Endocytosis of indium-tin-oxide nanoparticles by macrophages provokes pyroptosis requiring NLRP3-ASC-Caspase1 axis that can be prevented by mesenchymal stem cells

The biological effects of indium-tin-oxide (ITO) are of considerable importance because workers exposed to indium compounds have been diagnosed with interstitial lung disease or pulmonary alveolar proteinosis; however, the pathophysiology of these diseases is undefined. Here, mice intraperitoneally inoculated with ITO-nanoparticles (ITO-NPs) resulted in peritonitis dependent in NLRP3 inflammasome, with neutrophils recruitment and interleukin-1β (IL-1β) production. Withal peritoneal macrophages exposed ex vivo to ITO-NPs caused IL-1β secretion and cytolysis. Further, alveolar macrophages exposed to ITO-NPs in vitro showed ITO-NP endocytosis and production of tumor necrosis factor-α (TNF-α) and IL-1β, ensued cell death by cytolysis. This cell death was RIPK1-independent but caspase1-dependent, and thus identified as pyroptosis. Endocytosis of ITO-NPs by activated THP-1 cells induced pyroptosis with IL-1β/TNF-α production and cytolysis, but not in activated THP-1 cells with knockdown of NLRP3, ASC, or caspase1. However, exposing activated THP-1 cells with NLRP3 or ASC knockdown to ITO-NPs resulted in cell death but without cytolysis, with deficiency in IL-1β/TNF-α, and revealing features of apoptosis. While, mesenchymal stem cells (MSCs) co-cultured with macrophages impaired both inflammation and cell death induced by ITO-NPs. Together, our findings provide crucial insights to the pathophysiology of respiratory diseases caused by ITO particles, and identify MSCs as a potent therapeutic.


Results
Intraperitoneal inoculation in mice of ITO-NPs induces peritonitis dependent on NLRP3 inflammasome activation. MSU crystals cause an innate immune response via inflammasome, because mice deficient in ASC (a key component of inflammasome) show impaired neutrophil recruitment after intraperitoneal (i.p.) inoculation of MSU crystals as compared with WT mice 12 . Using a comparable approach, we found that mice with i.p. ITO-NP inoculation showed induced peritonitis with inflammatory cell recruitment ( Fig. 1a-g and Supplemental Fig. S1a,b). Indeed, among CD45 + CD11b + peritoneal cells of mice inoculated i.p. with ITO-NPs, we found a marked increase of neutrophils phenotypically characterized as CD11b + Gr-1 + Ly-6G + and inflamed monocytes characterized as CD11b + Gr-1 + Ly-6G − (Fig. 1b,c,f). Along with the efflux of neutrophils and inflamed monocytes, the level of IL-1β was significantly increased within peritoneal fluid of mice receiving ITO-NPs (Fig. 1g). By contrast, in control mice, CD45 + CD11b + peritoneal cells were negative for the Gr-1 marker, indicating an absence of neutrophils or inflamed monocytes (Fig. 1b,f). Among these peritoneal cells, a subset expressed a high density of CD11b antigens, and were further phenotypically characterized as CD11b + F4/80 + (Fig. 1e and Supplemental Fig. S1a,b), representing resting resident peritoneal macrophages 27 .
Peritonitis induced by ITO-NPs was diminished by impeding NLRP3 inflammasome activation by a sulfonylurea pharmacological inhibitor, glybenclamide (Gly) (Fig. 1d-g). Gly acts upstream of NLRP3 inflammasome by inhibiting K + efflux mediated via the Panx1 ion channel 28,29 . Here, Gly greatly reduced the number of neutrophils without affecting that of inflamed monocytes in mice receiving ITO-NPs (Fig. 1c-f). Thus, the peritoneal neutrophil/monocyte cell ratio was about 4:1 in mice receiving ITO-NPs but < 2:1 in mice receiving both Gly and ITO-NPs (Supplemental Fig. S1c). Furthermore, Gly restrained IL-1β production in ITO-NPs receiving mice (Fig. 1g). Here, our data indicate that ITO-NPs activate NLRP3 inflammasome in vivo to exacerbate inflammation.
Scientific RepoRts | 6:26162 | DOI: 10.1038/srep26162 macrophages exposed to ITO-NPs resulted in IL-1β and TNF-α production and release of LDH (Fig. 1h), similar results were observed when peritoneal macrophages were exposed to SiO 2 -NPs and MSU (Supplemental Fig. S1d). As well, to exclude any endotoxin-mediated effect resulting from contamination of ITO-NPs, splenocytes and bone marrow cells (BMCs) isolated from naïve mice were exposed to ITO-NPs or lipopolysaccharide (LPS) and examined for their viability and ability to secrete TNF-α . Splenocytes and BMCs responded vigorously to LPS in terms of TNF-α production but were unable to secrete TNF-α on exposure to ITO-NPs at 50, 100 and 500 μ g/mL, whilst cell viability was not disturbed for any treatments tested (Supplemental Fig. S2a,b). Thus, ITO-NPs were excluded from contamination with endotoxins and ITO-NPs could not affect splenocytes, or immature BMCs. Yet, in agreement with the findings obtained with resident peritoneal macrophages, ITO-NPs affected bone marrow-derived macrophages (BMDMs) by stimulating the release of IL-1β and TNF-α , and by decreasing cell viability through cytolysis (Supplemental Fig. S2c,d). Here, our findings indicate that ITO-NPs affect specifically macrophages. These unfavorable effects of ITO-NPs on macrophages might be related to their capability for particles uptake.
