Biotechnol. Prog. 1995, 11, 677-681
677
Growth Enhancement of Anchorage-Dependent and Anchorage-IndependentCells by Coimmobilization of Insulin with Poly(ally1amine)or Gelatin Ji Zheng, Yoshihiro Ito,* and Yukio Imanishi Division of Material Chemistry, Faculty of Engineering, Kyoto University, Kyoto 606-01, Japan
An anchorage-dependent cell, mouse fibroblast STO cell and an anchorage-independent cell, K562 cell, were cultured on a polymer membrane coimmobilized with insulin and adhesion factors, polfiallylamine), and gelatin. The former is cationic and electrostatically (non-biospecifically) enhances cell adhesion, and the latter is a hydrolyzate of collagen and biospecifically enhances cell adhesion. They were immobilized onto a surface-hydrolyzed poly(methy1 methacrylate) membrane by water-soluble carbodiimide. The adhesion of both STO and K562 cells was accelerated by the immobilization of poly(ally1amine) or gelatin. The insulin immobilization did not affect adhesion of either cell. Although the growth of STO cells was enhanced on the insulin-immobilized membrane, the growth of a 6 2 cells was not. However, the coimmobilization of insulin and a cell-adhesion factor accelerated growth of both cells. It was considered that an increased frequency of interaction between the immobilized insulin and the receptor led t o the cell growth acceleration.
Introduction A serum-free cell culture is one of the most important issues in biotechnology for mass production of bioactive substances. The serum-free cell culture was realized by the addition of serum extracts, which were considered to be growth factors (Barnes and Sato, 1979). On the other hand, the importance of biological and synthetic substrate for cell culture has been recognized (Hynes, 1992; Morla et al., 1994; Singhvi et al., 1994; Wong et al., 1994). Recently, we found that these growth factors were more active in the immobilized state than in the free (soluble) state and that the immobilized growth factors could be used repeatedly (Ito et al., l991,1992a,b, 1993,1994; Liu, et al., 1992a,b,1993a,b,c). However, the enhancement was not observed in the culture of anchorage-independent cells (Ito et al., 1992b). In the present study, poly(ally1amine) (PAA), which is a cationic polymer, or gelatin (GN), which is a hydrolyzate of collagen, was coimmobilized with insulin, which is a growth factor, in order to enhance frequency of the interaction of immobilized insulin with cells due to nonspecific or specific adhesion.
Materials and Methods Materials. Poly(ally1amine)hydrochloride (PAA, MW 60000) was a gift from Nittobo Co. (Tokyo, Japan). Insulin (bovine origin, No. 1-5500) and gelatin (porcine origin, No. G2500) were purchased from Sigma Co. (St. Louis). l-Ethyl-3-(3-(dimethylamino)propyl)carbodiimide hydrochloride, which is a water-soluble carbodiimide (WSC), ninhydrin, and dinitrofluorobenzene were purchased from Nacalai Tesque, Inc. (Kyoto,Japan). Rhodamine 6G and methyl methacrylate were purchased from Wako Pure Chem. Ind. Ltd. (Osaka, Japan). The Dulbecco's Modified Eagle medium (DMEM) and the RPMI 1640 medium were purchased from Nissui Pharm. Co. Ltd. (Tokyo,Japan). [125111nsulin(specific activity, 3480 kBq/pg) was purchased from DuPont Co. (Wilmington, DE).
