Inflammasome Components Asc and Caspase-1 Mediate Biomaterial

Inﬂammasome components Asc and caspase-1 mediate biomaterial-induced inﬂammation and foreign body response

Ahsan F. Malika, Rafaz Hoquea, Xinshou Ouyanga, Ayaz Ghania, Enping Hongb, Khadija Khanb, Laura Beth Mooreb, Gilbert Ngc, Fay Munroc, Richard A. Flavelld, Yan Shic, Themis R. Kyriakidesb,1, and Wajahat Z. Mehala,d,e,1

aSection of Digestive Diseases and bDepartments of Pathology and Biomedical Engineering, Yale University, New Haven, CT 06520; cDepartment of Microbiology, Immunology, and Infectious Diseases, University of Calgary, Calgary, AB, Canada T2N 1N4; dDepartment of Immunobiology, Yale University, New Haven, CT; and eSection of Digestive Diseases, Department of Veterans Affairs Connecticut Healthcare, West Haven, CT 06516

Edited by Shizuo Akira, Osaka University, Osaka, Japan, and approved October 21, 2011 (received for review April 5, 2011) Implantation of biomaterials and devices into soft tissues leads to complement has been shown to occur and enhance biomaterial– the development of the foreign body response (FBR), which can inflammatory cell interactions (6, 7). However, modulation of interfere with implant function and eventually lead to failure. The these interactions has not been shown to lead to long-lasting at- FBR consists of overlapping acute and persistent inflammatory tenuation of the FBR. phases coupled with collagenous encapsulation and currently In addition to macrophages, dendritic cells (DCs) have been there are no therapeutic options. Initiation of the FBR involves implicated in the foreign body response, primarily due to a com- macrophage activation, proceeding to giant cell formation, fibro- bination of products that present antigenic stimuli (8). Moreover, blast activation, and collagen matrix deposition. Despite the an in vitro study has shown that DC–biomaterial interactions can recognition of this sequence of events, the molecular pathways occur via engagement of multiple Toll-like receptors (TLRs) and required for the FBR have not been elucidated. We have identified lead to significant induction of IL-6 and regulated upon activa- that the acute inflammatory response to biomaterials requires tion, normal T-cell expressed, and secreted (RANTES) and mild nucleotide-binding domain and leucine-rich repeat-containing 3 induction of IL-1β and TNF-α (9). These observations raise the (Nlrp3), apoptosis-associated speck-like protein containing CARD possibility that DCs can also participate in the recognition of (Asc), and caspase-1, as well as plasma membrane cholesterol, biomaterials and serve as stimulators of the foreign body re- and Syk signaling. Full development of the FBR is dependent on sponse. However, to date the presence of DCs at the tissue– Asc and caspase-1, but not Nlrp3. The common antiinflammatory biomaterial interface in vivo has not been documented (8, 9). drug aspirin can reduce inflammasome activation and significantly Nevertheless, it is possible that DCs play a critical role in mod- reduce the FBR. Taken together, these findings expand the role of ulating cross-talk between innate and adaptive immunity, espe- the inflammasome from one of sensing damage associated molec- cially when engineered constructs contain immunogenic signals. ular patterns (DAMPs) to sensing all particulate matter irrespective There has been rapid recognition of the role of a set of cy- of size. In addition, implication of the inflammasome in biomate- tosolic proteins termed the inflammasomes in initiation of the fi rial recognition identi es key pathways, which can be targeted to inflammatory response to crystalline materials naturally found limit the FBR. in vivo, including uric acid and cholesterol (10–12). In addition to these biologically formed materials, other small particulates he use of biomaterials is an established part of medical such as alum, silica, asbestos, and nanoparticles have been shown Tpractice and such materials range from a single material such to result in inflammasome activation (13–15). These materials as silicone for breast implants to combinations of materials such have very different physical characteristics but are small enough as in sensors for measuring glucose concentration (1). The utility to be phagocytosed, and subsequent phagosome rupture has MEDICAL SCIENCES of implants and devices using biomaterials is limited, due to the been shown to result in inflammasome activation and production development of the foreign body reaction (FBR), which is initially of IL-1β and IL-18 (16). An alternative mechanism of inflam- an acute sterile inflammatory response, subsequently overlapp- masome activation by particulate matter has recently been ing with a chronic fibrotic response (2). Hallmarks of the FBR demonstrated. This mechanism depends on reorganization of include accumulation of macrophages at the tissue–implant in- cholesterol rafts by interaction of the particulate matter with the terface, formation of foreign body giant cells (FBGCs), and de- plasma membrane, resulting in Syk activation. The presence of position of a dense layer of collagenous matrix that isolates the an inflammasome-activating pathway independent of phagocy- implant. The clinical consequences of the FBR include pain, tosis prompted us to test the ability of biomaterials to activate the scarring, and for some devices such as glucose sensors, device inflammasome (17). In addition, we investigated the dependence failure due the development of fibrous encapsulation. Macro- of the FBR on individual components of the inflammasome. Using phage activation and fusion have been identified as critical cel- biomaterials that are too large to be phagocytosed, we investigated lular events in the FBR and recent studies have identified key molecular events in the formation of FBGCs, including induction of E-cadherin, Rac1 activation, and secretion of matrix metal- Author contributions: R.A.F., T.R.K., and W.Z.M. designed research; A.F.M., R.H., X.O., loproteinase-9 (MMP-9) (3). However, the initial critical events A.G., E.H., K.K., L.B.M., G.N., F.M., Y.S., and T.R.K. performed research; A.F.M., R.H., in macrophage–biomaterial interactions and the elicited down- X.O., A.G., E.H., K.K., L.B.M., G.N., F.M., R.A.F., Y.S., T.R.K., and W.Z.M. analyzed data; stream intracellular events have not been identified. Our current and A.F.M., T.R.K., and W.Z.M. wrote the paper. Conflict of interest statement: W.Z.M. and T.R.K. are co-inventors on a patent filed by understanding of this process