TY - JOUR
AU - Topsakal, Erdem
AB - Abstract This study evaluated the cytocompatibility of single- and poly-crystalline ZnO thin films using extract and direct contact methods. Exposure to poly-crystalline ZnO extract resulted in reduced cell viability, on average 82%/70% as measured by MTS/LDH assays, respectively. Direct exposure to both single- and poly-crystalline ZnO thin films resulted in reduced cell viability, which was attributed to anoikis due to inhibition of cell adhesion to the substrate by zinc. Intracellular zinc imaging suggests that single crystalline ZnO thin films do not result in a significant change in intracellular zinc concentrations. Overall, the results suggest that single-crystalline ZnO thin films have better short-term (24 h) cytocompatibility and support their potential to serve as a biocompatible sensor material. Graphical Abstract Open in new tabDownload slide This study evaluated the cytocompatibility of single- and poly-crystalline ZnO thin films using extract and direct contact methods. Introduction Continuous health monitoring using miniature implantable sensors would be immensely beneficial to patients suffering from chronic diseases, such as type I diabetes, and to medical professionals seeking new approaches to monitor and combat these diseases on a wider scale. Current challenges to the development of implantable sensors include long-term sensor functionality and material biocompatibility. A material gaining attention for sensor applications is zinc oxide (ZnO) due to its unique set of properties, which include a high isoelectric point (IEP = 9.5), electrochemical activity, and biocompatibility in the bulk form.1,2 ZnO has a long history of sensor applications in the form of bulk crystals, thick films, and sintered pellets.3 Recently ZnO-based nanostructures, such as nano-thin films, nanowires, and nanoparticles, have attracted particular attention owing to their high sensitivity to chemical species facilitated by their large surface-to-volume ratio.4 The high sensitivity of nanostructured ZnO to important constituents in blood as well as interstitial fluid, including glucose,5 cholesterol,6 and cortisol,7 is of particular interest for biomedical applications. For instance, the linear operation range for ZnO nanowire based sensors has been demonstrated to be between 0.2 and 415 mg dL−1, which covers the full range of glucose levels in human blood (30–300 mg dL−1).8 The toxicological profile of ZnO thin films is not well established to date despite demonstrations of their biosensing potential. Published data are scarce and vary due to different types of ZnO thin films tested and different in vitro models used. Carrascón et al. reported the formation of agglomerated structures of human mesenchymal stem cells when exposed to ZnO sol–gel film-coated silicon.9 Using RAW 264.7 mouse macrophages, Petrochenko et al. observed time-dependent toxicity for atomic layer deposited, nanotextured, ZnO-coated aluminum oxide membranes10 and significant toxicity for pulsed laser deposited, ZnO thin film-coated silicon.11 The toxicity of nanomaterials is commonly attributed to high reactivity and elevated metal ion concentrations due to decomposition. The chemical activity of metal oxides is dependent on their crystal surface orientation. Hexagonal ZnO, the stable ZnO structure under standard conditions, is composed of alternating zinc and oxygen planes along the direction perpendicular to the basal plane. The surface atomic density and surface polarization charge vary among the crystal planes of different orientations. Therefore, variations in surface orientation, which are linked to reactivity12 and stability, could correlate to variations in cytotoxicity. For instance, higher etching rates in acidic solution were reported for O-polar compared to Zn-polar surfaces.13,14 While the orientations of ZnO thin films and nanostructures sought for implantable biochemical sensors depend on the synthesis method and growth conditions, to the best of our knowledge, the published studies have not specified the surface orientation of the material under investigation. In this study, we examined the in vitro biocompatibility of Zn- and O-polar single-crystalline ZnO thin films and poly-crystalline ZnO thin films. L929 mouse