Polymer International 41 (1996) 419-425 Influence of Hard Segments of Polyurethane on Cell Growth P. C. Lee,' L. W. Chen,2* J. R . Lin,' K. H. Hsieh2& L. L. H. Huang3 Institute of Materials Science and Engineering, Department of Chemical Engineering, Center for Biomedical Engineering, National Taiwan University, Taipei, Taiwan (Received 14 May 1996; accepted 4 July 1996) Abstract: Polyurethanes (PU) with suitable soft segments have been found to be good blood-compatible polymers and have attracted much attention recently. In this study, various molar amounts of 4,4'-methylene bisphenyl isocyanate reacted with poly(tetramethy1eneoxide) were synthesized to explore the optimal ratio of hard/soft segments for cell attachment and proliferation in in uitro systems. Differential scanning calorimetry and dynamic mechanical analysis were used to determine the physical properties, hydrogen bonding index (HBI) and transmission electron microscopy to observe the phase-separation phenomena in the materials, and 3T3 fibroblast to evaluate the dependence of the cell proliferation at 37°C on the material properties. Our results show that cell attachment and proliferation are closely related to the cell growth surface, which in turn is controlled by (1) the ratio of hard to total segment concentration and (2) the recrystallization temperature (T,)of PU. To obtain a good cell growth surface, the ratio of hard to total segment concentration is found to be between 0.4 and 0.6, and HBI is between 1.5 and 2.1. Furthermore, when the T, of PU is near the physiology temperature, a stable surface for cell growth can be provided. The shorter molecules in the soft segment region can rearrange the molecular chain at 37°C. K e y words: polyurethane, cell growth, phase separation, hydrogen bonding index, cytotoxicity. INTRO D UCTlON c o - ~ o r k e r s . They ~ , ~ observed selective calcification by the 1000 molecular weight species, in contrast to the higher and lower molecular weight macroglycols, which reduced this behaviour. Thoma et aL6 hypothesized that the PTMO 1000 was capable of forming ring or crown structures, which can effectively allow passage of specific-sized molecules and hence play an important role in the biostability of polyurethane. Based on these results, PTMO 1000 was used as the soft segment in the polyurethane series in this study. A composition ratio of diisocyanate : polyol : chain extender = 2 : 1 : 1 has often been used to assess blood compatibility and cell ~ompatibility.~-'~ However, since we were particularly interested in the weight fraction of hard segment and its effect on cell growth, we changed Among polymers, polyurethane has been considered to have great potential for application in medical devices. Polyurethane is adopted for implant application because of such advantages as its high tensile strength, lubricity, good abrasion resistance, ease of handling and extruding and good 'bi~compatibility'.l-~ In this study, we focus on using poly(tetramethy1ene oxide) (PTMO 1000) to synthesize the polyurethane. In relation to the higher and lower molecular weight polyether macroglycols in polyurethanes, the behaviour of PTMO 1000 has been reported by Phillips, Thoma and * To whom all correspondence should be addressed. 419 Polymer ZnternationalO959-8103/96/$09.00 0 1996 SCI. Printed in Great Britain P. C. Lee et al. 420 the molar ratio of diisocyanate 4,4'-methylene bisphenyl diisocyanate, MOI) as a hard segment, and polyol (PTMO 1000) as a soft segment in this polyurethane series. The correlation of physical properties with cell growth was investigated. EXPERIMENTAL reaction was complete, the polymer solution was put under vacuum for 5min to degas the polymer solution, cast in moulds; cured at 70°C for 24 h, and post-cured at 110°C for 24h. The samples were placed under vacuum at room temperature to remove the residual solvent. All the specimens were conditioned at room temperature and 50% humidity for at least 2 weeks prior to testing. Materials Preparation of polymer film The materials used and their vendors were as follows. 