1 Fortsatt utveckling och validering av screeningtester för evaluering av industriellt och miljö producerade partiklars toxicitet (diarienr 150144 - Bo Nilsson) Aim The original aim of the study was to apply the whole blood technology that we previously have developed to elucidate the inflammatory responses of bloodcontacting biomaterials (for background see original application). Since sustained inflammation leads to cancerogenicity, genotoxicity and arteriosclerosis, we intended to delineate inflammatory processes that are induced by nanoparticles (NP). The technology has the advantage to be applicable to human tissue (any species can be used) and can be used characterize the mechanisms by which NP trigger the innate immune system of the blood (primarily the cascade systems of the blood), delineate the cross talk between NP and innate immune system that leads to the biological response and develop in vitro screening techniques to assess the toxicity of NP in contact with tissue fluids/blood plasma/whole blood. Our detailed aims were: 1. Identify the proteins of the cascade systems bound to NP after exposure to blood plasma and investigate activation mechanisms. Of particular interested is proteins of the complement (e.g. C3, C4), contact (e.g. FXII, HMWK) and the coagulation (e.g. fibrinogen) systems. 2. Assess the level of activation of the complement, coagulation and contact systems in blood plasma and whole blood elicited by NP. 3. Evaluate the biological response in whole blood with regard to erythrocytes, platelets and leukocytes. 4. Gauge the biological response of endothelial cells in contact with whole blood in the presence of NP. Results Overall the project has closely followed the initial aims. In order to study three different blood models have been developed (a-c) which all can be used as screening models of NP toxicity: Fig. 1: The blood model consists of tubing loops that are coated with heparin on the inside. Freshly drawn human blood, without the addition of anticoagulants, is added to the loop together with the test particles. The tubing is closed to form loops with a custom-made connector and then rotated at 37 C in a heat cabinet. Aliquots of blood are removed at different time points during incubation for analysis of activation markers of the innate immunity systems and cells of the blood.
2 1. Development of in vitro blood models: a. The whole blood loop model. Using our previous technology to incubate cells and biomaterials in contact with blood in the absence of anticoagulants 1, we have designed a test model for evaluating the toxicity of NPs (Fig. 1) 2. It allows analysis of any kind of cross talk between blood cells and plasma proteins as reflected in cascade system (complement, contact, coagulation, fibrinolysis systems activation etc) parameters, other plasma protein alterations and cell phenotypes (flow cytometry, cyto/chemokine generation, protein release etc). In its strictest application, it tests only direct NP-to-blood contact, but since plasma proteins are present in most body fluids at various concentrations, it has much wider applications and can be used as a general toxicity test. b. Blood-endothelial cell model. One of the major issues concerning nanoparticle toxicity is the interaction with epithelial and endothelial cells. In order to investigate the interaction between endothelial cells and blood, we have developed a chamber model, which utilizes cultured endothelial cells and freshly drawn non-anticoagulated blood from a donor 3. In the experimental models primary endothelial cell lines reflecting the heterogeneity of the endothelium are utilized, human dermal microvascular endothelial cells (HDMEC), human aortic endothelial cells (HAEC) or human umbilical vein endothelial cells (HUVEC) (Fig. 2). The endothelial cells are cultured until confluence in commercial available glass culture slides. As a demonstrator TiO 2 NPs (P25) was used. It was then shown that HDMEC responded with increased coagulation shown by significant increase in thrombin anti-thrombin (TAT) levels in comparison to HAEC. Fig. 2: The blood endothelial cell (EC) chamber. Schematic figure showing the combination of endothelial cells on glass culture slides with the blood chamber, connected and incubated on rotation in 37 C water bath (A). TAT and C3a levels in chambers with human dermal microvacular EC (HDMEC) and human aortic EC (HAEC) with and without nanoparticles (B). Clotting on HDMEC with nanoparticles an effect absent without nanoparticles (C). Loss of vwf in HDMEC after exposure to nanoparticles in comparison to higher vwfexpression remaining in the absence of nanoparticles (D).