fibroblast cells were exposed to extracts following a standard protocol,15 or grown directly on the samples for 24 h. Two cytotoxicity assays were used: the tetrazolium-based MTS assay to determine changes in metabolic activity, and the lactose dehydrogenase (LDH) assay to determine changes in cell membrane integrity. Little to no reduction in cell viability was noted for all but the polycrystalline ZnO/Si extract samples. Low cell adhesion with reduced metabolic activity and no significant changes in cell membrane integrity were observed when cells were grown in direct contact with samples. The extent of this effect was reduced when the samples were incubated in cell culture media before plating cells. Materials and methods Growth of ZnO thin films Single-crystalline ZnO thin films were grown on GaN(0001)/c-sapphire templates following the methods previously reported.16 Briefly, Ga-polar (0001) GaN was deposited onto c-plane sapphire by metal–organic chemical vapor deposition. After cleaning in aqua regia and aqueous HCl in order to remove potential metal contamination, the GaN templates were loaded into a plasma enhanced molecular beam epitaxy (PE-MBE) system wherein the GaN surface was further thermally cleaned followed by exposure to a Zn beam. A low-temperature nucleation layer was grown at T = 300 °C during which the polarity was controlled at the nucleation stage via the oxygen-to-Zn(vi/ii) flux ratio. The films nucleated under highly O-rich conditions (vi/ii > 1.5) were O-polar, while the material nucleated under oxygen poor conditions resulted in the Zn-polarity (see Ullah et al. for more details).16 This process was followed by an annealing step at 730 °C for 5 min before the final ZnO growth at 670 °C for 2 h was performed. For depositing polycrystalline ZnO films, the Si wafers were cleaned using the RCA 1 and the RCA 2 procedures prior to loading into the PE-MBE chamber. The deposition was performed at 200 °C without special surface preparation under O-rich conditions. In vitro biocompatibility testing L929 cells (mouse fibroblast cell line; American Type Culture Collection, ATCC) were cultured at 37 °C in a >90% relative humidity, 5% CO2 atmosphere using complete cell culture media which consisted of Dulbecco's Modified Essential Medium supplemented with 1% penicillin/streptomycin and 10% fetal bovine serum. The cell line was chosen based on the ISO-10993-5 protocol for biocompatibility testing of medical devices.15 For liquid extract testing, cells were seeded in 96 well plates at a density of 20 000 cells per well or 62 500 cells per cm2 and incubated for 24 h. Afterwards, the culture medium was removed, and the cells were exposed for 24 h to either serial dilutions of latex extract (positive control) or ZnO sensor extract in complete cell culture media. Extracts were prepared by incubating latex or the test samples for 24 h at 37 °C in 2 mL of complete cell culture media following a standard protocol.17 Unexposed cells served as a negative control. Triplicate wells were tested per exposure condition and three independent experiments were performed for each sample. For direct contact testing, cells were plated at a density of 50 000 cells per cm2 directly onto 0.4–0.5 cm2 samples, which included glass (negative control), latex (positive control), uncoated GaN(0001)/sapphire, O-polar ZnO thin film coated GaN(0001)/sapphire, Zn-polar ZnO thin film coated GaN(0001)/sapphire, uncoated silicon, and polycrystalline ZnO thin film coated silicon. The effect of sample pre-incubation in cell culture media for 24 h at 37 °C was also assessed. For both extract and direct contact exposure conditions, cell viability was measured after 24 h exposure using the LDH (CytoTox-ONE Homogeneous Membrane Integrity Assay, Promega Corporation) and MTS (CellTiter 96® AQueous One Solution Cell Proliferation Assay, Promega Corporation) assays. Fluorescence (ex 560 nm/em 590 nm) and absorbance (490 nm) readings were measured using a Cytation 3 imaging plate reader (Biotek). To determine intracellular zinc content, a separate set of direct contact samples was prepared for staining with FluoZin-3 AM (Thermo Fisher Scientific). After 24 h exposure, the cell culture media was removed and 0.5 mL of 1 mM FluoZin-3 dye in phosphate-free HEPES buffered Hanks