Preparation of Polymer Membrane. Poly(methy1 methacrylate) (PMMA) membranes were prepared as follows (Ito et al., 1992a; Liu et al., 1992b). The commercial methyl methacrylate was washed and distilled, and a toluene solution of the monomer (2 : 3 v/v) was subjected to polymerization initiated with 4,4'-azobis(isobutyronitrile) at 50 "C for 12 h. The PMMA obtained was purified by reprecipitations. A 10 wt % toluene solution of PMMA was eluted on a circular glass plate having a diameter of 15 mm and illuminated with an infrared lamp to remove the solvent by evaporation, obtaining a circular PMMA membrane having a diameter of 15 mm. The PMMA membrane was placed in a 4 N aqueous NaOH solution for 90 min at 50 "C for the surface hydrolysis. The membrane was then immersed in a 10 wt % aqueous citric acid a t room temperature overnight for neutralization. The membrane was washed with double-distilled water until the washing liquid reached pH 7.0 and stored in double-distilled water. Immobilization of Proteins. PAA or GN was immobilized onto the surface-hydrolyzedPMMA membrane by WSC. The surface-hydrolyzed polymer membrane was immersed in 2-(N-morpholino)ethanesulfonate-buffered solution (pH 4.5) containing WSC (1 mg/mL) and PAA (500 pg/mL) or GN (500 pg/mL) at 4 "C for 48 h. ARer the reaction, the PMMA membrane was repeatedly washed with phosphate-buffered saline (PBS) until the absence of PAA or GN in the washing liquid was confirmed by dinitrofluorobenzene(Dubin, 1960)or ninhydrin (Imanishi et al., 1988)as indicators, respectively. Then, insulin was immobilized onto the PAA- or GNimmobilized PMMA membrane (PAA-PMMA or GNPMMA) by WSC. The PAA-PMMA or GN-PMMA membrane was immersed in 2-(N-morpholino)ethanesulfonatebuffered solution (pH 4.5) containingWSC (1mg/mL) and insulin of different concentrations at 4 "C for 48 h. After immobilization of insulin, the membrane was repeatedly washed with PBS. The absence of releasing of insulin in the washing liquid was confirmed by [12511insulin,
8756-7938/95/3011-0677$09.00/0 0 1995 American Chemical Society and American Institute of Chemical Engineers
Biotechnol. Prog., 1995, Vol. 11, No. 6
678
which was immobilized under the same conditions as nonradioactive insulin. Determination of the Amount of Carboxyl Group, PAA, GN, and Insulin on the Surface of PMMA Membrane. The amount of carboxyl groups on the PMMA membrane surface was determined by using Rhodamine 6G as reported previously (Palit and Ghosh, 1962). The indicator was added to a benzene solution in which a polymer membrane was dissolved, and the absorption intensity at 515 nm was measured. The surface concentration of the amino group produced by the immobilization of PAA was determined by using dinitrofluorobenzeneas follows (Dubin, 1960; Liu et al., 1993~).PAA-PMMA was immersed into 0.066 M boratebuffered solution (pH 9.7) containing 0.12 vol % dinitrofluorobenzene at 65 "C for 30 min. M e r being washed with methanol to remove the adsorbed reagents, the membrane was dissolved in dioxane. The W absorbance of the solution at 350 nm was measured. The calibration was performed by using the dioxane solution of PAA of known concentrations. The amount of immobilized GN on PMMA membrane was determined using ninhydrin after hydrolysis as follows (Imanishi et al., 1988). A piece of the film was immersed in a 5 N HCl(1 mL) and autoclaved at 120 "C for 1 h. The resulting acidic solution was neutralized with a 5 N aqueous NaOH solution. A potassium acetatebuffered solution (5 mL) containing ninhydrin (0.4 g), hydrindantin (0.06 g), and methyl cellosolve (15 mL)was then added to the hydrolyzate solution, and the mixture was incubated in an autoclave at 100 "C for 3 min. After cooling, the absorbance a t 570 nm was measured. The calibration curve was obtained using GN of known concentrations. The amount of immobilized insulin was determined by using lZ5I-labeledinsulin as previously reported (Ito et al., 1991). Cell Adhesion and Growth. Mouse fibroblast cell STO (Ito et al., 1991) and human chronic myelogenous leukemia cell K562 (Carman et al., 1975;Sakakura, 1994) were subcultured in DMEM and RPMI 1640 medium wt % containing 10 vol % fetal calf serum, 5.0 x streptomycin, and 3 x w t % penicillin. The membranes were disinfected with 70 vol % ethanol, washed with disinfected phosphate-buffered saline (PBS, pH 7.41, and placed on a 24-well multiplate made of polystyrene. The STO cells were detached from the culture flask by the treatment with Ca2+-freeand M$+-free PBS [(-)PBSl containing 0.15 wt % trypsin (2000 unit/g) and 0.02 wt % ethylenediaminetetraaceticacid (EDTA),washed twice with (-)PBS, and then resuspended in a serum-free DMEM. The K562 cells were centrifuged, rinsed with (-)PBS, and then resuspended in fresh serum-free PRMI 1640 medium. Cell suspensions of two different concentrations(2 x lo4 and 2 x lo5 cells/mL)were prepared for cell-growth and cell-adhesion experiments, respectively. The cell suspension (1 mL) was added to each well containing a sample membrane and incubated in an atmosphere containing 5% COz at 37 "C for 48 and 2 h for cell-growth and cell-adhesion experiments, respectively. The number of cells was determined by measuring the amount of intracellular DNA as reported previously (Zheng et al., 1994). The relative cell adhesion is represented by the percentage of adhered cells against the initial number of cells in the suspension. The relative cell growth rate was obtained by dividing the cell number on the sample membrane by that on the hydrolyzed PMMA membrane after 48-h culture.