involves surface adsorption of Yale University, which is under license on the use of aspirin to reduce the foreign proteins present in edematous interstitial fluid, such as fibrino- body reaction. gen, which causes their denaturation and renders them adhesive This article is a PNAS Direct Submission. fl for in ammatory cells (4). For example, exposure of the cryptic 1To whom correspondence may be addressed. E-mail: [email protected] or themis. integrin-specific epitopes P1 and P2 in fibrinogen has been shown [email protected]. fl fl to in uence the accumulation of in ammatory cells in short-term This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. in vivo studies (5). In addition, surface-induced activation of 1073/pnas.1105152108/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1105152108 PNAS | December 13, 2011 | vol. 108 | no. 50 | 20095–20100 Downloaded by guest on September 23, 2021 the role of the inflammasome in phagocytosis-independent, resulting in nonspecific cytosolic aggregation of immunoreceptor biomaterial-induced inflammation and the FBR. tyrosine-based activation motif (ITAM) signaling domains and activation of a Syk kinase pathway (17). The involvement of this Results mechanism can be examined by depleting plasma cholesterol Inflammasome Components Nlrp3, Asc, and Caspase-1 Are Required with methyl-β-cyclodextrin (MβCD) or by inhibiting Syk activa- for the Acute Inflammatory Response to Biomaterials. To investigate tion with piceatannol. To do this an in vitro system was estab- the acute inflammatory response to biomaterials we injected lished, using peritoneal cells isolated from thioglycolate-primed sterile microspheres of poly(methyl methacrylate) (PMMA) with mice. Inflammasome activation was assayed by quantifying IL-1β a mean diameter of 153 μm into the peritoneal cavity of wild- in the supernatant by ELISA. IL-1β production was not induced type mice and quantified the inflammatory infiltrate 24 h later. in cells treated with the vehicle, microspheres, or LPS alone. The cell count in the lavage fluid of mice injected with micro- Microspheres in addition to LPS resulted in a significant rise in spheres was 5.09 × 107, which is significantly greater (P < 0.05) IL-1β levels, which was reduced in Nlrp3 KO macrophages and than of control mice injected with vehicle only (8.79 × 106 cells), by MbCD or piceatannol (Fig. 2B). These results suggest that the demonstrating an acute inflammatory response (Fig. 1A). The physical contact of the microspheres with the plasma membrane same experiment was repeated in mice with genetic deletion of is the vital interaction for inflammasome activation. Confirma- nucleotide-binding domain and leucine-rich repeat-containing tion for this suggestion was obtained by measuring the binding 3 (Nlrp3), apoptosis-associated speck-like protein containing force of a single cell to a single microsphere using atomic force CARD (Asc), and caspase-1. In comparison with wild-type mice, microscopy (AFM). The experimental design of the physical mice lacking Nlrp3, Asc, or caspase-1 displayed reduced cell interaction between the microsphere and the cell (Fig. 2C) and numbers that were similar to control mice injected with vehicle an actual microsphere on a cantilever (Fig. 2D) is shown. After (Fig. 1 A–C). Findings in mice deficient in Nlrp3, Asc, or cas- ∼50 s of oscillating contact, there was a dramatic increase in the pase-1 suggest that inflammasome activation plays a role in the binding force between the cell and the microsphere, which initial phases of the inflammatory response to biomaterials. Flow fluctuated over the duration of the experiment (Fig. 2 E and F). cytometric analysis of the peritoneal lavage fluid showed that in As predicted by the ELISA data (Fig. 2B), there was a significant wild-type mice injected with microspheres there was a relative decrease in the binding force of the cells to the microsphere in increase in Gr1+ neutrophils, in comparison with F4/80 macro- the presence of MbCD or piceatannol. phages, although both cell populations had an absolute increase in numbers (Fig. 1D). In mice lacking Nlrp3 or caspase-1, the Inflammasome Components Asc and Caspase-1, but Not Nlrp3 or relative increase in Gr1+ neutrophils was significantly reduced NLRC4, Are Required for the Development of the Foreign Body (Fig. 1E). To directly confirm the occurrence of caspase-1 acti- Reaction. To confirm that the role of the inflammasome was vation in response to biomaterials, we performed Western blot not limited to one biomaterial, and to test the role of inflam- analysis of peritoneal cell lysates from wild-type and caspase-1 masome components in the development of the FBR, we used KO mice injected with microspheres, which showed cleavage of the well-established model of s.c. silicone implantation. Silicone caspase-1 in the former (Fig. 1F). We further tested caspase-1 disks (6 mm diameter) were implanted for 4 wk, excised en bloc, activation in Nlrp3 and Asc KO mice and found that caspase-1 and processed for histological analysis. A central feature of the cleavage did occur, although to a much lesser degree than in FBR is the accumulation of macrophages and fusion into FBGCs wild-type mice. This suggests that although Nlrp3 and Asc play that surround the foreign body. Formation of FBGCs adjacent a required role, additional molecules proximal to caspase-1 can to the silicone implants was similar in wild-type, Nlrp3, Asc, or respond to biomaterials and participate in the partial activation caspase-1 KO mice (Fig. 3A). Immunohistochemical detection of caspase-1. of macrophages revealed that their presence was not reduced in the Nlrp3, Asc, or caspase-1 KO mice (Fig. 3 B and C). These Inflammasome Activation by the Microspheres Can Be Mediated findings demonstrate that the absence of these inflammasome Through Clustering of Lipid Rafts. Internalization of crystals has components does not inhibit macrophage localization and FBGC been demonstrated as one mechanism of inflammasome activa- formation in the FBR. tion, however the microspheres used in the present study were The thickness of the fibrotic reaction in the FBR is a clinically too large to be phagocytosed (Fig. 2A). Inflammasome activation relevant parameter as it can limit the longevity of biomaterials can also occur by membrane affinity triggered signaling (MATS), and complicate the function of medical devices. In contrast to which involves clustering of lipid rafts at the site of contact, the preservation of FBGC formation and macrophage density,