Endocytosis of ITO-NPs by alveolar macrophages provokes TNF-α and IL-1β production and caspase1-dependent cell death. Next, we compared the molecular mechanisms leading to inflammation and cell death in alveolar macrophages (AMs) exposed to ITO-NPs, SiO 2 -NPs and MSU crystals (Fig. 2a). Dysfunction of AMs is often associated with the pathogenesis of respiratory diseases 30 . Particularly, AMs prevent xenogenous compounds from accumulating into alveolar sacs, and AM endocytic potentials provide a vital defense 30 . We used transmission electronic microscopy (TEM) to investigate the uptake of ITO-NPs by MH-S cells, an established murine cell line of AMs. At 6-h post-exposure to 50 μ g/mL ITO-NPs, MH-S cells exhibited endosomes containing ITO-NPs appearing as aggregated electronic dense spots (Fig. 2b,c). At 24-h post-exposure, endosomes were enlarged and contained an increased amount of ITO-NPs (Fig. 2c). With higher concentrations of ITO-NPs, MH-S cells showed endosomes comprising an augmented amount of ITO-NPs (Supplemental Fig. S3). Thus, the size of endosomes containing ITO-NPs was dose and time-dependent. As well, MH-S cells stimulated with LPS and exposed to ITO-NPs exhibited endosomes containing ITO-NPs (Supplemental Fig. S4a). Additionally, MH-S cells were capable of uptake of SiO 2 -NPs and MSU crystals (Supplemental Fig. S4b), compounds with a strong ability to trigger activation of NLRP3 inflammasome 14 . MH-S cells exposed to ITO-NPs dose-and time-dependently secreted TNF-α (Fig. 2d). In parallel, intracellular TNF-α level was increased (Fig. 2e), which demonstrates de novo protein production by macrophages in response to ITO-NPs. As well, intracellular and secreted TNF-α level was increased in MH-S cells exposed to SiO 2 -NPs or MSU crystals (Supplemental Fig. S4d-f).
Further, we examined dexamethasone (DEX) as a potent immunosuppressive molecule that could prevent ITO-NPs-induced inflammatory signaling in macrophages. DEX diminished the intracellular and secreted TNF-α level in MH-S cells exposed to ITO-NPs (Fig. 2f), but secretion of the immunosuppressive cytokine IL-10 was increased (Fig. 2g). As well, DEX lessened TNF-α secretion in MH-S cells exposed to SiO 2 -NPs (Supplemental Fig. S4e). Next, we examined ITO-NP uptake blockade by using an inhibitor of actin polymerization, cytochalasin D (Cyto D). Cyto D prevented endocytosis of ITO-NPs in MH-S cells ( Fig. 2h and Supplemental Fig. S4c), and significantly decreased TNF-α secretion (Fig. 2h). However, and by contrast to DEX, Cyto D did not counter the intracellular overexpression of TNF-α in MH-S cells exposed to ITO-NPs (Fig. 2h). As compared with Cyto D, DEX may regulate inflammatory proteins at the transcriptional level in preventing inflammatory signaling induced by ITO-NPs. In accordance with this data, MH-S cells exposed to ITO-NPs or SiO 2 -NPs showed upregulated IL-1β expression ( Fig. 2i and Supplemental Fig. S4g), and DEX but not Cyto D markedly inhibited IL-1β overexpression (Fig. 2i). Here, blockade of endocytosis indicated that ITO-NPs could interact with surface receptors expressed on AMs, which was sufficient to transduce intracellular pro-inflammatory signaling. Moreover, DEX might counteract these inflammatory signaling at the transcriptional level.

ITO-NPs induce caspase1-dependent pyroptosis in alveolar macrophages involving scavenger receptor class-A (SR-A) CD204.
We next investigated ITO-NP-mediated cell death in AMs and whether caspase1, endocytosis and SRs are involved in this process. In addition to increasing pro-inflammatory cytokine production, ITO-NPs inhibited cell proliferation of MH-S cells (Supplemental Fig. S5a) and, LPS exacerbated this inhibition (Supplemental Fig. S5b). DEX did not prevent cell proliferation inhibition in MH-S exposed to ITO-NPs (Supplemental Fig. S5c). This cell proliferation inhibition was accompanied by a decrease in cell viability (Supplemental Fig. S5d,g), and consistently, LPS intensified the decreased cell viability (Supplemental Fig. S5e,h). Of note, DEX could not reverse the loss of cell viability induced by ITO-NPs in MH-S cells (Supplemental Fig. S5f,i). Furthermore, we examined the ability of 51 Cr-radiolabeled MH-S cells exposed to ITO-NPs to release 51 Cr into the extracellular milieu, which represents a direct measurement of AM plasma membrane rupture. 51 Cr-radiolabeled MH-S exposed to ITO-NPs dose-dependently released 51 Cr into the extracellular milieu (Fig. 3a), and LPS augmented the 51 Cr release (Fig. 3b). Of note, DEX did not prevent MH-S cells from releasing 51 Cr into the extracellular milieu with ITO-NP exposure (Fig. 3c). On the other hand, pharmacological blockade of caspase1 activity by Ac-YVAD significantly reduced 51 Cr release resulting from exposure of 51 Cr-radiolabeled MH-S cells to ITO-NPs (Fig. 3d). Consistently, exposed MH-S cells released LDH; Ac-YVAD but not DEX prevented the LDH release (Fig. 3e). Ac-YVAD diminished TNF-α secretion of exposed MH-S cells (Fig. 3f). Together, these data reveal that AMs exposed to ITO-NPs undergo a regulated cell death (RCD), with plasma membrane rupture, dependent on caspase1 that is evocative of pyroptosis 31 .