4,4'-Methylene diphenyl diisocyanate (MDI), poly(tetramethy1ene oxide) with average molecular weight 1000 (PTMO lOOO), dimethylformamide (DMF) and 1,Cbutanediol were purchased from Aldrich Chemical Company, Inc. (USA). Round glass coverslips, with diameter of 15mm, were purchased from Matsunami Glass Industries, Ltd (Japan). Dimethylsulphoxide (DMSO), 3-(4,5-dimethyl thiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) and RuO, were purchased from Sigma Chemical Company (USA). DMEM, RPMI medium 1640, fetal bovine serum (FBS), trypsin-EDTA and streptomycin/penicillin were all cell culture grade and were purchased from Gibco BRL Life Technologies, Inc. (USA). The polymer membranes were supported on a 24-well tissue culture to detect cell attachment and proliferation on the polymer surface. Films were cast onto optical coverslips using a 3% solution of polyurethane in DMF. The glass coverslips were first cleaned thoroughly with chromic acid, then doubly distilled water and detergents under sonication, and washed with ethanol and dried under vacuum at 50°C. After coating with polymer solution, the coverslips were dried in air at 70°C for 24 h and kept under vacuum at room temperature for 48 h. They were stored in a desiccator until used. Synthesis of polyurethanes with various molar ratios of hard segments This series of polyurethanes was prepared in degassed DMF solution using a two-stage solution polymerization method at 65°C under dry nitrogen. The molar ratios of MDI : PTMO 1000 : chain extender and their codes are shown in Table 1. A dry three-necked flask was used to prepare the prepolymer. 1,4-Butanediol, as chain extender, and 50wt% degassed DMF were added to promote the polymerization. No catalyst was added. The reaction was continued until the isocyanate functional group had disappeared from the Fourier transform infrared (FTIR) spectra. When the TABLE 1. Symbols of PUS synthesized with various MDI ratios ( M D I = 4.4'-methylene bisphenyldiisocyanate; polyol = PTMO 1000; chain extender = 1.4-butanediol) Symbol MDI : polyol : chain extender CE 1 :0:1 1 :I :o 2 : l :I 4:l :3 6:l :5 8:l :7 1-1 000 2-1 000 4-1 000 6-1 000 817 or 8-1 000 1 019 or 10-1 000 12111 or 12-1 000 110 or 211 or 41 3 or 615 or 10:1:9 1 2 : l :I1 Surface hydrophilicity analysis Analysis of the dynamic contact angle was performed on the polymer-coated film at 37°C by the Wilhelmy plate technique. Dynamic advancing and receding contact angles were calculated from the force-depth curve, based on measurements at advancing and receding rates of 25 mm min-l. Surface tension was measured for fresh films as well as for films soaked in octane at room temperature for 24 h. The instrument was calibrated using a Pt plate. All glass slides used in the experiment were cleaned with chromic acid and then washed several times with doubly distilled water. The size of sample used for testing was 2 cm x 3 cm. Measurements of five films were taken and the results averaged. The samples used as control glass surfaces were as f01lows.~(1) Hydrophilic glass surface: the glass was cleaned with cold chromium sulphuric acid for 24 h and rinsed thoroughly with distilled water. Thereafter the slides were put into 2% hydrofluoric acid for 2min and rinsed with distilled water. The cleaned glasses were stored in doubly distilled water to retain hydrophilicity. (2) Hydrophobic glass surface: dry glass slides were stored in dimethyloctadecylsilane (2% v/v) and washed in chloroform for 24 h. After soaking, the glasses were rinsed with chloroform three times. Dynamic mechanical analysis ( DM A ) and differential scanning calorimetry (DSC) measurements DMA and DSC were performed with a thermomechanical analyser (DuPont 100). Rectangular samples POLYMER INTERNATIONAL VOL. 41, NO. 4, 1996 Injuence of hard segments of PU 2.5in x 0.5in x 0.125in were torqued at 1Hz for DMA measurements. Under a 20mlmin-' dry nitrogen purge, the DSC samples (10 mg) were heated in sealed aluminium pans at 10"Cmin-' from room temperature to 200"C, kept for 3 min, cooled to - 150°C at the same rate, and then reheated to 250°C to obtain the second scan data. Attenuated total reflection (ATR)-FTIR analysis of polyurethane surface A Nicolet FTIR system with a resolution of 4cm-' was used to collect the ATR-FTIR spectra. The absorbance intensity at 17OOcm-' (bonded