for the hyperfusion phenotype, we inhibited actin polymerization with cytochalasin D at d1 after cytokine stimulation and found significantly reduced MGC formation indicated by ∼4-fold reduction in the fusion index in CD13^KO myeloid progenitors compared with vehicle, linking actin dynamics to the hyperfusion phenotype (Supplemental Fig. 1C).

FIGURE 4.

Actin cytoskeletal changes at the prefusion stage drive MGC formation in the absence of CD13. (A and B) Brightfield images of Giemsa-stained WT and CD13^KO flow-sorted BM macrophages grown on plastic indicate more actin protrusions at d2 after M-CSF + IL-4 + IL-13 treatment in CD13^KO (B) compared with WT (A) cells. (C and D) Average number (C) and length (D) of protrusions is significantly higher in the absence of CD13 at d2 after M-CSF + IL-4 + IL-13 treatment. (E and F) IF analysis of phalloidin-stained, TG-elicited WT and CD13^KO peritoneal macrophages grown on treated optical glass for 2 d after IL-4 + IL-13 treatment exhibited enhanced attachment of mononuclear cells along the length of the MGC in CD13^KO compared with WT cells. (G–J) TEM of WT and CD13^KO flow-sorted BM macrophages at d2 after M-CSF + IL-4 + IL-13 treatment indicates increased interaction between mononuclear cells and MGCs (G and H) with microprotrusions (I and J; green arrows) between two neighboring cells in CD13^KO (H and J) compared with WT (G and I) neighboring cells. (K) Quantification of microprotrusions/cell pair ∼1 μm apart. Data represent the average of two isolates and ∼26–33 fields/isolate evaluated by three individuals in a blinded manner. Data represent ±SD of two independent experiments. n = 3/genotype. Scale bars, (A and B) 50 μm; (E and F) 10 μm; (G and H) 10 μm; (I and J) 2 μm. *p < 0.05.

To explore the cell cytoskeletal changes required for proper cell–cell fusion at the organelle scale, we performed high-resolution TEM to determine potential differences in topography, morphology, and arrangement of cellular protrusions (3, 45). TEM has been used previously to study osteoclast but not MGC fusion (61). Because our standard growth and fusion conditions in 96-well dishes were challenging for TEM processing, we developed a method to grow MGCs on Aclar plastic (5 mm², suitable for 96-well dishes) and found that macrophages readily fused under these conditions, enabling EM analysis of sites of cell–cell contact for individual cell protrusions (61–63). Ultrastructural examination of fusing WT or CD13^KO myeloid cells at d2 after fusion (Fig. 4G, 4H) revealed a fusion pattern of mononuclear cells along the length of the CD13^KO MGC, similar to that shown by IF. Interestingly, TEM revealed microprotrusion-like structures of 100–150 nm in width and 1–2 μm in length that join two neighboring cells, which are not visible under a brightfield microscope (≥1 μm apart, observed as indicated by green arrows in Fig. 4I, 4J). Both fused and nonfused structures between two neighboring cells were quantified in 26 WT images and 33 CD13^KO images by three separate individuals in a blinded manner. This analysis indicated a significant increase in microprotrusions between fusing CD13^KO cells compared with the WT counterparts (Fig. 4K), supporting our notion that CD13 impacts cytoskeletal dynamics and protrusion formation during fusion.