Balanced Salt Solution (HHBSS) was added to each well containing a sample. The samples were incubated in the dark at room temperature for 30 min. After incubation, the dye solution was removed, and 0.5 mL of HHBSS was added to each well. Fluorescence images (ex 465 nm/em 525 nm) were taken using a Cytation 3 imaging plate reader. Results and discussion For the single-crystalline thin films, the total thickness was measured by Alpha-step Stylus Profiler to be between 175 and 365 nm, depending on polarity. All ZnO films featured atomically flat surfaces. Root mean square surface roughness for the single-crystal ZnO films was <1 nm from atomic force microscopy measurements. Polycrystalline ZnO films of around 400 nm in thickness as measured by the step profiler were deposited on Si(111) wafers. After 24 h exposure to serial dilutions of the liquid extracts from the uncoated GaN(0001)/sapphire, O-polar ZnO coated GaN(0001)/sapphire, and Zn-polar ZnO coated GaN(0001)/sapphire samples, no statistically significant change in cell viability was measured by the MTS and LDH (Fig. 1) assays compared to unexposed cells. A statistically significant increase in cell membrane damage was measured using the LDH assay for polycrystalline ZnO thin film coated silicon. This increase was also apparent for the uncoated silicon but was not statistically significant (Fig. 1). Fig. 1 Open in new tabDownload slide MTS (top) and LDH (bottom) results for L929 cells after 24 h exposure to increasing concentrations of sample extracts (n = 3). Statistically significant (p < 0.05) reductions in both metabolic activity and cell membrane integrity were measured for ZnO/Si as compared to the negative control. Visual inspection of the cells revealed morphological differences among cells grown directly on the ZnO coated samples with and without media pre-incubation (Fig. 2). Among the samples with no pre-treatment, less cell attachment and cell spreading was observed for the ZnO coated samples (Fig. 2, bottom row). These differences are less pronounced in the ZnO coated samples pre-incubated in cell culture media (Fig. 2, top row). These changes were consistent between the single-crystalline and polycrystalline ZnO coated samples (data not shown). Fig. 2 Open in new tabDownload slide Morphological differences between cells grown directly on samples with (top row) and without (bottom row) pre-incubation in cell culture media. Cells attached and spread onto the substrates when pre-incubated with media before seeding. Without pre-incubation with media, cells appear round and apoptotic after 24 h exposure to latex, O-polar, and Zn-polar ZnO thin film-coated samples. Magnification = 20×. Scale bar = 100 μm. The viability of the cells grown in direct contact with the samples with and without pre-treatment was further assessed using the MTS and LDH assays. Compared to cells grown on glass, reduced cell viability was observed for O-polar, Zn-polar, and poly-crystalline ZnO coated samples as measured by the MTS assay. This reduction in cell viability was less pronounced in the O-polar and poly-crystalline ZnO thin film coated samples that were pre-incubated in cell culture media. Membrane integrity measured by the LDH assay was comparable to the control samples for the O-polar ZnO coated samples after pre-incubation in cell culture media (Fig. 3). Fig. 3 Open in new tabDownload slide MTS (top) and LDH (bottom) results for L929 cells after 24 h direct contact exposure to samples (n = 3). Latex serves as a (+) control and percentages are with reference to cells grown on glass (−) control. Statistically significant (p < 0.05) improvements in cell viability were measured for O-polar single-crystalline and poly-crystalline ZnO thin films. FluoZin-3 staining of cells directly grown on the samples revealed similar intracellular zinc levels in the negative control (glass), the uncoated GaN/sapphire, the ZnO thin film coated GaN/sapphire (Fig. 4), the uncoated silicon, and the ZnO thin film coated silicon. Fig. 4 Open in new tabDownload slide Cytosolic zinc staining with FluoZin-3 in L929 cells grown directly on samples. Magnification = 20×. Scale bar = 100 μm. Cytotoxicity measured from extract testing involves dissolved chemicals that have leached from the sample during the extraction