C=O) ratioed to that at 1730cm-' (free bonded C=O) was used to define the hydrogen bonding index (HBI). Transmission electron microscopy (TEM) observations To prepare specimens for TEM observation, 0.3% polymer solutions were placed on a 200 mesh coated slot grid. Specimens were stained using vaporized RuO, for 30 min and observations, as well as photographs, were taken using a transmission electron microscope (JEM- 1200EXII). Evaluation of cytotoxicity of polyurethane extracts To examine the extracts of polymers from the cell culture, we used the ASTM F 624-93 method. The extracts were first completely immersed in sterilized DMEM at 37°C and then processed by shaking at 1000rpm for 120h in a borosilicate glass tube. After shaking, the extracts were immediately diluted with culture medium at a ratio of nine times its volume, and seeded with the fibroblast 3T3 cell line at a cell density of 5 x lo4 per well in the 24-well plate. The determination of fibroblast activation after contact with the extracts was assessed by the tetrazolium salt test (MTT). The details of MTT are reported in Refs 17 and 18; at the end of the incubation time, the fibroblasts were detached by trypsin and the trypsinization was terminated by DMEM with 10% serum. Then, 180pl/well of trypsinization solution and 20 pl/well of 0.5% MTT in RPMI medium were added to the 96-well plates. After incubation at 37°C for 3h, the culture plates were centrifuged, the supernatant was discarded, and the intracellular formazan crystals were solubilized with 100pl/well of DMSO. The absorbance of samples in each plate was determined at 570nm to obtain the optical density (OD) value. Direct contact cell to polyurethane to evaluate cell attachment and proliferation Polymer-coated glass coverslips, placed in a 24 multiwell cell culture plate, were sterilized with 70% alcohol POLYMER INTERNATIONAL VOL. 41, NO. 4, 1996 42 1 and 30min UV radiation. The 3T3 fibroblast cell line, with a 20th-pass approach was used in this study. Supplemented with 10% fetal calf serum (FCS), streptomycin (50mg1-') and penicillin (5 x lo4 units 1-I) in DMEM as medium, the cells were harvested with 0.15% trypsin/0.02% EDTA in PBS (pH 7.4). The cells were resuspended in the medium at a cell density of 5 x lo4 cells ml-'. The 1ml cell suspension was seeded on the polymer-coated coverslips placed on a 24-well tissue culture plate. After culture for a predetermined time, the cells were washed with PBS twice, with gentle shaking of the plate to remove the serum in the well. The cells were detached with 2 0 0 4 of 0.2% EDTA/ 0.15% trypsin in PBS by incubating at 37°C for 3 min. Then, 6 0 0 ~ 1DMEM/10% FCS was added to stop the trypsization. The cells were pipetted out thoroughly and the viability in suspension was measured by MTT assay. The cell growth rate was compared with that from PS tissue culture wells. Four samples were measured and the standard deviation was calculated. RESULTS AND DISCUSSION Polyurethane characterization In this study, the hard segment denotes the MDI chain extender region from the reaction of isocyanate and the chain extender. The soft segment denotes the major chain of PTMO. The calculated theoretical value of hard segment, HBI, recrystallization exotherm temperature (T,) and glass transition temperature (T,) of hard segment are shown in Table 2. Hydrogen bonding index in polyurethanes. In this study, the HBI was used to identify the interaction between the hard segment of N-H and the C=O group in forming hydrogen bonds in the polymer chain. As shown in Table 2, HBI increases with increase in the ratio of hard segment in PU. Compared with the PU of lower hard segment ratio, the PU of higher hard segment ratio has some intermolecules which exert a TABLE 2. Hard segment weight fraction, T,, Toand HBI in P T M O 1 OOO series PUS Sample Calculated hard segment weight fraction HBI T,("C) T,("C) 110 21 1 41 3 615 817 1019 12111 0.227 0.392 0.575 0.673 0.734 0.776 0.806 1.03 1.54 2.14 2.52 2.93 3.42 3.98 -50 -25 -10 0 2 6 10 - 30 46 75 100 86 97 P . C. Lee et al. 422 stronger interaction. Therefore, a separation is developed between the hard and soft segments. In other words, this higher ratio of ordered