Next, we sought to precise the molecular pathway involved in AM cell death induced by ITO-NPs. Cyto D prevented the LDH release in MH-S exposed to ITO-NPs (Fig. 3g), which demonstrates a critical role for ITO-NP endocytosis in triggering pyroptosis. Comparable protective effects of Cyto D were observed in MH-S cells exposed to SiO 2 -NPs (Supplemental Fig. S6a,b) or MSU crystals (Supplemental Fig. S6c,d). Furthermore, we Scientific RepoRts | 6:26162 | DOI: 10.1038/srep26162 examined whether necroptosis, a RCD that depends on receptor-interacting Ser/Thr-protein kinase 1 (RIPK1) 32 , or accidental cell membranes rupture due to oxygen species overproduction 33 , are involved in ITO-NPs-induced cell death. Necrostatin-1 (NEC-1), a potent RIPK1 inhibitor, did not neutralize LDH release in MH-S exposed to ITO-NPs (Fig. 3h), nor did neutralizing reactive oxygen species (ROS) with N-acetyl-(L)-cysteine (NAC), a non-exposed MH-S cells shows preserved cytoplasmic organelles, regular microvilli protruding, and well-delineated nucleus with heterochromatin. (c) MH-S cells exposed for 6 h to 50 μ g/mL ITO-NPs exhibited early endosomes containing ITO-NPs appearing as dark spots. At 24 h post-exposure to ITO-NPs at 50 μ g/mL, endosomes in MH-S contained an increased amount of ITO-NPs. (b,c) Original magnification x3000, x4000 and x8000, 1 representative TEM image is shown from at least 3. (d) MH-S cell production of TNF-α upon exposure to increasing doses, i.e. 50, 100 and 500 μ g/mL, of ITO-NPs. MH-S cells exposed to ITO-NPs at 500 μ g/mL and examined at 3, 6, 12 and 16 h for secretion of TNF-α . (e) Intracellular TNF-α level found in MH-S cells exposed ITO-NPs compared to unstimulated MH-S. (f) Secreted and intracellular TNF-α levels in MH-S cells treated with dexamethasone (DEX) and exposed to ITO-NPs. (g) IL-10 production in MH-S cells exposed to ITO-NPs with DEX treatment and compared to controls. (h) MH-S cells treated with or without Cyto D and exposed to ITO-NPs for 12 h were analyzed in 3 representative areas (RA) on x500 TEM images. The number of endocytic cells containing ITO-NPs was compared to the total number of nucleated cells in RA; data are given as percentage of endocytic cells. Cyto D effect was evaluated for TNF-α secretion and intracellular TNF-α expression in MH-S upon exposure to ITO-NPs. (i) Intracellular IL-1β level was evaluated in MH-S cells with ITO-NP exposure; with or without DEX and Cyto D treatment compared to controls. The concentration of ITO-NPs was 500 μ g/mL and time of   51 Cr increased in dose-dependently on exposure to 50, 100, and 500 μ g/mL of ITO-NPs. (b) MH-S cells treated with LPS and exposed to 50 μ g/mL ITO-NPs were compared to control for release of 51 Cr. (c) DEX treated MH-S were evaluated for 51 Cr release on exposure to ITO-NPs. (d) Blockade of caspase1 with Ac-YVAD in MH-S cells exposed to increasing doses of ITO-NPs were evaluated for preventing 51 Cr release. (e) LDH release was examined in MH-S cells exposed to 500 μ g/mL ITO-NPs for 24 h with Ac-YVAD or DEX treatment. (f) TNF-α secretion from MH-S cells exposed to 500 μ g/mL ITO-NPs with or without Ac-YVAD, 16 h postexposure, were measured and compared to controls. (g) The effect of ITO-NPs endocytosis blockade with Cyto D were evaluated in MH-S cells for LDH release at 16 and 24 h post-exposure to 500 μ g/mL ITO-NPs. (h) ROS scavenger N-acetyl-(L)-cysteine (NAC) and blockade of RIPK1 with necrostatin-1 (NEC-1) were tested in MH-S cells at 24 h post-exposure to 500 μ g/mL ITO-NPs for LDH release. (i) MH-S cells exposed to ITO-NPs for 8 h compared to unstimulated MH-S were analyzed after Annexin V and propidium iodide (PI) staining on flow cytometer. Dot plot of Annexin V and PI staining of MH-S cells exposed to ITO-NPs (red) compared to control (black). FSC/SSC shows MH-S with ITO-NPs uptake with increasing granularity compared to control and cells noted E(+ ) are for cells containing endosome. Percentage of E(+ ) cells are given. (j) Percentage of total cells being stained with Annexin V and PI are given. An increase in the percentage of cells double positive Annexin V + PI + among MH-S exposed to ITO-NPs, suggest a cell death with plasma membrane rupture. (a-j) n = 3-6 ± SEM, *p < 0.05.
ROS scavenger, diminish cell death by plasma membrane rupture in exposed MH-S cells (Fig. 3h). These data highlight a central role for caspase1 in inducing a RCD in AMs exposed to ITO-NPs, and excludes a cell death via RIPK1-mediated necroptosis or a cell death resulting from cell membranes damage due to ROS overproduction. Further, MH-S cells exposed to ITO-NPs weakly exhibited phosphatidylserine (PS) phospholipids on their outer leaflet membrane as revealed by a modest increase in the number of AnnexinV + PI − cells (Fig. 3i,j). Whereas, AnnexinV + PI + cells were further represented in MH-S exposed to ITO-NPs (Fig. 3i,j), indicating breakage in plasma membrane integrity and featuring pyroptotic cells. These data suggest a RCD occurring in MH-S cells on ITO-NP exposure that is not necroptosis nor apoptosis, but a RCD caspase1-dependent, with plasma membrane disruption, typical of pyroptosis 16 .