process. In the case of the ZnO thin film coated GaN/sapphire wafers, dissolution of the ZnO thin film could result in potentially toxic Zn2+ concentrations. Our extract MTS and LDH results suggest that either little dissolution occurred within 24 h or that the Zn2+ concentration apparently did not exceed the threshold of toxicity. While some have observed low dissolution rates of ZnO thin films in biological media with neutral pH values (∼7.4),18 there are also reports of substantial degradation rates of ZnO thin films. Han et al. reported that ZnO thin films can dissolve within a few hours in deionized water (pH ≈ 4.5–5.0), ammonia (pH ≈ 7.0–7.1, 8.7–9.0), and NaOH solution (pH ≈ 7.0–7.1, 8.7–9.0).19 Singh et al. reported that only 46% and 25% of the zinc remained in the film after 25 h in water and phosphate buffered saline, respectively.20 It should be noted that surface orientations of the thin films were not specified in these studies. To explain these seemingly conflicting observations, the dissolution rate and solubility of different ZnO crystallographic surfaces and potential variations in the material quality should be explored and taken into account. The intracellular zinc staining results suggest that little to no ZnO dissolution occurred. Comparisons between the ZnO coated sample, uncoated sample and the glass control revealed no differences. Instead, poor cell adhesion to the ZnO coated samples likely contributed to the reduced cell viability observed in the direct contact samples. The L929 cell line is an adherent cell line, making the cell survival and proliferation anchorage-dependent. Anoikis, or programmed cell death induced by inadequate cell–matrix interactions, can occur in anchorage-dependent cells if they remain detached from the extracellular matrix or culture substrate.21 Inhibition of cell adhesion by zinc via interaction with integrins, matrix receptor proteins in the cell membrane, has been reported as a possible mechanism.22 Binding of integrins depends on extracellular divalent cations, including Zn2+. For bulk ZnO substrates, a lack of cell adhesion was not reported for direct contact testing using the NIH 3T3 fibroblast, human umbilical vein endothelial, and bovine capillary endothelial cells.23 However for bulk metallic zinc, cell lysis was reported for human endothelial, human aortic smooth muscle, and human dermal fibroblast cells grown directly on metallic zinc foils.22 As mentioned before, the surface bonding structure and the associated chemical reactivity of zinc as well as the crystallinity may play a role in this apparent spectrum of results between bulk metallic zinc, ZnO thin films and bulk ZnO substrates. When the samples were pre-incubated with cell culture media for 24 h before conducting 24 h direct contact testing, the observed lack of cell adhesion and related cytotoxicity were reduced. We hypothesize that adsorption of serum proteins altered the surface of the wafers to enhance cell attachment. Reports on direct contact testing with metallic zinc wires support this hypothesis as protein layers were attributed to the biocompatibility measured when metallic zinc wires were implanted in vivo.22,24 Depending on the application of the ZnO thin films, cell attachment may or may not be desired when considering biofouling. Few groups have investigated the cytotoxicity of ZnO thin films. Carrascon et al. exposed human mesenchymal stem cells to ZnO sol–gel films deposited onto silicon wafers.9 Similarly, they reported the formation of agglomerated structures due to poor adhesion. Petrochenko et al. observed a time-dependent toxicity of atomic layer deposited ZnO coated membranes.10 No cytotoxicity was observed after 24 h of exposure but reduced cell viability was measured at 48 h based on the MTT assay. Petrochenko et al. later performed both leachate and direct contact testing of pulsed laser deposited ZnO thin films on silicon wafers.11 They also reported poor cell attachment of RAW 264.7 cells to the substrates which corresponded to a significant toxicity as measured by the MTT assay. In contrast to our results, macrophages exposed to leachates also exhibited significant cytotoxicity based on the MTT and LDH assays. This was attributed to the elevated zinc concentrations measured after 24 