CD13 mediates fusogen expression by post-transcriptional mechanisms

The “master fusogen” DC-STAMP and the tetraspanins CD9 and CD81 are common regulators of cell–cell fusion whose expression is transiently induced upon stimulation of MGC fusion, but it must be subsequently reduced to allow fusion to proceed (3, 18, 21, 23, 25, 64–66). Importantly, the sustained expression of fusogens has been shown to accelerate fusion both in vivo and in vitro (30, 41, 67). Indeed, prolonged CD81 expression has been shown to lead to abnormal fusion and MGC formation (68). To monitor fusogen trafficking, we tracked the loss and surface reexpression (recycling) of fusogens in WT and CD13^KO macrophages under fusogenic conditions by surface biotinylation and capture ELISA as described previously (40). Briefly, at d5 after fusion induction, proteins on the surface of TG-elicited peritoneal macrophages from WT and CD13^KO mice were surface biotinylated and allowed to internalize at 37°C for 30 min, followed by chase in complete medium over time to allow labeled proteins to recycle to the surface. At each time point of cell harvest, surface biotin was removed by MesNa treatment to eliminate the recycled fusogens, allowing us to measure internalized biotinylated fusogen levels. Next, cell lysates were incubated in wells coated with anti-CD9 or anti–DC-STAMP Ab, and captured biotinylated fusogen was measured by secondary mAb detection. The relative proportion of recycled CD9 and DC-STAMP at the surface was determined by the relative amount of internalized fusogen in cells during chase. Fusogen recycling was calculated by assuming that the sum of labeled internalized + recycled proteins = 100% and that the degree of recycling of the fusogen to the surface at each time point during chase = 100% minus the percentage of the labeled fusogen remaining inside the cell at each time point. In agreement with the sustained fusogen surface expression in CD13^KO osteoclasts (41), fusogen surface expression over time was significantly enhanced on CD13^KO cells (1.5–1.8-fold), suggesting that CD13 dictates the levels of these key fusion regulatory molecules (Fig. 5A, 5B). Furthermore, the transcript levels of these fusion regulators were equivalent in WT and CD13^KO fusion cultures (Fig. 5C), indicating that CD13 regulated fusogens by post-transcriptional mechanisms. To further verify this concept, we blocked de novo protein synthesis in WT and CD13^KO TG-Macs with CHX, followed by treatment with MG-132, to inhibit both the proteasomal and lysosomal degradation pathways (69). Quantification of surface expression by flow cytometry indicated that blocking protein synthesis for 6 h reduced the abundance of CD9 and DC-STAMP proteins on the surface to a greater extent in WT than in CD13^KO, suggesting that lack of CD13 stabilizes surface expression of DC-STAMP and CD9 by controlling protein turnover via proteasomal and lysosomal degradation pathways (Fig. 5D, 5E, Supplemental Fig. 2A, 2B).

FIGURE 5.

CD13 mediates protein stability via a ubiquitination pathway. (A and B) Surface expression of biotinylated CD9 (A) and DC-STAMP (B) measured by capture ELISA with anti-CD9 and anti–DC-STAMP Abs indicate that the percentage surface fusogen expression over time is higher in CD13^KO than in WT TG peritoneal macrophages at d3 after treatment with M-CSF + IL-4 + IL-13. (C) Expression of fusion-promoting transcripts is equivalent among genotypes. WT and CD13^KO flow-sorted BM progenitor cells were stimulated with M-CSF + IL-4 + IL-13 over time (d0, d3, and d5), and expression of fusion regulatory transcripts was analyzed by quantitative RT-PCR. The relative transcript level normalized to GAPDH indicated equivalent levels of DCSTAMP, CD9, and CD81 genes in both genotypes. (D and E) Flow-sorted BM myeloid progenitors under fusogenic conditions were treated with CHX (protein synthesis inhibitor) (100 μg/ml) ± 5 and 10 μM MG-132 (degradation inhibitor) for 8 h. Flow cytometric analysis of surface mean fluorescence intensity indicated a rapid decay of CD9 and DC-STAMP in WT compared with CD13^KO in the presence of CHX versus CHX + MG132. (F and G) Functionally blocking DC-STAMP and CD81 fusogens rescues abnormal fusion phenotype in CD13^KO indicated by the percentage fusion index. TG WT and CD13^KO peritoneal macrophages treated with blocking DC-STAMP and CD81 mAbs (20 μg/ml) in the presence of IL-4 + IL-13 at d1 were allowed to fuse for 7 d followed by immunohistochemical analysis (F) and measurement of the fusion index (G). IF analysis of Mac-3⁺ cells (red) showed a significant reduction in MGC formation in CD13^KO similar to the level of the WT cells by d7 after treatment. Data represent ±SD of three independent experiments. n = 3/genotype. Scale bar, 10 μm. **p < 0.01, *p < 0.05.