To identify the receptors expressed by AMs that could enable ITO-NPs to induce pyroptosis, we first examined whether MH-S cells could express the collagenous scavenger receptor class-A (SR-A) MARCO, CD204 and the non-collagenous SR-B CD36 and whether ITO-NPs could interact with SRs. MH-S cells expressed SR-A CD204 and SR-B CD36, but SR-A MARCO were almost not detectable on cell surface ( Fig. 4a and Supplemental Fig. S7a). Furthermore, MH-S cells exposed to ITO-NPs and SiO 2 -NPs showed a marked decrease in SR-A CD204 surface expression and to a lesser extent SR-B CD36 expression (Fig. 4a-d and Supplemental Fig. S7b-e). Hence, we examined whether neutralizing SR-A MARCO, CD204 or SR-B CD36 with specific monoclonal antibodies (mAbs) could reverse ITO-NP-and SiO 2 -NP-induced pyroptosis. Remarkably, blockade of SR-A CD204 with 2F8 mAb significantly prevented the death of MH-S cells exposed to ITO-NPs or SiO 2 -NPs as measured by LDH release (Fig. 4e and Supplemental Fig. S7f). However, blockade of SR-B CD36 or SR-A MARCO with JC63.1 or ED31 mAbs, respectively, did not prevent plasma membrane disruption of MH-S cells exposed to ITO-NPs and SiO 2 -NPs ( Fig. 4f and Supplemental Fig. S7g-i). Thus, our data indicate that ITO-NPs may interact with AMs through SR-A CD204 leading ultimately to caspase1-dependent pyroptosis.
Then, we considered caspase1-dependent pyroptosis in activated monocytes, to elucidate the molecular pathway of pro-inflammatory cytokines production and LDH release following ITO-NPs endocytosis. As described for AMs, DEX inhibited production of both TNF-α and IL-1β in PMA-activated THP-1 cells exposed to ITO-NPs but did not prevent cell death (Fig. 5f). Importantly, Cyto D obstructed endocytosis of ITO-NPs in PMA-activated THP-1 cells, restrained IL-1β secretion and prevented LDH release (Fig. 5g,h). By contrast, NAC and NEC-1 did not impede cell death in PMA-activated THP-1 cells exposed to ITO-NPs (Fig. 6a). Consistent with our findings regarding AMs, pharmacological blockade of caspase1 activity using Ac-YVAD, significantly restricted cell death and diminished IL-1β and TNF-α secretion in PMA-activated THP-1 cells exposed to ITO-NPs (Fig. 6a,b). Similar beneficial effect of Ac-YVAD was found in PMA-activated THP-1 cells with SiO 2 -NP or MSU crystal exposure (Supplemental Fig. S9a-b). To reveal whether caspase1 is specifically indispensable to pyroptosis in activated-monocytes ensuing ITO-NPs endocytosis, and to exclude a role of caspase4 and caspase5 in participating in this event 17 , we examined the response of THP-1 genetically knockdown for caspase1 (i.e. deficient in caspase1, abridged def-Casp1) exposed to ITO-NPs. We found that while PMA-activated wild-type (WT) THP-1 exposed to increasing doses of ITO-NPs produced IL-1β in a dose-dependent manner, PMA-activated THP-1 knockdown for caspase1 were unable to secrete IL-1β (Fig. 6c). As well, PMA-activated THP-1 caspase1 knockdown were prevented from plasma membrane rupture resulting of ITO-NPs exposure, as assessed by an absence of LDH release compared to WT control (Fig. 6d).
Thereby the pyroptosis is achieved through plasma-membrane rupture, and thus of the release of intracellular contents that can include molecules such as alarmin IL-1α 34,35 . IL-1α is considered to have a potent role in inducing tissue damage and fibrosis 34,36 . Therefore, we examined the ability of activated monocytes exposed to ITO-NPs to produce and release IL-1α . IL-1α production and release was significantly upregulated in PMA-activated THP-1 cells exposed to ITO-NPs, and similar results were found with SiO 2 -NP or MSU crystal exposure ( Fig. 6E and Supplemental Fig. S9c,d). Impeding caspase1 activity or endocytosis with Ac-YVAD or Cyto D prevented IL-1α release of PMA-activated THP-1 cells exposed to ITO-NPs (Fig. 6f,g). These findings indicate that both the release of IL-1α and the cell death through cytolysis, induced by ITO-NPs in macrophages, is dependent on activity of caspase1. As well, endocytosis of ITO-NPs by PMA-activated THP-1 cells was required for IL-1α release, which is consistent with our data showing that blockade of endocytosis, prevented the rupture of plasma membrane in macrophages exposed to ITO-NPs, thus impeding intracellular content release of IL-1α .
PMA-activated monocytes with NLRP3 or ASC knockdown show ITO-NP uptake with apoptotic cell death rather than pyroptosis. Because NOD-like receptor NLRP3 and its adaptor ASC are critical components of the NLRP3 inflammasome permitting caspase1 activation 4,14,37 , we next investigated whether ITO-NPs could specifically require NLRP3-ASC-Caspase1 axis for inducing pyroptosis in macrophages. We compared the effects of ITO-NPs in THP-1 cells with stable knockdown of NLRP3 (def-NLRP3) or ASC (def-ASC) and WT THP-1 cells. On exposure to ITO-NPs, PMA-activated THP-1 cells with NLRP3 or ASC deficiency were unable to release LDH, thus showing for unaltered plasma membranes (Fig. 7a). Furthermore, PMA-activated  THP-1 WT cells secreted IL-1β and TNF-α with increasing doses of ITO-NPs (Fig. 7b,c); whereas PMA-activated THP-1 cells with NLRP3 or ASC deficiency were unable to secrete IL-1β or TNF-α (Fig. 7b,c).