A † B † C

) * Vehicle * †

6 50 50 50 * ) Beads † ) 6 40 6 † 40 40 (X 10 30 (X 10 30 30 † 20 20 * 20 * 10 * 10 10 Fig. 1. PMMA microspheres cause an acute in-

Cell Counts (X 10 0 Cell Count n=4 n=6 n=6 n=7 Cell Count 0 n=6 n=13 n=7 n=13 0 n=8 n=10 n=2 n=8 flammatory infiltrate, which is dependent on WT Nlrp3 ko WT ASC ko WT Casp1 ko Nlrp3, Asc, and caspase-1. (A–C)Reductionin peritoneal infiltrate in Nlrp3, Asc,andcaspase-1 D E F KO mice compared with wild-type (WT) controls. WT + Vehicle WT+ Beads Nlrp3 ko + Beads Casp1 ko +Beads GR1+ cells F4/80+ cells (D) Relative increase in neutrophils in WT mice WT

ASC ko kDa in response to microspheres. (E)Reducedneu- + Casp1ko 4.2% NLRP3 ko 2.8% 20.2 % 11.9 % 60% caspase 10 trophil inﬁltrate in Nlrp3 KO and caspase-1 GR1 Cells 40% β actin 43 KO mice compared with WT. (F) Western blot * †‡ 20% ‡ showing speciﬁc caspase-1 activation in protein

+ 66.4% 56.1% † 49.8% % of Cells * 57.1% 0% lysates from the inﬂammatory inﬁltrate elicited

F4/80 Cells WT + WT + Nlrp3 ko Casp1 ko by microspheres from wild-type mice, with re- †‡ Vehicle Beads + Beads + Beads duction in Nlrp3 and Asc KO mice. * P ≤ 0.05.

20096 | www.pnas.org/cgi/doi/10.1073/pnas.1105152108 Malik et al. Downloaded by guest on September 23, 2021 Φ A B 900 NLRP3ko Piceatannol(50mM) 750 MbCD (2 mM) 600 †‡Φ 200 * ‡ 20µm † 100 *

C Cantilever 0 Epoxy Vehicle + + - + + + + + - Bead Beads - + -- + + + + - Macrophage LPS - - +++ + + + + Glass Disk ATP ------+ D E Untreated (7) Fig. 2. PMMA microspheres induce IL-1β MbCD (4) production from peritoneal macrophages in 8 a membrane lipid- and Syk-dependent man- 3 Piceatannol (2) ner. (A) Single microsphere, which is much 7 larger than attached macrophages and can- 6 not be phagocytosed. (B) Microspheres in- β ﬁ 500μm duce IL-1 , which is signi cantly less in the 5 absence of Nlrp3 and can be reduced in wild- type macrophages by the cholesterol in- 4 hibitor MbCD (50 μm), and the Syk inhibitor F † piceatannol (2 mM). (C) Experimental design 2 * 3 3 for measuring the binding force between