To establish a direct link between CD13-dependent fusogen expression and fusion, we treated WT and CD13^KO peritoneal macrophages with DC-STAMP and CD81 blocking mAbs at 20 μg/ml daily for 7 d after IL-4 + IL-13 addition (68, 70, 71). IF analysis of Mac-3⁺ MGCs revealed that in comparison with isotype control, blocking the fusion-promoting proteins clearly reversed the hyperfusion phenotype in CD13^KO cells to a level similar to WT at d7 after cytokine stimulation (Fig. 5F, 5G). Together, these results confirm that manipulation of fusogens by either fusogen-blocking mAbs or lack of CD13 correlates with the hyperfusion phenotype.

CD13 regulates ubiquitination of fusogens

Although little is known about the post-translational fate and regulation of the fusogens in giant cell formation, it is thought that ubiquitination and degradation of fusion proteins are critical to ensure proper cell–cell fusion into multinucleated cells (72, 73). This process involves the tagging of fusion proteins with ubiquitin, a small protein that marks proteins for degradation by the proteasome (74). Various studies have shown that ubiquitination is a multistep, tightly controlled enzymatic process, and, when dysregulated, it leads to abnormal cell function, including fusion and multinucleated cell formation (75–78). Therefore, to determine if CD13 participates in the ubiquitination pathway, WT and CD13^KO peritoneal macrophages treated with different doses of the degradation inhibitor MG-132 for 3 h were lysed and incubated on a PROTAC assay plate (LifeSensors) for 2 h to capture polyubiquitinated proteins. Bound, polyubiquitinated DC-STAMP or CD9 protein was detected by ELISA and was significantly enhanced (2-fold) in WT cells, resulting in decay of the fusogens compared with CD13^KO cells (Fig. 6A, 6B), indicating that CD13 controls protein levels via a ubiquitination pathway.

FIGURE 6.

Loss of CD13 results in reduced ubiquitination of fusogen and mislocalization of Cullin-4A and CD9. (A and B) WT and CD13^KO TG macrophages were treated with increasing doses of proteasomal degradation inhibitor MG-132 (2.5–10 μM) for 0–3 h. Cells were lysed and subjected to an ELISA-based protein ubiquitination assay using a PROTAC assay plate containing polyubiquitinated capture reagent. Bound, polyubiquitinated target proteins were detected with anti–DC-STAMP (A) or CD9 (B) mAb and indicated reduced ubiquitination of fusogens in CD13^KO compared with WT cells. (C–E) Cullin-4A and CD9 are mislocalized in CD13^KO compared with WT TG-Macs at d3 under fusogenic conditions. Cytoplasmic expression of CUL-4A (red) in WT (C) is markedly altered in CD13^KO (D) myeloid cells undergoing fusion. (E and F) Zoomed images of colocalization of CUL-4A (red) and fusogen CD9 (green) in the WT (arrows) but not in the CD13^KO cells, further confirmed by Pearson’s correlation coefficient (r). (G) Original magnification ×63; oil. DAPI, blue. Data represent ±SD of three independent experiments. n = 3/genotype. Scale bar, 10 µm. *p < 0.05.

CD13 is required for the association and localization of Cullin-4A and fusogens

In unpublished mass spectrometry studies, we had previously determined that the scaffold protein Cullin-4A (CUL-4A) coimmunoprecipitated in a complex with CD13 (not shown) in myeloid cells. CUL4 is a core component of the CRL4 E3 (Cullin RING ubiquitin E3) ligase complex, which is responsible for the recruitment, ubiquitination, and degradation of numerous proteins (79). Defects in the expression and/or localization of CUL4 protein have been shown to compromise ubiquitination and subsequent target protein turnover (80–82). As a scaffold, CD13 organizes molecular complexes by bringing key proteins into close proximity to promote their interactions and situating these complexes at the proper location for optimal function. Therefore, lack of CD13 could disrupt the interaction and localization of key complexes (83). IF analysis to examine potential effects of CD13 on CUL-4A, CD9, and DC-STAMP localization during fusion of WT and CD13^KO TG-elicited peritoneal macrophages indicated that in contrast to the primarily juxtamembrane expression in WT cells, the expression pattern of CUL-4A is more diffuse throughout the cytoplasm in the absence of CD13 (Figs. 6C–6F, 7A–7D). Importantly, colocalization of CUL-4A and CD9 (Pearson’s correlation coefficient, WT versus CD13^KO; r = 0.7 versus 0.3) and of CUL-4A and DC-STAMP (Pearson’s correlation coefficient, WT versus CD13^KO; r = 0.52 versus 0.2) was significantly reduced in CD13^KO cells as compared with WT (Figs. 6D, 6F, 6G, 7B, 7D, 7E), in agreement with CD13 acting as a scaffold to promote the association of the fusogenic proteins and CUL-4A for proper ubiquitination and degradation.