Nonetheless, ITO-NPs increased cell death among both THP-1 WT and NLRP3 or ASC knockdown THP-1 cells with PMA activation, as assessed by optical microscope cell count on trypan blue dye exclusion (Fig. 7d). As cell viability decline was not associated with LDH release in NLRP3 or ASC knockdown THP-1 cells with PMA activation and exposed to ITO-NPs (Fig. 7a), we examined whether these cells were subjected to apoptosis. We found that WT PMA-activated THP-1 cells, upon exposure to ITO-NPs, exhibited a modest increase in AnnexinV + PI − apoptotic cells as compared to Annexin V + PI + pyroptotic cells (Fig. 7e). On the contrary, NLRP3 or ASC knockdown THP-1 cells with PMA activation, ITO-NP exposure resulted in a marked increase of Annexin V + PI − apoptotic cells, while Annexin V + PI + pyroptotic cells were not augmented compared to WT control (Fig. 7e). Consistently, TEM of PMA-activated THP-1 cells with NLRP3 or ASC deficiency exposed to ITO-NPs were capable of ITO-NPs uptake, and often exhibited apoptotic morphological features among macrophages with endosomes containing ITO-NPs, such as membrane blebs and apoptotic bodies (Fig. 7f). Conversely, among PMA-activated THP-1 WT cells exposed to ITO-NPs, unfit endocytic cells often appeared with disrupted plasma membranes (Fig. 7f). Our data suggest that circumvention of NLRP3 or ASC inflammasome components in macrophages exposed to ITO-NPs could prevent pyroptosis but facilitate alternatively apoptosis. Moreover, ITO-NP endocytosis by macrophages did not directly or, indirectly through ROS overproduction, causes damage to cell membranes, but rather demonstrated that NLRP3 inflammasome activation and caspase1 function are specifically required for pyroptosis. Blockade of inflammation using DEX or even specific blockade of NLRP3 inflammasome components might not be optimal in preventing cell death in macrophages exposed to ITO-NPs.

Mesenchymal stem cells prevent pyroptosis in alveolar macrophages and activated monocytes
with ITO-NP, SiO 2 -NP or MSU crystal exposure. Next, we examined whether mesenchymal stem cells (MSCs) could prevent both the inflammatory response and cell death in macrophages undergoing pyroptosis (Fig. 8a). MSCs possess immunosuppressive and regenerative properties capable of controlling inflammatory disorders such as graft versus host disease, Crohn's disease and chronic obstructive pulmonary disease 25,26,38 . Here, primary MSCs were derived from adult human bone marrow mononuclear cells 39 . MSCs were expanded in vitro and established as CD14 -CD34 -CD45non-hematopoietic cells and phenotypically characterized as CD73 + CD90 + CD105 + cells (Fig. 8b). On TEM, MSCs were large with a clear cytoplasm containing an eccentric nucleus and were morphologically discernable from macrophages (Fig. 8c,e). MSCs co-cultured with MH-S or PMA-activated THP-1 cells exposed to ITO-NPs were analyzed by TEM. MH-S and PMA-activated THP-1 cells LDH release as measured in PMA-activated THP-1 cells exposed to ITO-NPs with caspase1 activity blockade using Ac-YVAD, compared with neutralizing ROS with NAC or RIPK1 with NEC-1 (b) IL-1β and TNF-α secretion in PMA-activated THP-1 cells exposed to ITO-NPs with caspase1 activity blockade using Ac-YVAD. (c) Effect of increasing doses of ITO-NPs in PMA-activated THP-1 with a stable knockdown in caspase1 (def-Casp1) compared to wild-type (WT) PMA-activated THP-1 in terms of IL-1β secretion and (d) LDH release. (e) Intracellular expression and release of IL-1α alarmin in PMA-activated THP-1 exposed to ITO-NPs compared to controls. (f) Caspase1 activity blockade with Ac-YVAD or (g) endocytosis of ITO-NPs blockade with Cyto D, reduced IL-1α release by PMA-activated THP-1 cells exposed to ITO-NPs. Concentration of ITO-NPs was 500 μ g/mL, and time of exposure was 16 h, unless stated otherwise. (a-g), n = 3-4 ± SEM, *p < 0.05.
Scientific RepoRts | 6:26162 | DOI: 10.1038/srep26162 Figure 7. Endocytosis of ITO-NPs by PMA-activated THP-1 cells with NLRP3 or ASC knockdown does not elicit pyroptosis but rather apoptosis. (a) LDH release capabilities as measured in PMA-activated THP-1 either wild-type (WT) or NLRP3 knockdown (def-NLRP3), or ASC knockdown (def-ASC) on exposure to increasing doses of ITO-NPs. (b) PMA-activated THP-1 WT cells exposed to increasing doses of ITO-NPs but not cells with NLRP3 knockdown or ASC knockdown released IL-1β dose-dependently. (c) Similarly, PMA-activated THP-1 WT cells exposed to increasing doses of ITO-NPs but not cells with NLRP3 knockdown or ASC knockdown released TNF-α dose-dependently. (d) PMA-activated THP-1 WT cells or with NLRP3 knockdown or ASC knockdown exposed to ITO-NPs were assessed for cell viability by trypan blue dye exclusion cell count on light microscope. Results are given as number of cells alive and dead cells (×10 6 ) with or without exposure to ITO-NPs. (e) PMA-activated THP-1 WT cells or cells with NLRP3 knockdown or ASC knockdown exposed to ITO-NPs were stained with Annexin V and PI to determine the relative number of single positive cells Annexin V + PI − (apoptotic cells) compared to the number of cells double positive Annexin V + PI + (pyroptotic cells). Results are given as a percentage of total cells with or without exposure to ITO-NPs. (f) TEM analysis of PMA-activated THP-1 WT cells exposed to ITO-NPs exhibited almost exclusively endocytic cells containing ITO-NPs with plasma membrane being ruptured; black arrow shows ITO-NPs being released after plasma membrane has dislocated (red arrow). By contrast, PMA-activated THP-1 cells with NLRP3 knockdown or ASC knockdown exhibited cells with capabilities of ITO-NPs uptake (black arrow), showing morphological features of apoptosis, such as membrane blebs and apoptotic bodies (blue arrow). TEM original magnification x3000 and x8000; 1 representative image is shown in at least 3. Concentration of ITO-NPs was 500 μ g/mL, and time of exposure was 16 h, unless stated otherwise. (a-f) n = 3-8 ± SEM, *p < 0.05. TEM abbreviations, Cyto = cytoplasm, N = nucleus and E = endosome.