3 Fig. 3: (A). Particles in suspension are incubated in EDTA blood plasma. Plasma proteins are adsorbed to the particles. (B) after wash the particles atre incubated with anti-c4 and anti- C4BP which are labeled for instance with an enzyme or a fluorescence probe. (C) After an additional wash, the relative binding of C4 and C4BP is quantified and the ratio between these proteins is calculated. The ratio C4/C4BP reflects the down-stream cyto/chemokine release induced by the particles in during incubation in whole blood. c. Protein fingerprinting. We examined protein adsorption (protein finger printing) in plasma and complement activation/cytokine release in whole blood incubated with well-characterized polymer particles 4. Strong correlations were found between the ratios of adsorbed plasma proteins C4/C4BP and FXIIa/ C1INH after exposure of the material to blood (EDTA) plasma and 10 (mainly pro-inflammatory) cytokines, including IL-17, IFN-g, and IL-6. We thereby demonstrate for the first time that there is a direct correlation between downstream biological effects and the proteins initially adhering to an artificial surface after contact with blood. The outline of a simple and robust protocol for NP testing is presented in (Fig. 3). By studying the protein finger print, we have also described a fundamentally new activation mechanism of complement component C3, which does not involve any other complement components or proteolytic cleavages by other proteases 5, 6. We believe that this type of contact activation is a common feature on various materials and, in particular, nanomaterial surfaces, which potentially can facilitate uptake of NPs by phagocytic cells and thereby help the take up of NPs into the body. Fig. 4: The work flow applied in order to evaluate and validate the in vitro models
4 Testing of various types of NPs: a. We evaluated the toxicity of TiO 2 NPs considered to have relatively low toxicity and to be inert in several applications. In the whole blood loop model (Fig. 1 and 2), we stepwise identified that the TiO 2 NPs at very low concentrations activate FXII and the contact system, and demonstrated that the TiO 2 NPs are highly thrombogenic and prone to induce inflammation 2. The steps that were evaluated were: NP physicchemical characterization, identification of proteins in the corona, activation of FXII, coagulation activation, platelet activation and finally induction of cytokine/chemokines (Fig. 4). All these events were verified using a specific FXIIa inhibitor, demonstrating the thromboinflammation was propelled by the contact/kallikrein system. b. In other studies using the whole blood loop model we have demonstrated that unlike the TiO 2 NPs 2, iron oxide 7. and diesel particles do not trigger the contact/kallikrein system to the same extent. Despite these observations both type of particles are thrombogenic with high levels of TAT generation and platelet consumption (Fig. 5). This is a striking difference, which probably reflects a different mechanism of activation. One mechanism by which these particles may elicit thrombogenic activation is that they are phagocytized by phagocytic cells (in the blood granulocytes and monocytes), which thereby activate and release tissue factor (TF; the physiological activator of the coagulation system). This uptake of particles or particle aggregates may be accelerated by the binding of C3 as described below (see above c). Another interesting finding is that the environment) that contains the iron oxide particles greatly influences the biological effect. Iron oxide particles in saline bind much more proteins and activates the coagulation system much stronger than the particles in a phosphate buffer. Analyzes of these two particles continues, particularly, with studies of potential phagocytosis by and oxidative stress of the granulocytes induced by both Fig. 5: Comparison of TiO 2 and Fe 2O 3 NPs with regard to FXIIa activation and platelet consumption. The TiO 2 NPs activated FXII in a dose dependent fashion, while Fe 2O 3 NPs did not activate FXII in an NaCl environment. If PO4 was added a robust FXII activation was observed. Despite this activation, the TiO 2 NPs induced a much stronger FXII-dependent platelet consumption compared to any of the Fe 2O 3 NP preparations demonstrating a totally different biological activity.
5 Fe 2 O 3 and diesel particles. Meeting with Swedish Chemicals Agency : A meeting with Swedish Chemicals Agency (Kemikalie-inspektionen) took place the 13 th of May 2014. Conclusion The present grant period has been devoted to develop three in vitro blood models, which are candidates for being used as screening assays for evaluation of NP toxicity. In the project we have used these assays to evaluate three types of NPs and other types of particle matter i.e. TiO 2, Fe 2 O 3 and diesel particles (not consistently below 100 nm). All three particles showed distinctive biological responses demonstrating that their potential effects on the body is quite varying and that they will pose distinctive stress profiles on the body. The new assays are now ready to screen various types of NP and PM in order to map their biological response profiles, which will make it possible to correlate the initial response with in vivo adverse reactions. To further improve the generality of the assays, parallel exposure of particles in vitro and in vivo will be performed. This will be done in order to be able to correlate the biological response in the in vitro assays with adverse reactions in vivo both in animal and human models. The results have been published in high impact journals Published articles and manuscripts specific for the project: 1. Ekdahl KN, Hong J, Hamad O, Larsson R and Nilsson B. Evaluation of the blood compatibility of materials, cells and tissues: Basic concepts, test models and practical guidelines. Adv Exp Biol Med. 2013;734:257-270. 2. Ekstrand-Hammarström B, Hong J, Davoodpour P, Sandholm K, Ekdahl K, Bucht A and Nilsson B. TiO2 nanoparticles tested in a novel screening whole human blood model of toxicity trigger adverse activation of the kallikrein system at low concentrations. Biomaterials in press. 2015. 3. Nordling S, Hong J, Fromell K, Edin F, Brännström J, Nilsson B and Magnusson P. Vascular repair utilizing immobilized heparin conjugate for protection against early activation of inflammation and coagulation. Thrombosis Haemostasis in press. 2015. 4. Engberg AE, Nilsson PH, Huang S, Fromell K, Hamad OA, Mollnes TE, Rosengren-Holmberg JP, Sandholm K, Teramura Y, Nicholls IA, Nilsson B and Ekdahl KN. Prediction of inflammatory responses induced by biomaterials in contact with human blood using protein fingerprint from plasma. Biomaterials. 2015;36:55-65. 5. Klapper Y, Hamad OA, Teramura Y, Leneweit G, Nienhaus GU, Ricklin D, Lambris JD, Ekdahl KN and Nilsson B. Mediation of a non-proteolytic activation of complement component C3 by phospholipid vesicles. Biomaterials. 2014;35:3688-96. 6. Klapper, Y, Maffre, P, Shang, L, Ekdahl, KN, Nilsson, B, Hettler, S, Dries, M, Gerthsen, D, Nienhaus, GU Low affinity binding of serum proteins to lipid-coated
quantum dots as observed by in-situ fluorescence correlation spectroscopy. Advanced Science (1 st revision). 7. P Davoodpour, J Hong, Ekstrand-Hammarström B, KN Ekdahl, B Nilsson Iron oxide (Fe 2 O 3 ) nanoparticles at low concentrations induce coagulation without contact system involvement in whole human blood. (Preliminary manuscript) 6