Binding Force (pN) x 10 Binding a single microsphere and a single macro- 1.5 2 * phage using an AFM. (D) Single microsphere 1 1 on the cantilever of the AFM. (E) Binding force between a single microsphere and cell 0 0.5 over a period in the presence and absence of † 0 100 200 300 Time MbCD or piceatannol. (F) Readings of binding 0 ﬁ Binding Force (pN) x x 10 (pN) Force Binding (sec) af nity in all experiments depicted in the E †‡ϕ Untreated MbCD Piceatannol are averaged. * P ≤ 0.05.

the development of the FBR was significantly reduced in mice the drinking water was able to almost completely inhibit the deficient in Asc or caspase-1 (Fig. 3 D–F). Histological analysis acute peritoneal infiltrate 24 h after injection of microspheres showed significantly reduced collagen deposition in Asc or cas- (Fig. 4A). We then sought to determine whether aspirin had pase-1 KO mice, but not in Nlrp3 KO mice. The mean capsule a direct effect on macrophages. At 60 mg/L aspirin in the thickness in wild-type mice on the dermal side was 160 ± 10.2 drinking water the average serum concentration was 62 μM. MEDICAL SCIENCES μm, and 84.3 ± 5.7 μm in caspase-1 KO, 81.5 ± 8.6 μm in Asc KO Using LPS/bead-activated peritoneal macrophages, we evalu- and 163.7 ± 4.7 μm in Nlrp3 KO mice. For the side facing the ated the ability of 10–150 μM aspirin to inhibit IL-1β production body wall (Fig. 3D), the thickness was 44.6 ± 7.5 μm in wild-type and found that, even at 10 μM, aspirin was effective (Fig. 4B). mice, and 20.8 ± 2.8 μm in caspase-1 KO, 17 ± 3.8 μm in Asc To further test whether aspirin can limit the FBR, mice were KO, and 42 ± 5.5 μm in Nlrp3 KO mice. It is interesting to note continuously placed on drinking water containing aspirin at that although Nlrp3 was required for the early inflammatory a concentration of 60 mg/L immediately after implantation of response to foreign materials, for the full chronic FBR either the silicone discs. Consistent with our observations in mice Nlrp3 has no role, or there is redundancy with other NLRs that lacking inflammasome components, the formation of FBGCs can compensate in the absence of Nlrp3. Because NLRC4 is was not impaired by aspirin (Fig. 4C). However, collagen de- known to be proximal to Asc, we investigated the FBR in position, measured as capsule thickness, was significantly re- NLRC4 KO mice and found that it was normal (Fig. S1). duced by aspirin (Fig. 4 D–F). The mean FBR in WT mice on thedermalsidewas98± 15.1 μm, and 48.6 ± 6.7 μm in mice on Aspirin Can Reduce the Acute Inflammation and the Foreign Body aspirin. In juxtaposition to the body wall the thickness was 45.8 ± Response. Aspirin is widely recognized to have antiinflammatory 6.7 μm in WT and 17.3 ± 3.3 μm in mice on aspirin. This was properties. In addition to antiplatelet effects, and its ability to comparable to the reduction of the FBR in mice lacking Asc inhibit production of prostaglandins by Cox-1 and prostanoids or caspase-1. by Cox-2, aspirin also increases production of the potent anti- To determine whether either of the two signaling pathways inflammatory molecules lipoxins (18, 19). We have previously required for the production of IL-1β is reduced by aspirin, the shown that aspirin can inhibit the inflammasome pathway and degree of caspase-1 activation and pro–IL-1β up-regulation in- reduce production of IL-1β. In view of this inhibition by aspirin, duced by PMMA microspheres in vivo was determined. As and the above demonstration that acute inflammation to foreign shown in Fig. S2, there was significantly less activation of cas- materials and the FBR are dependent on the inflammasome pase-1 by PMMA microspheres in the peritoneal infiltrate of pathway, we tested the ability of aspirin to modulate these mice treated with aspirin in the drinking water (60 mg/mL). In responses. Aspirin at a concentration of 60 mg/L or 0.6 mg/L in contrast, there was no reduction in the up-regulation of pro–IL-

Malik et al. PNAS | December 13, 2011 | vol. 108 | no. 50 | 20097 Downloaded by guest on September 23, 2021 A B C 80

60 WT WT Nlrp3 ko 40 Macrophage counts/hpf 20

0 WT Nlrp3 ASC Casp1 ASC ko Casp1 ko Casp1 ko ko ko ko D

DERMAL SIDE

E BODY Fig. 3. Macrophage recruitment and WALL foreign body giant cell formation is in- tact in the absence of Nlrp3, Asc, or caspase-1, but the FBR is reduced. (A) WT Nlrp3 ko ASC ko Caspase-1 ko Formation of foreign body giant cells stained “reddish” with dark nuclei adjacent to the silicone is normal in Nlrp3 KO, * † F G * † Asc KO, and caspase-1 KO mice. (B and C) 150 50 The density of tissue macrophages is 120 † 40 unchanged in Nlrp3 KO, Asc KO, and 90 * 30 caspase-1 KO mice compared with WT. * † – 60 20 (D G) Capsule thickness and collagen of ﬁ 30 10 the FBR stained blue is signi cantly reduced in Asc KO and caspase-1 KO but 0 0 not Nlrp3 KO mice compared with WT. † WT Casp1 ko ASC ko Nlrp3 ko WT Casp1 ko ASC ko Nlrp3 ko * P ≤ 0.05.