chains of hard segments will exhibit more phase separation. Morphology of polyurethanes observed by T E M . To observe the hard segment distribution in the structure, the vapour-stained KuO, materials were examined by TEM. The results are shown in Fig. 1. Since RuO, is a good staining agent for benzene rings, the hard segments are observed in the black or grey region of the TEM, while the soft segments are in the white domain. Figure 1A shows the TEM picture of sample 110. Because the weight fraction of hard segments is less in this system, they are dispersed in the structure. The stained area is increased with increase in value of the hard segments. The 211, 413 and 615 polyurethanes possess hard segments in the dispersed phase, but the ( A ) 110 (B)2I1 (IT) 817 817 polyurethane has a continuous phase of hard segments. The 817 material possesses a calculated hard segment weight fraction of 0.73 and HBI of 2.93. Continuous phases of hard segments were also observed in the 1019 and 12111 composites which possess of HBIs 3.42 and 3.98, respectively. Mechanical properties of polyurethanes. The dynamic mechanical spectra of flexural storage moduli ( E ) of samples based on PTMO 1000 are shown in Fig. 2. It can be seen that increasing the MDI content of the polyurethane significantly increases E' through the rubbery plateau region. Furthermore, the onset of the rubbery plateau region is in a lower temperature region for the lower ratios of MDI. These trends indicate that the samples with lower MDI weight fractions have higher mobility to rearrange the hard-segment chains in the bulk at the physiological temperature of 37°C. This assumption is reasonable, because our cell culture system at 37°C is close to the annealing condition such that molecular chains can be easily rearranged as the MDI ratio is lowered. On the other hand, the polyurethanes with high hard segment ratios cannot move freely, especially samples 615, 817, 1019 and 12 111. At the cell culture incubation temperature, 37"C, only 110 and 211 have soft segments which move rather easily, resulting in a surface suitable for the attachment and proliferation of cells. DSC was used to investigate the melting temperature (T,) of crystallites in these polymers. The results are shown in Fig. 3. The endotherms show that AH and melting temperature increased as the hard segment content increased. The 615, 817, 1019 and 12 111 polyurethanes have an obvious T, at about 200°C. Also observed in Fig. 3 are the recrystallization exotherm temperatures (K), which are also related to the weight fraction of hard segments in the samples. In 211 polyurethane, the T, is 40°C which is near to the physiology temperature. The 615 and polyurethanes with higher 10 (C) 413 Amplitude (P-ol-O.&O mm (F) 1019 IIG ) I 2 1 I I (D)615 Fig. 1. TEM pictures of PU samples: (A) 110; (B) 211; (C) 413;(D) 615;(E) 817;(F)1019;(G)12 111. 7 ! -200 -100 0 100 Temperature YC) Fig. 2. Flexural storage moduli (E') versus temperature for various MDI ratios. POLYMER INTERNATIONAL VOL. 41, NO. 4, 1996 Influence of hard segments of PU 423 of higher mobility of the molecular chains in the 211 component. In vitro evaluation of cell growth on polyurethanes -=a& -0.5- I ' . 0 -1.0- g . -1.5- -z.s! -150 . , . . -lw , -60 . 0 sb 100 Temperature ('c) 160 260 z Fig. 3. DSC scans for various MDI ratios at a heating rate of 10°C min- '. hard segment ratio possess T, higher than 75°C. Since there is a molecular weight distribution of the polymer chains, the lower T, of materials with lower weight fractions of MDI is due to thermally induced recrystallization of hard segments for the shorter molecular chains existing in the soft segments. Surface tension of polyurethanes. The results of surface tension measurements of specimens immersed in octane for 24h at 37°C are shown in Fig. 4. Compared with the control specimens (hydrophilic and hydrophobic glasses), all the polyurethanes possess intermediate behaviour between hydrophilic and hydrophobic, implying the existence of both hydrophilic and hydrophobic molecules in the molecular structure. The 211 polyurethane possesses higher surface tension and contains more hydrophilic groups on the surface than other