but not MSCs were capable of ITO-NP endocytosis (Fig. 8d,e). Furthermore, TEM of co-cultured MSCs and macrophages showed tight cell-to-cell interaction between MSCs and macrophages exhibiting endosomes containing ITO-NPs, without displaying plasma membrane rupture or apoptosis morphological features (Fig. 8e). Furthermore, MH-S cells exposed to ITO-NPs released a reduced amount of LDH, which was directly proportional to the number of MSCs present (Fig. 8f). As well, TNF-α secretion decreased significantly when MSCs were co-cultured with ITO-NP-exposed MH-S cells (Fig. 8g). Similar results were found for MH-S exposed to (e) TEM of MSCs co-cultured with MH-S or PMA-activated THP-1 cells exposed to ITO-NPs. TEM reveals tight cell-to-cell contacts (orange arrow) between MSCs and macrophages containing ITO-NP endosomes (black arrow); original magnification ×3000 and ×8000, 1 representative TEM image in 3 is shown. (f) MSCs co-cultured with MH-S cells exposed to ITO-NPs exhibited LDH release that was inversely proportional to the number of MSCs present. (g) MSCs effect on TNF-α release by MH-S cells exposed to ITO-NPs. (h) MSCs co-cultured with PMA-activated THP-1 cells exposed to ITO-NPs were tested for LDH release, and the release of LDH was inversely proportional to the number of MSCs present. (i) TNF-α release by PMA-activated THP-1 cells exposed to ITO-NPs in coculture with MSCs. (j) MSCs co-cultured with PMA-activated THP-1 cells exposed to ITO-NPs were examined for IL-10 release. MSCs alone are represented as an empty green circle on the graph. Concentration of ITO-NPs was 500 μ g/mL, and exposure time 16 h, unless stated otherwise. (b-j) n = 3-6 ± SEM, *p < 0.05. TEM abbreviations, Cyto = cytoplasm, N = nucleus and E = endosome. SiO 2 -NPs and MSU crystals on co-culture with MSCs (Supplemental Fig. S10a,b). Consistently, MSCs alleviated LDH release and TNF-α secretion resulting from exposure to ITO-NPs in PMA-activated THP-1 cells (Fig. 8h,i). Similar protective effect of MSCs was found in PMA-activated THP-1 cells upon exposure to SiO 2 -NPs or MSU crystals (Supplemental Fig. S10c,d). Concomitantly, we found a significant increase in secretion of the immunosuppressive cytokine IL-10 in PMA-activated THP-1 cells co-cultured with MSCs and exposed to ITO-NPs, SiO 2 -NPs or MSU compared to controls (Fig. 8j and Supplemental Fig. S10e). Therefore, MSCs possess potent ability to protect macrophages against both arms of pyroptosis -inflammation and cell death -induced by ITO-NPs, SiO 2 -NPs or MSU crystals and might favor tissue regeneration by increasing IL-10 secretion to further promote tissue healing 40 .

Discussion
Cummings K. and colleagues revealed that occupational inhalation of InO and ITO particles cause respiratory diseases 1 . These respiratory diseases are characterized by a rapid evolution: the median age at diagnosis is 35 years old, with cases of premature death 1,41,42 . To provide further insights into the pathophysiology for respiratory diseases; we examined the inflammatory ability of ITO-NPs in vivo in mice and in vitro in AMs and monocyte-derived macrophages.
We first show that ITO-NP-caused peritonitis in mice depended on NLRP3 inflammasome. Resident macrophages first in contact with ITO-NPs may undergo cell death via cytolysis, which can have a considerable impact by inducing prolonged inflammation with tissue damage and fibrosis. Consistently, a mouse model of bleomycin-induced pulmonary fibrosis showed that ATP released from injured cells amplified NLRP3 inflammasome 43 . Previous studies have shown that extracellular ATP availability and subsequent activation of the P2× 7 R/Panx1 pathway predominantly enhanced neutrophil accumulation in inflammatory diseases 10,12,44,45 . This agrees with our results showing that blockade of NLRP3 inflammasome with glybenclamide in the ITO-NPmediated peritonitis mouse model principally reduced the influx of neutrophils and IL-1β production.
Then, we examined the ITO-NP effects in AMs and monocyte-derived macrophages. Particularly, AMs exposed to ITO-NPs, SiO 2 -NPs, or MSU crystals provoked a regulated cell death (RCD) distinguished by cytolysis that depended on caspase1 activity; this RCD was identified as pyroptosis 16 . Necroptosis, which is a RIPK1-dependent and caspase1-independent RCD that resembles to pyroptosis, was not involved in mediating cell death of AMs exposed to ITO-NPs. Indeed, blockade of RIPK1 with NEC-1 failed to rescue AMs from cytolysis. As well, blockade of ROS with NAC in AMs exposed to ITO-NPs did not impede cytolysis, which demonstrates the major role of caspase1 in inducing pyroptosis. Furthermore, ITO-NPs and SiO 2 -NPs induced pyroptosis in AMs via scavenger receptor class-A (SR-A) CD204, and endocytosis of these particles was required for execution of pyroptosis. The blockade of CD204, but not of SR-A MARCO or SR-B CD36, reduced the cytolysis induced by ITO-NPs or SiO 2 -NPs in AMs. Thus, SR-A CD204 could have a role in AMs as an extracellular pattern-recognition receptor [46][47][48] , acting upstream of NLRP3 inflammasome, to detect and facilitate endocytosis of ITO-NPs like SiO 2 -NPs.