1β expression. The use of selective Cox-1 and Cox-2 inhibitors Recently, a study using atomic force microscopy described a individually, or in combination, did not reduce IL-1β production complementary model where extracellular particles activate the by microsphere-stimulated macrophages (Fig. S3). inflammasome via physical interaction with the plasma membrane, resulting in reconfiguration of plasma membrane choles- Discussion terol and colocalization and activation of cytoplasmic ITAMS Foreign bodies that come in contact with vascularized tissues and Syk activation. We considered this MATS model relevant for elicit a FBR that progresses through a stereotypical sequence of biomaterials because in most applications biomaterials are too events involving activation of complement and deposition of large to be phagocytosed. Consistent with this hypothesis, we plasma proteins, followed by a recruitment of neutrophils and observed that microspheres can activate the inflammasome and macrophages. The latter adhere to the surfaces of foreign bodies induce an inflammatory infiltrate that is dependent on the and form FBGCs (2). Over a period of weeks to months, the fl fi in ammasome components Nlrp3, Asc, and caspase-1. These are response progresses to recruitment of broblasts that deposit the same components that are required for initiation of in- collagen, leading to encapsulation of the foreign body. Despite flammation to microscopic particulate matter, and the inflam- the established sequence of events outlined above, the molecular masome appears to be a common pathway for initiating inflam- components required for sensing the foreign material and initi- mation in response to physical materials irrespective of their size. ating the FBR have not been identified. In the present study we To confirm that the described mechanism of perturbation of show that biomaterials trigger activation of the inflammasome β plasma membrane cholesterol was involved in this process we leading to generation of active caspase-1 and secretion of IL-1 . β Our findings are consistent with previous demonstrations show- show that IL-1 production can be inhibited by MbCD and also ing that the inflammasome is required for the sterile inflamma- by the Syk inhibitor piceatannol. These results suggested these tory response to a wide range of particulate matter including agents should ameliorate the binding force between microspheres uric acid, silica, and alum. These particulates have a very dif- and the plasma membrane. This was tested by atomic force mi- fi ferent structure from each other, but have in common the fact croscopy and both MbCD and piceatannol signi cantly reduced that they are small enough to be phagocytosed. In fact an im- the binding force. More importantly, these studies were per- portant mechanism of inflammasome activation is phagocytosis formed in serum-free media suggesting that the microspheres with subsequent rupture of the phagosome (16). These findings, could not acquire a proteinaceous coat. Thus, in this model we and the proposed phagosome-dependent model, have resulted envision direct interaction of macrophages with the biomaterial in the inflammasome being identified as a sensor of microscopic surface. Interestingly, microparticles like alum can also induce particulate matter. However, our data clearly show that 150-μM syk activation in DCs, which raises the possibility that these cells microspheres can activate the inflammasome and induce an in- have a similar ability to respond to biomaterials (20). However, flammatory infiltrate. Thus, we presume that the microspheres the DC response to alum was shown to be caspase-1 independent, trigger activation in the absence of phagosomal lysis. implicating the involvement of a separate pathway.

20098 | www.pnas.org/cgi/doi/10.1073/pnas.1105152108 Malik et al. Downloaded by guest on September 23, 2021 A * †‡ B * †‡ C ) 200 6 50 † ‡ 40 WT (X 10 150 30 pg/ml)

20 ( * β 100 10 † ‡ 0 IL-1 50 WT+ASA Cell Count n=4 n=8 n=4 n=8 * Water Plain Plain ASA ASA 0 + + + + ASA in Water --0.6mg/L 60mg/L LPS+Beads Beads - + + + ASA - 10μM 60μM 150μM D F 120 * DERMAL Fig. 4. Oral administration of aspirin re- SIDE 90 duces the acute inflammatory response to 60 * † biomaterials and the FBR. (A) Aspirin reduces 30 † the peritoneal acute inflammatory infiltrate elicited by microspheres. (B) Aspirin reduces

E Capsule thickness (μm) 0 IL-1β production by macrophages in response BODY WALL WT WT+ASA WT WT+ASA to LPS/beads. (C) Aspirin does not inhibit DERMIS BODY WALL foreign body giant cell formation in response to s.c. silicone implants. (D–F) Aspirin reduces the FBR in response to s.c. silicone †‡ WT WT + ASA implants. * P ≤ 0.05.