polyurethanes. It is suggested that this is a consequence Cytotoxicity of extract. Several methods have been reported for detecting the cytotoxicity of materials. In the present study, we used (1) extract dilution assay and (2) direct contact assay on cell cultures to evaluate the in uitro cytotoxicity and biocompatibility of materials. In addition, we used MTT assay to determine the cell viability in examining cell attachment and proliferation. It is a rapid and reproducible colorimetric method, which is based on the cleavage of a yellow tetrazolium salt from purple formazan crystals by mitochondria1 enzymes of metabolically active cells. As shown in Fig. 5, no oytotoxicity was detected in the extracts of the polyurethanes since the cell viability results were similar to the control value of the tissue culture well. Therefore, we can rule out extract toxicity in these materials and conclude that the properties of the material surface have a dominant influence on the experimental results. Cell attachment and proliferation. To compare cellular growth on these materials, the 3T3 fibroblast cell line was directly seeded on the PU series. In Fig. 6, the results show that all the materials can allow attachment (i.e. seeding after 2 h) of the 3T3 fibroblast. It can also be seen that the polyurethanes have less cell attachment than the tissue culture well and hydrophilic glass. Among the polyurethanes, the 211 and 413 have better surfaces supporting cell attachment. Furthermore, cell attachment decreases with increasing MDI ratio and 110 polyurethane, i.e. 12 111 PU has the least number of cell attachments. The results of cell proliferation of the PU series at 37°C after 72 h and 144h are shown in Fig. 40 0.40 I 35 6 - h 30 a 5 I 1 25 - E 1 0.25 o v) 5 2 20 g 15 -m I 10 rn e 0.20 -1-1000 3 > -2-1000 -4-1000 0.15 x 0 0.10 x t t 1 2 4 6 8 10 12 Hi Hb Molar ratio of MDI per PTMO 1000 Fig. 4. Surface tension of PTMO 1000 series PU, where Hi denotes a hydrophilic glass surface and Hb denotes a hydrophobic glass surface. POLYMER INTERNATIONAL VOL. 41, NO. 4, 1996 0.00 6-1000 -8-1000 - A - 10-1000 5 0 T.C. Well -CE c 5 0 CI 0.30 12-1000 ' 24h 48h 72h 96h 120h 144h Growth Time ( hours) Fig. 5. Cell proliferation in extracts of PTMO 1OOO series PU by MTT assay. P . C. Lee et al. 424 0.035 0.030 0.025 0 b m T 0.020 0, 4 0.015 0 0.010 > 0 0.005 0.000 2 1 a 6 4 H.G. T.C. well 12 10 Molar ratio of MDI per PTMO 1000 Fig. 6. Cell attachment of PTMO 1000 series PU after 2 h. 7. The molecular chains on the surface may achieve an equilibrium state during a longer-term incubation at 37°C. After 144 h proliferation, the cells almost achieve confluence on each of the PUS,except for 110 PU. With a low ratio of hard segments, 110 PU has the worst cell growth behaviour after 72 and 144h of cell proliferation, implying that the softest material may not be a good supporting matrix for cell growth owing to the lack of hard segments in the structure. It seems that suitable hard segments are a necessary element for cell growth. On the other hand, 211 PU has the best cell growth among this polyurethane series. The same trend in cell attachment and cell proliferation on the various PUS is that increasing the MDI ratio causes a decrease in cell numbers on the surface. Based on the results of TEM and cell proliferation for this polyurethane series, it appears that the cells can grow perfectly on the disperse phase of hard segments, but not on the continuous hard segment surface. However, the requirement 0.40 i OlMN 0.35 E 0.30 E 0 [;; c m 0.25 Q, 0.20 > 0.15 -m3 6 0 0.10 0.05 0.00 Glass T.C. C.E. I 2 4 6 8 10 12 well Molar ratio of MDI per PTMO 1000 Fig. 7. Cell proliferation of PTMO 1000 series PU after 12 and 144 h. for cell growth is not only a continuous soft segment domain, but also a ratio of hard/total segments, which is between 0.4 and 0.6. The 211 and 413 PUS fall into this range. The general concept of enhanced interaction with hydrophilic materials has been fostered by experiments, demonstrating that cells usually attach much more readily to glass than to hydrophobic surfaces such as Teflon,’ siliconized glass,” polystyrene” or parafilm.” Although the importance of substrate interfacial