I
t
2N
n
se4
0.4 I
t
. I I
0
E
Eo
0.3
Ea z0 3. 0.2 Y -
0 . 0 b '
I
0
100
"
200
"
*
300
400
'
I
500
" 600
Insulin concentration in feed ( pg/ml) Figure 1. Immobilization of insulin on
(0)PMMA, (0) GNPMMA, and (O),PAA-PMMA membranes. Bars represent the standard deviation. n = 6.
Results Immobilization of insulin onto PAA-PMMA and GN-PMMA Membranes. Insulin was immobilized on three types of membranes, which were PMMA containing moUcm2,PAA (39.6pg/cm2)carboxyl groups of 1.8 x PMMA, and GN (12 pg/cm2)-PMMA. The amount of insulin immobilized on the surface increased with increasing the feed concentrations, as shown in Figure 1. No significant difference in the amount of immobilized insulin on these membranes was observed. Cell Adhesion on Membranes. Figure 2 shows the adhesion of STO and K562 cells onto PMMA, PAAPMMA and GN-PMMA membranes on which insulin was immobilized. By comparing parts a and b of Figure 2, it is found that the adhesion of STO cells was higher than that of K562 cells. Under the conditions without insulin immobilization, adhesion of the cells was accelerated by the immobilization of PAA or GN and the effect of PAA was slightly stronger than that of GN. On the other hand, the immobilization of insulin did not affect the adhesion of either cell. Although the insulin immobilized onto the PAA-PMMA membrane did not affect the adhesion of either cell, the insulin immobilized onto the GN-PMMA membrane reduced the adhesion of both kinds of cells. Cell Growth on Membranes. Figure 3 shows the relative cell growth of STO and K562 cells on PMMA, PAA-PMMA, and GN-PMMA membranes in the presence of free insulin. With increasing concentration of free insulin, the growth rates of both kinds of cells were accelerated. In the absence of insulin, the growth rate of STO cells on the PAA-PMMA membrane was about three times as large as that on the PMMA membrane. The growth acceleration of K562 cells by insulin addition was not so large on any of the membranes. On the other hand, the growth of STO cells markedly depended on the nature of the membrane. For STO cells, the growth rate on the PAA-PMMA membrane was higer than that on the GN-PMMA membrane and, without the addition of insulin, comparable to that on the PMMA membrane in the presence of 20 pg/mL of insulin. Figure 4 shows the relative growth rate of cells on the PMMA, PAA-PMMA, and GN-PMMA membranes on which insulin was also immobilized. The growth rate of STO cells was accelerated by insulin immobilization (InsPMMA membrane), while that of K562 cells was not accelerated. However, when PAA or GN is coimmobilized (Ins-PAA-PMMAand Ins-GN-PMMA),the growth rates
Biotechnol. Pfog., 1995,Vol. 11, No. 6
a
679
I
I
I "
u.0
0.1
0.2
0.4
0.3
0
0
Amount o f immobilized insulin ( pg/cm 45
1
)
1
10
20
Concentration o f free insulin ( pg/ml ) 1.6
I
0.9
u.0
0
~~
1
0.1
0.2
0.3
0.4
Amount o f immobilized insulin ( Ir g/cm 2 , Figure 2. Adhesion of STO (a) and K562 (b) cells onto (0) PMMA, (0) GN-PMMA, and ( 0 )PAA-PMMA membranes. Bars represent the standard deviation. n = 8.
of both kinds of cells were accelerated with increasing the amount of coimmobilized insulin. The effect of PAA or GN was stronger with the STO cells than with the K562 cells. The growth of STO cells depended on the nature of the membrane. On the Ins-PAA-PMMA membrane the growth rate of STO cells was 3.5 times as large as that on the PMMA membrane and comparable to that in serum-containing media. Figure 5 summarizes the results of Figures 3 and 4 by plotting the amount of insulin in the wells as abscissa. The same effects on ceIl growth acceleration were observed by smaller amounts (1/10to l/lOO) of insulin in the immobilized state than in the free (soluble) state, except for K562 on insulin-immobilizedPMMA.