Further, endocytosis of ITO-NPs by macrophages results in NLRP3 inflammasome activation, resulting in the production of IL-1β , IL-1α and TNF-α , and cytolysis, both events -inflammation and cell death -being caspase1-dependent. Moreover, ITO-NP-induced pyroptosis was abrogated in PMA-activated THP-1 cells with knockdown of NLRP3, ASC or caspase1. Thus, NLRP3 inflammasome was indispensable in mediating caspase1-dependent pyroptosis in macrophages exposed to ITO-NPs. Concurrently, we have excluded other inflammasome pathways such as NLRP1, NLRC4, NLRP6 or AIM2 14,37,49 participating in caspase1-dependent pyroptosis induced by ITO-NPs. In addition, PMA-activated THP-1 cells with caspase1 knockdown and ITO-NP exposure were unable to undergo pyroptosis. This finding excludes the implication of caspase4 or caspase5, caspases known for inducing pyroptosis independently to NLRP3 inflammasome and caspase1 50,51 , in the cell death of PMA-activated THP-1 cells with ITO-NP exposure. Altogether, it appears that pyroptosis induced in macrophages upon exposure to ITO-NPs is strictly dependent on endocytosis of the particles, NLRP3 inflammasome activation and caspase1 function. However, both NLRP3 and ASC knockdown THP-1 cells, with PMA activation and with ITO-NP exposure underwent a cell death with features of apoptosis. While PMA-activated THP-1 cells with NLRP3 or ASC deficiency and ITO-NP exposure did not secrete IL-1β /TNF-α or undergo cytolysis, cell viability has declined with an increase in AnnexinV + PI − apoptotic cells. Moreover, TEM of PMA-activated THP-1 cells knockdown with NLRP3 or ASC and ITO-NP exposure exhibited morphological features of apoptosis, with endocytic cells showing membrane blebs and apoptotic bodies. This finding agrees with the hypothesis that NLRP3 inflammasome might play a role as a molecular switch in guiding inflammatory (i.e., pyroptosis) and non-inflammatory (i.e., apoptosis) RCD 52,53 .
Thus, controlling prolonged inflammation induced by xenogenous compounds implies targeting both inflammation and cell death 54 . Although inflammation can be impeded by directly targeting inflammatory pathways, cell death might still occur among immune cells, thus disorganizing the innate immunity 16,54 . Indeed, DEX moderates lung fibrosis in rodents 55,56 , and glucocorticoids are used as treatment for certain respiratory diseases 57 ; however, the actual benefit of DEX is still poorly defined 58 . In our study, we examined DEX properties in the context of pyroptosis induced by ITO-NPs, SiO 2 -NPs and MSU crystals in macrophages. We ascertained that DEX inhibits the production of IL-1β and TNF-α in macrophages exposed to ITO-NPs, SiO 2 -NPs or MSU crystals. Yet, DEX did not neutralize caspase1-dependent cell death induced by ITO-NPs in macrophages. Consistently, in a model of Mycobacterium tuberculosis (MTb) infection of microglial cells, which vigorously activate caspase1 and result in IL-1β and TNF-α upregulation, pre-treating microglial cells with DEX and infection with MTb downregulated IL-1β and TNF-α transcription 59 . In contrast, DEX can increase NLRP3 transcriptional expression in macrophages 60 , rendering macrophages even more responsive to caspase1-dependent cell death. Agreeing with this, our findings indicate that blockade of inflammatory pathways by DEX or by circumventing NLRP3 Scientific RepoRts | 6:26162 | DOI: 10.1038/srep26162 inflammasome somehow did not protect macrophages exposed to ITO-NPs against cell death. Thus, we sought an approach with MSCs to control both arms of pyroptosis -inflammation and cytolysis while preventing apoptosis.
Here, we show that MSCs can prevent pyroptosis in alveolar macrophages and monocyte-derived macrophages induced by ITO-NPs, SiO 2 -NPs and MSU crystals. MSCs cooperate with macrophages containing endosomes with ITO-NPs likely via cell-to-cell contact as assessed by TEM, ultimately resulting in decreased pro-inflammatory cytokines release, cell death and enhanced IL-10 immunosuppressive cytokine production. Hence, IL-10 is critical for both reducing inflammation and favoring tissue healing 61,62 . Most importantly, MSCs reduced cytolysis of macrophages exposed to ITO-NPs without increasing apoptosis. The anti-apoptotic properties of MSCs towards immune and non-immune cells are well established [63][64][65] . Here, we show that MSCs possess potent anti-pyroptotic effect in macrophages.
Altogether, our findings indicate that respiratory diseases induced by ITO-NPs can be mediated NLRP3-ASC assembly and caspase1-dependent pyroptosis in macrophages, and identifies MSCs as promising therapeutic to suppress pyroptosis (Supplemental Fig. S11). These conclusions may be extended to other inflammatory disorders involving pyroptosis such as in silicosis and gout. Further investigations on the molecular pathways involved in MSCs for preventing pyroptosis in cells are critical to improve MSC-based therapeutic settings in inflammatory diseases and in regenerative medicine. Cells. Peritoneal cells were from naïve Balb/c mice and seeded at 2 × 10 6 cells/mL complete medium, allowed to adhere in culture dishes for 1 h in an incubator. Adherent cells were peritoneal macrophages assessed as CD11b + F4/80 + cells by flow cytometer. Splenocytes were from spleen and bone marrow cells were obtained by flushing tibiae and femora of naïve Balb/c mice. Bone marrow-derived macrophages (BMDMs) were obtained by culture of bone marrow cells in a BMDM medium consisting of RPMI 1640, 10% FBS, 1% penicillin/streptomycin, 0.01% 2-β mercaptoethanol (Sigma-Aldrich), and 20 ng/mL of recombinant murine GM-CSF (Biolegend). After 4 days, BMDMs medium was replaced, and adherent cells were harvested after 8 days as BMDM. MH-S and THP-1 wild-type (WT), both from ATCC, were used as alveolar (mouse) and monocyte-like (human) macrophages. THP-1 cells knockdown in NLRP3, ASC and caspase1 were from Invivogen. Cells were cultured in RPMI 1640 complete medium. Hygromycin B selective antibiotic (Sigma-Aldrich) was used at 100 μ g/mL for knockdown THP-1 cells. Lipopolysacharride (LPS) from Escherichia coli 0111:B4 (Sigma-Aldrich) was used at 100 ng/mL in some experiments. Phorbol 12-myristate 13-acetate (PMA) from Sigma-Aldrich, was used at 20 ng/ mL to activate THP-1 cells; about 1 h before experiments.