tension has often been cited in cell adhesion, the surface tension values we have measured are only minimally affected. The reason for this is that the measurement of surface tension is an averaged value of the macrophase on the surface, but cell attachment and proliferation is a microphase factor of the specimen surface. Hence, the microphase, observed by TEM, is a direct and significant factor. In view of the above results it is reasonable to expect that physical and chemical properties are equally important in yielding a suitable environment for cell growth. In this context, two factors are considered to be crucial in determining the cell attachment and cell proliferation properties of the surface. The first factor is the surface morphology, which has been demonstrated from TEM, and the existence of an optimally dispersed hard segment phase in the PUS. The second factor is the hydrophilic property, which in principle can be induced readily when the polymer chains have high mobility at 37°C. Hence the high weight fraction of hard segments has a lower cell growth than 211 and 413 PUS. CONCLUSION In this study, the bulk properties of a series of hard segment polyurethanes and cell interaction on these materials were studied by DSC, DMA and TEM. The results show that cell proliferation is affected by mechanical factors of the materials such as T,, T, and rigidity. A lower Tg for the lower weight fraction of hard segments gives a higher mobility to the soft segments in the molecular structure. With T, close to 37”C, the hard segments of 21 1 and 413 PUS can rearrange the molecular surface to provide a stable morphology for cell growth. The softest material, 110 PU, and the hardest material, 12 111 PU, are not suitable for cell attachment and proliferation. The results also indicate the absence of a cytotoxicity response for extracts from the polyurethanes studied. Furthermore, there was not a direct relation between surface tension and cell proliferation, since surface tension is measured for the macrostructure of the surface, while cell growth depends on the microphase structure of the surface, which can only be detected by TEM. The domain of hard-soft segment distribution is found to affect cell attachment or proliferation. The cells tended to grow on the optimally disPOLYMER INTERNATIONAL VOL. 41, NO. 4, 1996 InJIuence of hard segments of PU persed hard segments of PUS, especially on the surface with weight fraction of hard segments between 0.4 and 0.6. The corresponding HBI is between 1.5 and 2.1. These PUS may possess suitable chemical structures and physical properties for cell growth. The phaseinverse region is at the hard segment ratio 0.7 (817 PU); however, a continuous phase of hard segments is not suitable for cell attachment and proliferation. REFERENCES 1 Tobushi, H., Hayasi, S. & Kojima, S., JSME J., 35 (1992) 296. 2 Bamford, C. H. & Wayne, R. P.;Polymer, 10 (1969) 661. 3 Droscher, M. & Wegner, G., Polymer, 19 (1978) 42. 4 Lm,C. C. & Baliga, S., J. Appl. Polym. Sci., 31 (1986) 2483. 5 Phillips, R. E., Amith, M. C. & Thoma, R. J., J. Biomater. Appl., 3 (1988) 207. POLYMER INTERNATIONAL VOL. 41, NO. 4, 1996 425 6 Thoma, R. J., Tan, F. R. & Phillips, R. E., J. Biomater. Appl., 3 (1988) 180. I Leonard, P., J. Biomater. Sci. Polym. Edn, 6 (1994) 225. 8 Pangman, W. J., US patent 2,842,775 (1958). 9 Pangman, W. J., US patent 3,559,214 (1971). 10 Boretos, J. W. & Pierce, W. S., Science, 158 (1967) 1481. 11 Stokes, K. & Cobian, K., Biomaterials, 3 (1982) 225. 12 Hergenrother, R. W., Wabers, H. D. & Cooper, S . L., Biomuterials, 14 (1993) 449. 13 Stokes, K., J. Biomuter. Appl., 3 (1988) 248. 14 Szycher, M., J. Biomuter. Appl., 3 (1988) 383. 15 Christ, R. E., Buchen, S. Y., Fend, D. A., Knight, P., Solomon, K. D. & Apple, D. J., J. Biomed. Muter. Res., 26 (1992) 607. 16 Blais, P., J. Appl. Biomater, l(l990) 197. 17 Ciapette, G. & Genni, E., Biomaterials, 15(1994) 63. 18 Mosmann, T., J . Immunol. Method, 65 (1983) 55. 19 Taylor, A. C., Exp. Cell Res. Suppl., 8 (1961) 154. 20 Weiss, L. & Blumenson, L. E., J. Cell Physiol., 70 (1967) 23. 21 Grinnel, F., Milam, M. & Srere, P. A., Aarch. Biochern., 153(1972) 193. 22 Harris, A., Exp. Cell Res., 77 (1973) 285.