Discussion Insulin Immobilization. A novel method for cell culture in protein-free media has been developed by using immobilized growth-factor proteins as described in previous papers (Ito et al., l991,1992a,b,1993,1994).It was found that the immobilized insulin enhanced the growth of anchorage-dependent cells, although the immobilization of other proteins such as albumin, y-globulin, fibrinogen, and collagen was ineffective. Anti-insulin inhibited cell growth on the insulin-immobilizedmaterial because the antibody covered the immobilized insulin. Recently, Reddy et al. (1994)found by using a cell containing internalization-deficientmutant receptor that the biological signal of epidermal growth factor (EGF) was transmitted without internalization. In addition it was reported that the extracellular matrices induced
10
20
Concentration o f free insulin ( p g/ml) Figure 3. Growth of STO (a)and K562 (b) cells on (0)PMMA, (0)GN-PMMA, and (0)PAA-PMMA membranes in the presence of free insulin. Bars represent the standard deviation. n = 8.
phosphorylation of integrin or the neighbor proteins without internalization (Hynes, 1992). These results demonstrate that the biological signals are transmitted without internalization of ligandheceptor complexes. In the present study the growth rate of STO cells increased with the increase of the amount of immobilized insulin, although the immobilized insulin did not affect cell adhesion. On the other hand, neither adhesion nor growth of K562 cells was enhanced on the insulinimmobilized PMMA membrane. The same has been reported with a mouse hybridoma cell, which is anchorage-independent. The growth rate of K562 cells in the presence of immobilized insulin was less than that in the presence of free insulin. These results indicate that the signal transmission is triggered by adhesion of cells onto extracellular matrix and that the growth acceleration is not due to released insulin. In fact, the very small amount of insulin in the immobilized state accelerated cell growth. The reasons why the immobilized insulin markedly accelerated cell growth are considered to be as follows. (i)The immobilized insulin provided a high local concentration of insulin, which should induce multivalent simultaneous stimulation, leading to enhanced complex formation with receptors and also to promoted crosslinking of the complex. (ii) The immobilized insulin inhibited down-regulation of cells by internalization to be decomposed and persistently activated cells. (iii) Unknown nonbiospecific interactions must be taken into consideration. For example, the physicochemical and morphological natures of immobilized surfaces and the influences on cell membrane fluidity, etc., should be important.
680
4.0
0) Y
a
P)
Y
&
E
9
1
Biofechnol. Prog., 1995, Vol. 11, No. 6
(a)
Immobilized Free
h
5
t
3
L
EM
M Q)
-2 i+
0)
>
. Y I
cp
. I
Y (P
I
B
0
0.0
0.1
0.2
0.3
0.4
Amount of immobilized insulin ( pg/cm 1.5
.1
2,
I
1.6
3
Q) Y
1
m
1.5
I
fbl Immobilized
&
E
9
9
3
E
j:&
1.4 1.3
M
M
* Q)
.I
Y
. I
Y
ep
2
Ez
a I
10
100
Amount of insulin ( pg/well)
1.1 le2[
1 .o
1
d
Free
V."
u.0
0.1
0.2
0.3
0.4
Amount of immobilized insulin ( pg/cm
2>
Figure 4. Growth of STO (a)and K562 (b) cells on (0)PMMA, (0)GN-PMMA, and (0)P a - P M M A coimmobilized with insulin. Bars represent the standard deviation. n = 8.