We find the involvement of syk signaling in biomaterial rec- unilateral ureteral obstruction, and bleomycin (BLM)-induced ognition by macrophages intriguing because the interaction with lung injury (24–27). biomaterials leads to macrophage fusion and formation of We had previously shown that continuous administration of FBGCs. In fact, syk has also been shown to be required for fu- aspirin could inhibit the archetypal inflammasome-dependent sion as Syk KO macrophages do not fuse in response to IL-4/ inflammation by uric acid crystals (28). Based on our findings GM-CSF (21). Similarly, the fusion of monocytes/macrophages regarding the requirement for Asc and caspase-1 in the FBR, we to form osteoclasts on bone surfaces involves Syk (22). Specifi- asked whether aspirin could inhibit the acute inflammatory re- cally, studies have linked Syk signaling in osteoclast maturation sponse to microspheres. As we demonstrate here, administration fi and it is conceivable that in addition to production of proin- of aspirin in the drinking water signi cantly reduced the peri- fl fi β flammatory signals, activation of syk interfaces with the fusogenic toneal in ammatory in ltrate, and also reduced IL1 production fi machinery. by macrophages in vitro, indicating that the effect was speci c for fl these cells. Of the two main pathways required for the pro- The in ammatory response to a foreign material examined by β the above experiments with microspheres involved the early duction of IL-1 , we have demonstrated a reduction in caspase-1 phase of the FBR. To study the long-term role of the inflam- activation by aspirin. Aspirin has a number of complex in vivo effects that include inhibition of NF-κβ, Cox-1, and platelets, as masome in the FBR, we used s.c. implantation of silicone discs well as switching the catalytic activity of Cox-2, resulting in the for 4 wk. In the absence of Nlrp3, Asc, or caspase-1 the ability of generation of a family of molecules termed aspirin-triggered macrophages to migrate to the site of the biomaterial and to lipoxins (ATLs) (18). Pharmacological inhibition of Cox-1 and/or form FBGC was not impaired. In addition, macrophage density Cox-2 did not reproduce the inhibitory effect of aspirin on mi-

in the tissue adjacent to the biomaterial was not decreased. In crosphere-induced inflammasome activation, and it is of great MEDICAL SCIENCES fi contrast to the macrophage response, there was a very signi cant interest to examine the role of the family of ATLs in our system. reduction in the thickness of the FBR in the mice lacking Asc We further tested whether continuous aspirin administration and caspase-1, but not in mice lacking Nlrp3 or NLRC4. These could influence the FBR to silicone implants and found a sig- data lead to the question of why Nlrp3 is required for the acute nificant reduction in thickness of the FBR. Of potential clinical inflammatory response to biomaterials, but not the chronic fi- importance, the thickness of the FBR was reduced to less than brotic response. We propose that this is due to the mechanism of 100 μM, which could allow for enhanced diffusion of small MATS-induced activation, which by forming cholesterol rafts molecules (29). induces clustering of a variety of molecules with cytoplasmic Overall, our findings constitute a unique demonstration of the ITAMS. These molecules will have varying degree of efficiency participation of the inflammasome in cell–biomaterial inter- in activating Asc/caspase-1, with Nlrp3 being very efficient. In the actions and in the development of the FBR. These results are setting of a few hours, such as acute inflammation, the low- also significant and unexpected because they extend the role of efficiency molecules cannot make a significant contribution; inflammasome components to biomaterials, irrespective of size, however, over a period of several weeks, are allowed the low- and show that inflammasome activation can be triggered by efficiency molecules could provide a significant activation signal physical membrane contact. Due to the pervasive use of bio- to Asc/caspase-1. In addition to Nlrp3 and NLR4C, AIM2 and materials, identification of the molecular components required NLRP1 are known to form Asc-dependent inflammasomes, and for the FBR, and the demonstration of a clinically translatable a number of other NLRs have not been fully characterized yet method for reducing the FBR have broad applicability. (23). Taken together, these findings suggest that the loss of Materials and Methods inflammasome components hinders the progression of the FBR Animals. Nlrp3, Asc, IPAF, and caspase-1 KO mice were on the C57BL/6 possibly through the attenuation of profibrotic signals. Consis- −/− background and were maintained in specific pathogen-free facilities. Asc fl − − tent with this suggestion, the inability to form the in ammasome and Nlrp3 / mice were originally from Millennium Pharmaceuticals. Control has been shown to be associated with reduced fibrosis in injury C57BL/6 mice were purchased from NCI, and all mice were used at 8–12 wk models of liver fibrosis, myocardial ischemia/reperfusion, of age. All procedures were performed in accordance with the regulations