h and 48 h incubation in culture media. An influential parameter appears to be ZnO thin film stability, which is dependent upon the thin film microstructure (fine or coarse grains), crystallinity (single-crystal or polycrystalline material), and even fabrication method (sol–gel, atomic layer deposition, pulsed laser deposition, molecular beam epitaxy), most likely through defect density. This reinforces the need to assess manufacturing approaches when fabricating nanomaterials for biomedical applications and suggests that improved biocompatibility can be achieved through refined fabrication. Conclusions Zn- and O-polar single-crystalline and poly-crystalline ZnO thin films grown by molecular beam epitaxy exhibit short-term (24 h) cytocompatibility. Extract testing revealed no significant changes in cell viability. Although reduced cell viability was observed when cells were plated on samples without pre-incubation, incubation in culture media prior to direct contact exposures reduced the inhibition of cell attachment to the samples. This suggests that biocompatibility can be enhanced with pre-processing steps. Comparison with other studies suggests that the manufacturing/preparation method can also be modified to improve biocompatibility. Future work will examine the correlations between the material quality, surface properties, and stability/dissolution, and will explore toxicity for longer exposure durations. Conflicts of interest There are no conflicts to declare. Acknowledgements This work was supported in part by startup funds provided to Dr Lewinski by the School of Engineering at Virginia Commonwealth University and the Commonwealth Health Research Board (CHRB) grant no. FP00001405. References S. K. Arya , S. Saha , J. E. Ramirez-Vick , V. Gupta , S. Bhansali and S. P. Singh , Recent advances in ZnO nanostructures and thin films for biosensor applications: Review , Anal. Chim. Acta , 2012 , 737 , 1 – 21 . Google Scholar Crossref Search ADS PubMed WorldCat A. Wei , L. Pan and W. Huang , Recent progress in the ZnO nanostructure-based sensors , Mater. Sci. Eng., B , 2011 , 176 , 1409 – 1421 . Google Scholar Crossref Search ADS WorldCat A. Jones , T. A. Jones , B. Mann and J. G. Firth , The effect of the physical form of the oxide on the conductivity changes produced by CH4, CO and H2O on ZnO , Sens. Actuators , 1984 , 5 , 75 – 88 . Google Scholar Crossref Search ADS WorldCat N. Izyumskaya , A. Tahira , Z. H. Ibupoto , N. Lewinski , V. Avrutin , Ü. Özgür , E. Topsakal , M. Willander and H. Morkoç , Review—Electrochemical Biosensors Based on ZnO Nanostructures , ECS J. Solid State Sci. Technol. , 2017 , 6 , Q84 – Q100 . Google Scholar Crossref Search ADS WorldCat M. Ahmad , C. Pan , Z. Luo and J. Zhu , A Single ZnO Nanofiber-Based Highly Sensitive Amperometric Glucose Biosensor , J. Phys. Chem. C , 2010 , 114 , 9308 – 9313 . Google Scholar Crossref Search ADS WorldCat S. P. Singh , S. K. Arya , P. Pandey , B. D. Malhotra , S. Saha , K. Sreenivas and V. Gupta , Cholesterol biosensor based on rf sputtered zinc oxide nanoporous thin film , Appl. Phys. Lett. , 2007 , 91 , 063901 . Google Scholar Crossref Search ADS WorldCat R. D. Munje , S. Muthukumar , A. Panneer Selvam and S. Prasad , Flexible nanoporous tunable electrical double layer biosensors for sweat diagnostics , Sci. Rep. , 2015 , 5 , 14586 . Google Scholar Crossref Search ADS PubMed WorldCat M. Ahmad , C. Pan , L. Gan , Z. Nawaz and J. Zhu , Highly Sensitive Amperometric Cholesterol Biosensor Based on Pt-Incorporated Fullerene-like ZnO Nanospheres , J. Phys. Chem. C , 2010 , 114 , 243 – 250 . Google Scholar Crossref Search ADS WorldCat R. Delgado Carrascón , D. Gallach Pérez , J. P. Carcía Ruiz and M. Manso Silvan , Compared Biocompatibility of ZnTiO3, ZnO and TiO2 Sol-Gel Films with Human Mesenchymal Stem Cells , MRS Adv. , 2016 , 1 , 737 – 742 . Google Scholar Crossref Search ADS WorldCat P. E. Petrochenko , S. A. Skoog , Q. Zhang , D. J. Comstock , J. W. Elam , P. L. Goering and R. J. Narayan , Cytotoxicity of cultured macrophages exposed to antimicrobial zinc oxide (ZnO) coatings on nanoporous aluminum oxide membranes , Biomatter , 2013 , 3 , e25528 . Google Scholar Crossref Search ADS PubMed WorldCat P. E. Petrochenko , Q. Zhang , R. Bayati , S. A. Skoog , K. S. Phillips , G. Kumar , R. J. Narayan and P. L. Goering , Cytotoxic evaluation of nanostructured zinc oxide (ZnO) thin films and leachates , Toxicol. in Vitro , 2014 , 28 , 1144 – 1152 . Google Scholar Crossref Search ADS PubMed WorldCat S. A. Ansari , M. M. Khan , S. Kalathil , A. Nisar , J. Lee and M. H. Cho , Oxygen vacancy induced band gap narrowing of ZnO nanostructures by an electrochemically active biofilm , Nanoscale , 2013 , 5 , 9238 – 9246 . Google Scholar Crossref Search ADS PubMed WorldCat M. Mehta and C. Meier , Controlled etching behavior of O-polar and Zn-polar ZnO single crystals , J. Electrochem. Soc. , 2011 , 158 , H119 – H123 . Google Scholar Crossref Search ADS WorldCat V. Petukhov , A. Bakin , A.-H. El-Shaer , A.-C. Mofor and A. Waag , Etch-pit density investigation on both polar faces of ZnO substrates , Electrochem. Solid-State Lett. , 2007 , 10 , H357 – H361 . Google Scholar Crossref Search ADS WorldCat E. ISO , 10993-5. Biological evaluation of medical devices-Part 5: Tests for in vitro cytotoxicity , International Organization for Standardization , Geneva , 2009 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC M. B. Ullah , V. Avrutin , S. Q. Li , S. Das , M. Monavarian , M. Toporkov , Ü. Özgür , P. Ruterana and H. Morkoç , Polarity control and residual strain in ZnO epilayers grown by molecular beam epitaxy on (0001) GaN/sapphire , Phys. Status Solidi RRL , 2016 , 10 , 682 – 686 . Google Scholar Crossref Search ADS WorldCat E. ISO , 10993-12. Biological evaluationof medical devices–Part 12: Sample preparation and reference materials (ISO 10993-12: 2007) , International Organization for Standardization , Geneva , 2008 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC S. Liu , E. Killen , M. Lim , C. Gunawan and R. Amal , The effect of common bacterial growth media on zinc oxide thin films: identification of reaction products and implications for the toxicology of ZnO , RSC Adv. , 2014 , 4 , 4363 – 4370 . Google Scholar Crossref Search ADS WorldCat J. Han , W. Qiu and W. Gao , Potential dissolution and photo-dissolution of ZnO thin films , J. Hazard. Mater. , 2010 , 178 , 115 – 122 . Google Scholar Crossref Search ADS PubMed WorldCat M. Singh , G. Palazzo , G. Romanazzi , G. P. Suranna , N. Ditaranto , C. Di Franco , M. V. Santacroce , M. Y. Mulla , M. Magliulo , K. Manoli and L. Torsi , Bio-sorbable, liquid electrolyte gated thin-film transistor based on a solution-processed zinc oxide layer , Faraday Discuss. , 2014 , 174 , 383 – 398 . Google Scholar Crossref Search ADS PubMed WorldCat S. M. Frisch and R. A. Screaton , Anoikis mechanisms , Curr. Opin. Cell Biol. , 2001 , 13 , 555 – 562 . Google Scholar Crossref Search ADS PubMed WorldCat E. R. Shearier , P. K. Bowen , W. He , A. Drelich , J. Drelich , J. Goldman and F. Zhao , In Vitro Cytotoxicity, Adhesion, and Proliferation of Human Vascular Cells Exposed to Zinc , ACS Biomater. Sci. Eng. , 2016 , 2 , 634 – 642 . Google Scholar Crossref Search ADS PubMed WorldCat J. Lee , B. S. Kang , B. Hicks , T. F. Chancellor Jr. , B. H. Chu , H.-T. Wang , B. G. Keselowsky , F. Ren and T. P. Lele , The control of cell adhesion and viability by zinc oxide nanorods , Biomaterials , 2008 , 29 , 3743 – 3749 . Google Scholar Crossref Search ADS PubMed WorldCat P. K. Bowen , R. J. Guillory Ii , E. R. Shearier , J.-M. Seitz , J. Drelich , M. Bocks , F. Zhao and J. Goldman , Metallic zinc exhibits optimal biocompatibility for bioabsorbable endovascular stents , Mater. Sci. Eng., C , 2015 , 56 , 467 – 472 . Google Scholar Crossref Search ADS WorldCat Footnotes † Electronic supplementary information (ESI) available. See DOI: 10.1039/c8tx00061a © The Author(s) 2018. Published by Oxford University Press on behalf of Royal Society of Chemistry. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model) © The Author(s) 2018. Published by Oxford University Press on behalf of Royal Society of Chemistry. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com
TI - Influence of ZnO thin film crystallinity on in vitro biocompatibility
JF - Toxicology Research
DO - 10.1039/c8tx00061a
DA - 2018-09-01
UR - https://www.deepdyve.com/lp/oxford-university-press/influence-of-zno-thin-film-crystallinity-on-in-vitro-biocompatibility-hU6VHfmQyT
SP - 754
EP - 759
VL - 7
IS - 5
DP - DeepDyve
ER -