Mesenchymal stem cells (MSCs) and co-culture with macrophages.
Approval for collecting procedures and use of human biological samples were obtained from the medical ethics committee at the University Hospital of Toulouse (France), with written informed consents obtained from all donors. Human marrow cells were from femoral heads adults undergoing orthopedic surgery. Mononuclear cells contained in marrow cells were isolated by Ficoll-Histopaque 1077 (Sigma-Aldrich), plated in culture flasks at 10 5 cells/cm 2 , in α -MEM (Gibco), 10% heat-inactivated FBS, and 1 ng/mL human recombinant basic fibroblast growth factor (b-FGF, RD Systems). The culture was depleted of hematopoietic cells by successive medium removal made every other day. A confluent spindle/fibroblastic cell monolayer was obtained about 15 day after initial plating and were considered as MSCs passage 0 (P0). MSCs were used at P4 to P8 for phenotyping, characterization and co-culture experiments. MSC were allowed to adhere before co-culture with macrophages in RPMI 1640 complete medium.
Cell viability. Cells were plated at 10 5 cells per well in a final volume of 200 μ L. CCK-8 solution (Sigma-Aldrich) was added then plates were incubated for 4 h and read on a microplate reader (Multiskan, Thermo Scientific) at 450 nm. Data were obtained as OD and given as % of control, with control = cell in complete medium and, % of control = (OD sample/OD mean of control) ×100. LDH release. 10 5 cells/well were stimulated in round bottom 96 well-plates in a volume of 200 μ L. 100-μ L of supernatant from each well was transferred into flat-bottom 96-well plates, then catalyst and dye solutions were added (LDH Detection Kit, Takara). Then 100 μ L HCl 1N/well were added, and plates were read on a microplate reader (Multiskan, Thermo Scientific) at 450 nm. Results are given as % of maximum release, with medium = no cells and TX-100 = cells + 1% Triton X-100, finally % of maximum release = (OD sample − OD medium)/(OD TX-100 − OD medium) ×100.
Transmission electronic microscopy (TEM). Cellular samples were fixed in 0.9% NaCl, 1% glutaraldehyde (Wako) for 1 h at 4 °C. Cells were pelleted and re-suspended in 1 mL of a 37 °C pre-warmed 1.1% agarose solution (Bio-Rad). After centrifugation (3000 rpm at 37 °C), pellets were placed at 4 °C and cut into small specimens, washed in 5% sucrose PBS (Wako) and post-fixed with 5% sucrose, 1.5% osmic acid (Nisshin EM) in 100 mM phosphate buffer (Wako) pH 7.3, for 1 h at 4 °C. Post-fixed specimens were washed in 5% sucrose PBS and dehydrated in a graded series of ethanol (Wako) and propylene oxide (Wako), then embedded in epoxy resin EPON 812 (TAAB Laboratories). Ultrathin sections were obtained with an ultra-microtome (Leica EM UC7) and subjected to double staining with a uranyl-acetate (Wako) and lead-citrate (Nacalai Tesque). Samples were observed under a JEM-1400Plus Electronic Microscope (JEOL). TEM were analyzed by collecting images of at least 3 representative areas (RA) at magnification of ×500. One RA contained 10-20 nucleated cells. Then representative images were taken at higher magnifications.
Total protein extraction and titration. Dry cell pellets were obtained after PBS washing and centrifugation of tested cells (1500 rpm, 8 min at 4 °C). Tris-HCl lysis buffer consisting of 50 mM Tris-HCl, 300 mM NaCl and 0.5% Triton X-100, pH 7.6, was added, 250 μ L of lysis buffer/pellet. Cells were vortexed and kept for 1.5 h on ice. Cell lysates were sonicated with 3 pulses of 5 s, minimum power (Ultrasonic, UTH-80), and centrifuged at 16,000 g for 10 min at 4 °C. The cell extracts were placed into new test tubes. 10 μ l of cell extracts was diluted into 990 μ l PBS and used for determination of total protein by Bradford assay (Micro BCA, Thermo Scientific). Cell extracts were diluted with PBS from 1/4 to 1/8 and tested in ELISA for a given cytokine. ELISA data are given as picogram (intracellular cytokine) per milligram (pg/mg) total protein.

ELISA.
Samples for ELISA were supernatants or cell lysates of cells stimulated under conditions with cells plated at 10 6 cells in 1 mL final volume per well, in 24-well plates. Cell culture supernatants or cell lysates or biological fluids from mice were assayed for mouse IL-1β , TNF-α and IL-10 (Mouse ELISA MAX Deluxe, Biolegend). Cell culture supernatants or cell lysates were assayed for human IL-1α , IL-1β , TNF-α and IL-10 (Human ELISA MAX Deluxe, Biolegend). Statistical analysis. Data are given as mean ± SEM, unpaired t test or one-way ANOVA with a multiple comparison test were performed using GraphPad Prism 5.0 (GraphPad Software, Inc). P < 0.05 was considered statistically significant.
Ethics statement. Approval for collecting procedures and use for research purposes of bone-marrow samples were obtained from the medical ethics committee (CPP# DC-2012-1612) at the University Hospital of Toulouse (France), and written informed consents were obtained from all donors. The institutional animal care and use committee at Kochi Medical School (Japan) reviewed and approved all animal experiments in this study. Animal care, use, and treatment were strictly in compliance with guidelines and regulations.