FIGURE 7.

Cullin-4A and DC-STAMP are mislocalized in CD13^KO. (A–D) IF analysis of WT and CD13^KO TG peritoneal macrophages under fusogenic condition at d3 indicated that CUL-4A (red) and DC-STAMP (green) show different expression patterns and do not colocalize in CD13^KO (C and D; zoomed image) compared with WT (A and B; zoomed image) cells, confirmed by Pearson’s correlation coefficient (r). (E) Original magnification ×63; oil. DAPI, blue. Data represent ±SD of three independent experiments. n = 3/genotype. Scale bar, 10 µm. *p < 0.05. (F) Anpep (CD13) is one of the most highly induced genes in FBGCs (GC) compared with immediate precursor (IP) and M2 macrophages (M2) by transcriptome analysis of scRNA-seq in mesh-induced FBR.

CD13, ubiquitination, and actin binding protein transcripts are enriched in FBGCs

In a recent study, single-cell RNA-sequencing (scRNA-seq) analysis of cells at sites of mesh-induced FBR showed that Actn1 [α-actinin, CD13-interacting cytoskeletal protein (37)], Dcstamp (fusogen), and Uchl1 (ubiquitin C-terminal hydrolase) are highly expressed in FBGCs compared with M2 macrophages or precursors (12). Importantly, further mining of the RNA-seq data revealed that Anpep (CD13) is highly expressed in giant cells compared with the precursor populations (Fig. 7F), together supporting our results that CD13-dependent mechanisms regulating ubiquitination and cytoskeletal rearrangement pathways are important in FBGC formation. Taking these data together, we conclude that CD13 is a critical regulator of myeloid cell fusion, both in the formation of MGCs in vitro and FBGC formation in vivo (this study) and in osteoclast fusion in vitro and in vivo (41). Mechanistically, CD13 limits formation of microprotrusions and actin-based protrusions at the prefusion stage while facilitating the association of fusogens with the CRL4 complex during fusion. Therefore, loss of CD13 leads to hyperfusion, exaggerated microprotrusions between neighboring cells, compromised ubiquitination, and subsequent degradation of fusogens by the CRL4 complex, contributing to accelerated fusion (Fig. 8). Together, our results suggest that CD13 may be a therapeutic target to control the damaging inflammation observed at sites of implantation and may prevent implant failure.

FIGURE 8.

Schematic of CD13-mediated giant cell fusion by a post-transcriptional regulation of fusogens. Loss of CD13 results in increased protrusion and microprojection formation between neighboring cells but comparable fusogen expression among genotypes at the prefusion stage (A and B), with reduced fusogen ubiquitination/degradation during fusion (C and D) augmenting fusogen surface expression and ultimately forcing aberrant multinucleated giant cell formation (E and F).

Discussion

Implantable medical devices have become invaluable tools to prevent, monitor, treat, diagnose, or alleviate numerous health conditions and serve to improve our overall quality of life (1). Not surprisingly, these devices compose a large and rapidly growing sector of the health care market, and, considering our aging population, their demand will only increase. Implantable devices span all medical specialties and include artificial joints, heart valves, cardiac pacemakers, biologic meshes, cardiac and glucose sensors, insulin infusion pumps, breast, dental and cochlear implants, and intraocular lenses, among hundreds of others. These implants invariably invoke a deleterious response from the host that can limit diffusion of nutrients, drugs, or analytes; impair communication regarding function; and lead to tissue destruction and device failure (84). Published studies implicate the FBR in the failure of 65% of cochlear implants (85, 86), 20–40% of intraocular lens implants (87), and surgical mesh implants (88, 89), and breast implants have been linked to consequent inflammation-related leukemias (90, 91). Clearly, the FBR is a fundamental barrier that must be surmounted, compelling a focus on uncovering the essential regulatory mechanisms and novel molecules and pathways that represent tomorrow’s therapeutic targets (84). Controlling the FBR is crucial for any further refinement of the performance, biocompatibility, and longevity of implanted medical devices, as well as the success of emerging technologies such as implanted cellular therapies necessary for critical advancement and improvement of life-enhancing medical devices (92).