Coimmobilization of Insulin with Adhesion Factors. We have reported that the coimmobilization of insulin with adhesion factors including fibronectin (It0 et al., 1993), collagen (It0 et al., 1993, 19951, and poly(allylamine) (Zheng et al., 1994) is effective for growth acceleration of anchorage-dependent cells. The present study compared the effects of PAA and GN. The former is a cationic macromolecule and is believed to electrostatically enhance cell adhesion (Chaturvedi and Picciano, 1989). GN is a hydrolysate of collagen and biospecifically enhances cell adhesion. When insulin was immobilized on a GN-PMMA membrane, the cell adhesion decreased with the increase of the amount of immobilized insulin. However, this was not the case in the insulin immobilizationonto the PAAPMMA membrane. The deteriorating effect of insulin immobilization onto the GN-PMMA membrane may be explained by a coverage of the active site of GN with immobilized insulin. Since PAA has no specific active site, the insulin immobilization should not affect cell adhesion so much. Coimmobilizationof a cell-adhesionfador (PAA or GN) with insulin accelerated the growth of anchorageindependent K562 cells as well as anchorage-dependent STO cells. The growth of STO cells markedly depended on the nature of the membrane, while the growth of K562 cells was insensitive to the nature of the adhesion factor (PAA or GN). Two different explanations have been proposed for the synergistic effect of different biosignals (It0 et al., 1995). One of them is an increased frequency of interaction of immobilized insulin with its receptor protein in the cell
0.9
.1
1
10
100
Amount of insulin ( pg/well) Figure 6. Growth of STO (a)and K562 (b) cells on (A)PMMA, (W GN-PMMA, and ( 0 )PAA-PMMA in the presence of insulin and on (A) PMMA, (0)GN-PMMA, and (0)PAA-PMMA coimmobilized with insulin. Bars represent the standard deviation. n = 8.
membrane, stemming from the enhanced adhesion. The other is based on a postulated interaction between the activated insulin receptor and the activated receptor (integrin) for the extracellular matrix. Taking into consideration the experimental observation that the nonspecific adhesion factor PAA was more effective than the specific adhesion factor GN, the former explanation might be dominant. However, recently Vuori and Ruoslahti (1994) found the evidence demonstrating association between integrin and insulin receptor substrate-1 for the synergistic action of growth factor and extracellular matrix receptors. Therefore the latter explanation cannot be denied. Nature of the Cell Line. As mentioned above, the potentiality of immobilized insulin depended on the nature of the cell line. Anchorage-independentcells did not grow on the insulin-immobilized matrix because the cells do not interact with the immobilized insulin. However, the growth of anchorage-independent K562 cells was accelerated by coimmobilization of a celladhesion factor, PAA or GN, with insulin. This result should be attributed to an increased frequency of interaction between the immobilized insulin and its receptor. It was demonstrated that any kind of cell line could be cultured by controlling the nature of the membrane. For the design of a bioreador for cell culture to produce biologically active substances, attention should be paid to the nature of the cell line and extracellular matrix. In this respect, the design of the cell line by transformation (e.g., from an anchorage-independent cell to an anchor-
Biotechnol. Prog., 1995,Vol. 1 1 , No. 6
681
age-dependent cell) will be interesting to use the merit of an extracellular matrix.
Acknowledgment The authors thank Dr. M. Saito of Radioisotopic Center of Kyoto University for the kind help and advice on the radioisotope experiments. The authors also thank Mr. K. Kawamura of Akita Sumitomo Bake Co. for providing K 562 cells.