Malik et al. PNAS | December 13, 2011 | vol. 108 | no. 50 | 20099 Downloaded by guest on September 23, 2021 adopted by the National Institutes of Health and approved by the animal incubated overnight at 37 °C. AFM was performed as described previously care and use committee of Yale University. except for increasing hold time to 10 s in relative intermittent contact

mode in a temperature (37 °C)- and CO2 (5%)-controlled humidified cham- Biomaterial Induced Peritoneal Inflammation. Ultrapure PMMA microspheres ber (17). Force curves were obtained over 300 s and analyzed using JPK (Bang Labs; BB05N/5438) were obtained as dry powder and washed exten- image processing software. For inhibitor studies, MbCD (10 mM) or picea- sively in 70% ethanol and endotoxin-free water. A 0.5-mL suspension of tannol (50 μM) were present for 30 min before AFM reading. 5 × 105 PMMA bead in PBS supplemented with 0.02% (vol/vol) ethanol and – 0.05% (wt/vol) hydroxypropyl cellulose (6 10 cps; TCI) was injected i.p. using SEM Analysis. SEM was performed by plating 2 × 106 cells per well in a 24-well a 22-gauge needle and sterile techniques. At 24 h after i.p. injection, mice plate containing a sterile 12-mm glass disk as described previously. After 60 were euthanized and peritoneal fluid was lavaged, cell counts determined, min, medium was removed and replaced with warm fresh media. Micro- and flow cytometric analysis performed using antibodies to Gr1 and F4/80 spheres were added and shaken by hand for 2–3 min and then allowed to (BD Biosciences) using a FACS Calibur flow cytometer. Cellular protein lysates were also prepared for Western blot analysis of caspase-1 activation. The settle. After 60 min incubation, media were removed, samples were washed following antibody was used: caspase-1 p10 (M-20) from Santa Cruz (sc-514, twice with PBS, and 1% EM grade glutaraldehyde (Electron Microscopy rabbit polyclonal IgG). Sciences) was added. After overnight incubation (4 °C), samples were dehydrated first by ethanol gradient followed by hexamethyldisilazane In Vitro Macrophage Activation by PMMA Microspheres. C57BL/6 mice were (HMDS; Electron Microscopy Sciences) gradient. Samples were sputter coated injected i.p. with 4% thioglycolate (Fluka; B2551) under aseptic conditions. with gold-palladium (Techniques Hummer II Sputter Coater) and analyzed After 3 d, mice were euthanized and inflammatory monocytes/macrophages using SEM (environmental scanning electron microscope, XL30). were collected by peritoneal lavage. Cells were plated into 12-well plates at a density of 3 × 106 per well. After 3 h, the media (DMEM high glucose) was Foreign Body Reaction. All procedures were performed in accordance with the replaced with fresh media and cultured with 100 ng/mL of LPS (Sigma; regulations adopted by the National Institutes of Health and approved by the L9764) for 16 h. The media was changed afterward. For positive control animal care and use committee of Yale University. Six-millimeter disks were wells, 5 mM ATP (Sigma; A2383) was freshly added for exactly 20 min, and punched out from sterile 100% medical grade silicone sheet with 1-mm then the medium was changed. A ratio of one PMMA bead per 20 cells was thickness (Invotec) using a sterile Acu-Punch (Acuderm). Discs were implanted used to add in the respective wells. A total of 2 mM MbCD (Sigma; C4555) s.c. as described previously (30). After 4 wk, discs were removed and the and 50 μm piceatannol (Sigma; P0453) were added 40 min before addition of tissue stained by Masson’s trichrome stain according to standard protocols. PMMA microspheres in the respective wells. The plates were then incubated for 5 h in the incubator, and the supernatant collected to assay IL-1β. Cox Administration of Aspirin. For the acute experiments with PMMA micro- inhibitors sc-560 and sc-58125 were purchased from Cayman Chemicals. spheres, aspirin was provided in the drinking water for 3 d before bead injection. For the chronic FBR experiments, aspirin was provided in the ELISA. High-binding 96-well ELISA plates (BD Biosciences; 353279) were used as per manufacturer instructions. The capture antibody was antimouse IL-1β drinking water immediately after implantation of silicone discs. (R&D; MAB401), and the detection antibody was polyclonal biotinylated antimouse IL-1β antibody (R&D; BAF401). ACKNOWLEDGMENTS. We thank Eleni A. Skokos and Salma Kamal for technical assistance and Drs. Jordan Pober and Tarek Fahmy for a critical reading of the manuscript. This work was supported by National Institutes Atomic Force Microscopy. Arrow TL-1 cantilevers (Nanoworld) were func- of Health Grant R01DK076674-01A2 and a Veterans Administration Merit × 5 tionalized with PMMA tips under AFM control. A total of 5 10 lavaged Award (to W.Z.M.) and RO1 GM072194-01 (to T.R.K.), R21AI089963 (to Y.S.), macrophages were plated on 25-mm glass disks in six-well plates. After and 1K08DK092281-01 (to R.H.). R.A.F. is an Investigator for the Howard 60 min, media were removed and replaced with warm fresh media and Hughes Medical Institute.