Monocytes and macrophages have many functions, ranging from phagocytosis of invading pathogens, triggering the adaptive immune response, to regulating lipid and iron metabolism (93). Certain subsets of macrophages are able to fuse their membranes, forming syncytial MGCs that share a common cytoplasm and plasma membrane (94). The three primary MGC subsets are osteoclasts, FBGCs, and Langhans giant cells (95), which represent distinct lineages and result from distinct signaling pathways. Recently, it has been reported that macrophages and tumor cells fuse to form tumor–macrophage fusion cells in the circulation of patients with metastatic cancer (96). In addition, M2 macrophage and tumor cell fusion gives rise to radioresistant cancer cells in solid tumors (97), underscoring that pathologic cell–cell fusion may occur in numerous contexts. Osteoclast fusion is regulated primarily by the receptor activator of NF-κB ligand (RANKL) signaling pathway. RANKL is produced by osteoblasts and stromal cells and binds to its receptor RANK on preosteoclasts, promoting their fusion into multinucleated osteoclasts. Although osteoclast fusion is a physiological process that enables bone remodeling by breaking down and resorbing bone tissue, dysregulated osteoclast activity has been associated with osteoporotic bone loss and osteolysis in multiple myeloma–induced bone resorption (98). Foreign body–induced fusion is a pathological process triggered by the presence of foreign materials that activate macrophages to fuse and form FBGCs in an attempt to phagocytose and eliminate foreign particles (7, 99, 100). Formation of these lineages can be recapitulated in vitro by stimulation of BM progenitors with appropriate cytokines, RANKL for osteoclasts and IL-4 and IL-13 for FBGC-like MGCs (44). Although osteoclast fusion has been studied intensely, little is known about fusion in other MGC subtypes (11). However, because aseptic loosening is the basis for revision surgery in nearly 55% of hip revisions (101) and 60.2% of knee revisions (102), it is imperative to optimize primary outcomes in implantable devices by controlling the FBR.

The mechanism of the FBR depends on the physical and biochemical properties of the implant, the implant site, and host immune responses (50). It is a complex, multifactorial process, and, to date, which components of the FBR are causal and/or predictive of implant failure remain unresolved (8, 9, 103). Despite the fact that macrophages are key players in inflammation, studies focusing on preventing FBR at the stage of inflammation mediated by FBGCs are lacking, partly due to some ambiguity regarding the precise contribution of FBGCs and the controversies surrounding their role in FBR. Numerous investigations have indicated that FBGCs contribute to the degradation of foreign particles through secretion of ROS, MMPs, and acids (14) and are a source of cytokines and chemokines (15, 16) and ROS (10–13). Alternatively, two studies found that FBGCs were dispensable for FBR-induced encapsulation (8, 9) where both CCL-2– and MMP-9–null mice had reduced FBGC numbers but normal capsule thickness, although MMP-9–null mice reportedly showed defects in collagen deposition and assembly, illustrating the complexity of the FBR. The increased number of FBGCs at the peri-implant site in our CD13^KO model is accompanied by increased collagen deposition and amplified intracellular and systemic TGF-β levels, hallmarks of fibrosis and indicative of a more intense FBR. This would indicate that FBGCs are a critical and causal component of FBR progression in vivo and supports CD13 as a potential target in minimizing the deleterious effects of the FBR. Alternatively, it is possible that a lack of CD13 could alter the FBGC phenotype, perhaps making it more aggressive, which also leads to exacerbated fibrosis independent of FBGC numbers. Further analysis to uncover potential differences in the transcriptional profiles of the cell types at the implant site in WT versus CD13^KO mice by scRNA-seq is currently underway in our laboratory to address this possibility.