Literature Cited Barnes, D.; Sato, G. Growth of a Human Mammary Tumour Cell Line in a Serum-Free Medium. Nature 1979,281,388389. Carman, B. L.; Bismarck B. L. Human Chronic Myelogenous Leukemin Cell-Line with Positive Philadelphia Chromsome, Blood 1975,45, 321-334. Chaturvedi, N.; Picciano, P. L. Tissue Immobilization and Cell Culturing System and Method for Affixing Biologically Active Moieties to a Substrate. European Patent Application .. 89111726.9, 1989. Dubin, D. T. The Assay and Characterization of Amines by Means of 2.4-Dinitrofluorobenzene. J. Biol. Chem. 1960.235. 783-786. Hynes, 0. Integrins: Variety, Versatility, and Interactions in Cell Adhesion. Cell 1992, 69, 11-25. Imanishi, Y.; Ito, Y.; Liu, S. Q.; Kajihara, M. Design and Synthesis of Biocompatible Polymeric Materials. J. Macromol. Sci., Chem. 1988, A.25, 555-570. Ito, Y.; Liu, S. Q.; Imanishi, Y. Enhancement of Cell Growth on Growth-Factor-Immobilized Polymer Film. Bwmateriuls 1991, 12,449-453. Ito, Y.; Liu, S. Q.; Nakabayashi, M.; Imanishi, Y. Cell Growth on Immobilized Cell-Growth factor. 2. Adhesion and Growth of Fibroblast Cells on Poly(methy1 methacrylate) Membrane Immobilized with Proteins of Various Kinds. Biomaterials 1992a, 13, 789-794. Ito, Y.; Uno, T.; Liu, S. Q.; Imanishi, Y. Cell Growth on Immobilized Cell-Growth Factor. 8. Protein-Free Cell Culture on Insulin-Immobilized Microcarrier. Biotechnol. Bioeng. 1992b, 40,1271-1276. Ito, Y.; Inoue, M.; Liu, S. Q.; Imanishi, Y. Cell Growth on Immobilized Cell-Growth Factor. 6. Enhanced Cell Growth on Cell Growth Factor and Adhesion Factor Co-immobilized Materials J. Biomed. Mater. Res. 1993,27, 901-907. Ito, Y.; Liu, S. Q.; Orihara, T.; Imanishi, Y. Cell Growth on Immobilized Cell-Growth Factor. Interactions of Fibroblast Cells with Insulin Immobilized on Polymer Membranes of Different Wettabilities, J . Bioactive Compatible Polym. 1994, 9, 170-183. Ito, Y.; Zheng, J.; Imanishi, Y. Serum-Free Cell Culture on Insulin-Immobilized Porous Collagen Beads. Biotechnol. Bioeng. 1995, 45, 144-148. I
,
Liu, S. Q.; Ito, Y.; Imanishi, Y. Enhanced Cell Growth on Immobilized Cell-Growth Fador. 1.Acceleration of the Growth of Fibroblast Cells on Insulin-Immobilized Polymer Matrix in Culture Medium Without Serum. Biomaterials 1992a, 13, 50-58. Liu, S. Q.; Ito, Y.; Imanishi, Y. Cell Growth on Immobilized Cell Growth Factor. 4. Interaction of Fibroblast Cells with Insulin Immobilized on Poly(methy1 methacrylate) Membrane. J . Biochem. Biophys. Methods 1992b, 25, 139-148. Liu, S. Q.; Ito, Y.; Imanishi, Y. Cell Growth on Immobilized Cell Growth Factor. 5. Interaction of Immobilized Transferrin with Fibroblast Cells. Znt. J . Biol. Macromol. 1993a, 15,221-226. Liu, S. Q.; Ito, Y.; Imanishi, Y. Cell Growth on Immobilized Cell Growth Factor. 7. Protein-Free Cell Culture by Using Growth Factor-Immobilized Polymer Membrane. Enzyme Microb. Technol. 1993b,15, 167-172. Liu, S. Q.; Ito, Y.; Imanishi, Y. Cell Growth on Immobilized Cell Growth Factor. 9. Covalent Immobilization of Cell Growth Bovine Endothelial Cells. J . Biomed. Mater. Res. 1993c, 27, 909-915. Morla, A.; Zhang, Z.; Ruoslahti, E. Superfibronectin is a Functionally Distinct Form of Fibronectin. Nature 1994,367, 193-194. Palit, S. R.; Ghosh, P. Quantitative Determination of Carboxyl Endgroups in Vinyl Polymers by the Dye-Interaction Method. J . Polym. Sci. 1962, 58, 1225-1232. Reddy, C. C.; Wells, A.; Lauffenburger, D. A. Proliferative Response of Fibroblasts Expressing Internalization-Deficient Epidermal Growth Factor (EGF) Receptors Is Altered via Differential EGF Depletion Effect. Biotechnol. Prog. 1994,10, 377-384. Sakakura, T. Extracellular Matrices; Yohdosha Publ. Co.: Tokyo, 1994; pp 87. Singhvi, R.; Kumar, A.; Lopetz, G. P.; Stephanopoulos, G. N.; Wang, D. I. C.; Whitesides, G. M.; Ingber, D. E. Engineering Cell Shape and Function. Science 1994,264, 696-698. Vuori, K.; Ruoslahti, E. Association of Insulin Receptor Substrate-1 with Integrins, Science 1994, 266, 1576-1578. Wong, J. Y.; Langer, R.; Ingber, D. E. Electrically Conducting Polymers Can Noninvasively Control the Shape and Growth of Mammalian Cells. Proc. Natl. Acad. Sci. U.S.A. 1994,91, 3201-3204. Zheng, J.; Ito, Y.; Imanishi, Y. Cell Growth on Immobilized CellGrowth Factor, Insulin and Polyallylamine Co-Immobilized Materials. Biomaterials 1994, 15, 963-968. Accepted April 14, 1995.@
BP9500201 @Abstractpublished in Advance ACS Abstracts, June 1, 1995.