1. Morais JM, Papadimitrakopoulos F, Burgess DJ (2010) Biomaterials/tissue interactions: 16. Hornung V, et al. (2008) Silica crystals and aluminum salts activate the NALP3 in- Possible solutions to overcome foreign body response. AAPS J 12:188–196. flammasome through phagosomal destabilization. Nat Immunol 9:847–856. 2. Anderson JM, Rodriguez A, Chang DT (2008) Foreign body reaction to biomaterials. 17. Ng G, et al. (2008) Receptor-independent, direct membrane binding leads to cell- – Semin Immunol 20:86 100. surface lipid sorting and Syk kinase activation in dendritic cells. Immunity 29:807–818. 3. Helming L, Gordon S (2009) Molecular mediators of macrophage fusion. Trends Cell 18. Chiang N, Serhan CN (2006) New mechanism for an old drug: aspirin triggers anti- Biol 19:514–522. inflammatory lipid mediators with gender implications. Compr Ther 32:150–157. 4. Luttikhuizen DT, Harmsen MC, Van Luyn MJ (2006) Cellular and molecular dynamics in 19. Flower R (2003) What are all the things that aspirin does? BMJ 327:572–573. the foreign body reaction. Tissue Eng 12:1955–1970. 20. Flach TL, et al. (2011) Alum interaction with dendritic cell membrane lipids is essential 5. Hu WJ, Eaton JW, Ugarova TP, Tang L (2001) Molecular basis of biomaterial-mediated for its adjuvanticity. Nat Med 17:479–487. foreign body reactions. Blood 98:1231–1238. 21. Helming L, et al. (2008) Essential role of DAP12 signaling in macrophage pro- 6. Tang L, Wu Y, Timmons RB (1998) Fibrinogen adsorption and host tissue responses to gramming into a fusion-competent state. Sci Signal 1:ra11. plasma functionalized surfaces. J Biomed Mater Res 42:156–163. 22. Mócsai A, Ruland J, Tybulewicz VL (2010) The SYK tyrosine kinase: A crucial player in 7. Szott LM, Stein MJ, Ratner BD, Horbett TA (2011) Complement activation on poly – (ethylene oxide)-like radiofrequency glow discharge-deposited surfaces. J Biomed diverse biological functions. Nat Rev Immunol 10:387 402. fl Mater Res A 96:150–161. 23. Bauernfeind F, et al. (2011) In ammasomes: Current understanding and open ques- 8. Kou PM, Babensee JE (2011) Macrophage and dendritic cell phenotypic diversity in tions. Cell Mol Life Sci 68:765–783. the context of biomaterials. J Biomed Mater Res A 96:239–260. 24. Kawaguchi M, et al. (2011) Inflammasome activation of cardiac fibroblasts is essential 9. Shokouhi B, et al. (2010) The role of multiple toll-like receptor signalling cascades on for myocardial ischemia/reperfusion injury. Circulation 123:594–604. interactions between biomedical polymers and dendritic cells. Biomaterials 31: 25. Vilaysane A, et al. (2010) The NLRP3 inflammasome promotes renal inflammation and 5759–5771. contributes to CKD. J Am Soc Nephrol 21:1732–1744. 10. Shi Y, Evans JE, Rock KL (2003) Molecular identification of a danger signal that alerts 26. Gasse P, et al. (2009) Uric acid is a danger signal activating NALP3 inflammasome in – the immune system to dying cells. Nature 425:516 521. lung injury inflammation and fibrosis. Am J Respir Crit Care Med 179:903–913. 11. Martinon F, Pétrilli V, Mayor A, Tardivel A, Tschopp J (2006) Gout-associated uric acid 27. Watanabe A, et al. (2009) Inflammasome-mediated regulation of hepatic stellate fl – crystals activate the NALP3 in ammasome. Nature 440:237 241. cells. Am J Physiol Gastrointest Liver Physiol 296:G1248–G1257. 12. Duewell P, et al. (2010) NLRP3 inflammasomes are required for atherogenesis and 28. Imaeda AB, et al. (2009) Acetaminophen-induced hepatotoxicity in mice is dependent activated by cholesterol crystals. Nature 464:1357–1361. on Tlr9 and the Nalp3 inflammasome. J Clin Invest 119:305–314. 13. Eisenbarth SC, Colegio OR, O’Connor W, Sutterwala FS, Flavell RA (2008) Crucial role 29. Novak MT, Yuan F, Reichert WM (2010) Modeling the relative impact of capsular for the Nalp3 inflammasome in the immunostimulatory properties of aluminium adjuvants. Nature 453:1122–1126. tissue effects on implanted glucose sensor time lag and signal attenuation. Anal Bi- – 14. Dostert C, et al. (2008) Innate immune activation through Nalp3 inflammasome oanal Chem 398:1695 1705. sensing of asbestos and silica. Science 320:674–677. 30. Kyriakides TR, et al. (2004) The CC chemokine ligand, CCL2/MCP1, participates in 15. Demento SL, et al. (2009) Inflammasome-activating nanoparticles as modular systems macrophage fusion and foreign body giant cell formation. Am J Pathol 165: for optimizing vaccine efficacy. Vaccine 27:3013–3021. 2157–2166.

20100 | www.pnas.org/cgi/doi/10.1073/pnas.1105152108 Malik et al. Downloaded by guest on September 23, 2021