Cell–cell fusion is tightly controlled at the molecular level, involving the coordination of multiple factors, some that are shared and others specific to the cell and tissue type. Although various cell types and biological events involve cell fusion, the process has yet to be reduced to a well-ordered universal scheme. However, all fusion requires the initial steps of cell migration toward neighboring cells; organization, recognition, and engagement of membrane adhesion molecules (104); and the assembly of specialized actin-based protrusions extending between cells that colocalize with actin polymerization proteins (105). These steps also depend on the spatial and proximate coordination of components of key signaling pathways and their target adhesion molecules, cell surface receptors, intracellular kinases, and transcription factors, together culminating in the fusion event (106). We have shown that CD13 is a cell–cell adhesion molecule (35–37, 107), regulates cell migration (40, 60), controls the assembly of actin structures (40, 108), and limits osteoclast fusion (41) and thus could contribute to FBGC fusion and FBR at a number of steps. However, we have also demonstrated that in some of these cases, CD13 acts as a transmembrane anchor to bind and properly position signaling complexes in the correct membrane context to enable subsequent cellular functions. Anchor proteins in turn often bind to scaffold proteins that assemble various signaling components into functional signaling complexes. We have shown that CD13 promoted ARF6 GTPase activity by binding to the scaffold IQGAP and positioning the IQGAP/ARF6/EFA6 signaling complex at the leading edge of the cell membrane, promoting ARF6 GTPase cycling and cell migration (40). In this regard, we consider CD13 as a transmembrane anchor that coordinates various signaling steps that allow fusion to occur. This is supported by the clear mislocalization of the CUL4 and CD9 proteins in CD13^KO FBGCs, suggesting that CD13-dependent membrane tethering of the CUL4 complex is required for fusogen degradation and normal cell–cell fusion.

Cullin-RING ubiquitin ligases (CRLs) are protein complexes that regulate various biological processes by ubiquitination, the post-translational attachment of the ubiquitin protein to lysine residues on target proteins, primarily marking them for degradation. CRLs contain multiple core subunits: a Cullin protein scaffold, a RING domain-containing catalytic subunit that transfers ubiquitin to the target protein, and an interchangeable substrate recognition (SR) module that confers ubiquitination specificity by recruiting specific target proteins to the CRL complex. CRLs are activated by neddylation, the attachment of the protein NEDD8 to a specific lysine residue on the Cullin subunit, which enhances the CRL ubiquitin ligase activity. Conversely, CRL activation is blocked by the Cullin-associated NEDD8-dissociated (CAND) proteins. Specifically, CAND1 binds to CRL complexes and blocks the neddylation site, thus preventing the activation of CRLs and the binding of SRs. In this study, we found that fusogen protein expression was prolonged in cells lacking CD13, which mechanistically was due to decreased fusogen ubiquitination in CD13^KO cells. We had previously uncovered a potential link between CD13 and the ubiquitin ligase machinery in a proteomic screen of CD13-associated proteins where the Cullin scaffold protein (Cullin 4A), an SR module (DDB1), CAND1 (Cullin-associated NEDD8-dissociated 1), and the E1 ligase UBA1 were present in a complex coimmunoprecipitated with CD13. The mislocalization of Cullin-4A and the CD9 fusogen in the absence of CD13 supports this finding and suggests that CD13 may play a role both in the assembly of the CRL4-fusogen complex and in directing it to the proper intracellular site for ubiquitination and subsequent degradation. Alternatively, it is possible that the presence of both the inhibitory CAND1 and the substrate recruiter DDB1 in the Cullin-4A and CD13 complex signifies that both active (DDB1-bound) and inactive (CAND1-bound) CRL4 associate with CD13, but CD13 does not impact complex assembly and only guides the CRL to the proper position to promote subsequent processes. This interesting hypothesis is currently under investigation in our laboratory.

This article is featured in Top Reads, p. 503

Disclosures

The authors have a financial conflict of interest in that the SL13 mAb has been licensed to EMD Millipore.