microheterophase structure, permeability, and biocompatibility of a-b-a triblock copolymer membranes...

Microheterophase structure, permeability, and biocompatibility of A-B-A triblock copolymer membranes composed of poly( y-ethyl L-glutamate) as the A component and polybutadiene as the B component

H. Sato,* A. Nakajima, and T. Hayashi Research Center for Medical Polymers 6 Biomaterials, Kyoto University, Kawaharacho, Shogoin, Kyoto 606, Japan

G . -W. Chen lnstitute of Chemistry, Academia Sinica, Beijing, China

Y. Noishiki lnstitute for Thermal Spring Research, Okayama University, Misasa, Tottori 682-02, Japan

A-8-A block copolymers (EBE) consisting of poly(y-ethyl L-glutamate) as the A component and polybutadiene as the B component were synthesized. The microheterophase structure of these block copolymers, as observed microscopically, was less distinct than that of the block copolymer (GBG) composed of poly( y-benzyl L-glutamate) as the A component. This was attributed to the better compatibility of the ethyl group with polybutadiene block, compared with the benzyl group. The observed micelle size of various microheterophase structures ranged from 350 to 480 A, which agreed well with the micelle size estimated from a thermodynamic

study. From the measurement of water permeation through the EBE block copolymer membranes, the contributions of the interfacial region between the a-helical polypeptide domain and polybutadiene domain to the permeability coefficient were concluded to be higher than those of poly(y-methyl L-glutamate)-polybutadiene block copolymer (MBM) membranes, i.e., more hydrophilic domain seemed to exist in EBE membranes than in MBM membranes. The results of in vivo tests on the tissue compatibility indicate that the EBE block copolymer membranes have good biocompa tibility.

INTRODUCTION

In our previous papers,'-3 the microheterophase structure and properties of A-B-A block copolymers (GBG) consisting of poly( y-benzyl L-glutamate) as the A component and polybutadiene as the B component in a helicogenic solvent, chloroform, were discussed as a function of the polybutadiene content. The GBG membranes were found to have high water permeability

*To whom correspondence should be addressed.

Journal of Biomedical Materials Research, Vol. 19, 1135-1155 (1985) 0 1985 John Wiley & Sons, Inc. CCC 0021-9304/85/091135-21$04.00

1136 SAT0 ET AL.

despite their low degree of swelling in water. The mechanism of water permeation can be partially elucidated by examining block copolymer membranes composed of various kinds of ester groups in the polypeptide portion and of various contents of polybutadiene.

In another paper14 the A-B-A block copolymers (EBE) consisting of poly(y-ethyl L-glutamate) as the A component and polybutadiene as the B component were synthesized in a similar manner as the GBG block copolymer series, and their molecular characterization was performed. In this work, we will discuss the water permeability and the biocompatibility of EBE membranes in comparison with block copolymers MBM5r6 composed of poly(y-methyl L-glutamate) as the A component.

The foreign body reaction of GBG and MBM block copolymers was observed to be almost the same degree as that of their homopolypeptides, i.e., poly( y-benzyl L-glutamate) and poly(y-methyl L-glutamate), which is remarkably high. This result seemed to be related to polybutadiene and polypeptide contents. The microheterophase structure of block copolymers was not found to result always in good biocompatibility. The ethyl ester of poly(L-glutamic acid) is expected not to be harmful in vivo even on biodegradation' and, therefore, the EBE block copolymers may be particu- larly useful for biomedical application.

EXPERIMENTAL

Materials

A homopolypeptides, poly( y-ethyl L-glutamate), abbreviated PELG, and A-B-A type block copolymers EBE composed of PELG as the A component and polybutadiene as the B component were synthesized4 according to the method reported in the previous paper.' The N-carboxy anhydride (NCA) of y-ethyl L-glutamate (7-ELG) was prepared from L-glutamic acid as a starting material. A cycloaliphatic secondary amine-terminated polybutadiene (ATPB), mol wt 3600, was used as the middle block for all the EBE block copolymers reported here.

The EBE block copolymers were prepared by the polymerization of various amounts of NCA of yELG onto ATPB in a mixture of dioxane and methylene dichloride (1 : 2 v/v) at 20°C for 72 h. The molar percentage of y-ELG in block copolymers was estimated from the results of elemental analysis carried out in the Organic Microanalysis Center in Kyoto University, and of infrared (IR) spectra.

Infrared, x-ray diffraction, and electron microscope measurements

The IR spectra of block copolymer films were measured with a Shimadzu Model-30A infrared spectrophotometer. The compositions of block copolymers were calculated from the relative intensity at 970 cm-' to that at

A-B-A TRIBLOCK COPOLYMER MEMBRANES 1137

1120 cm-l. Wide-angle x-ray diffraction of polymer films was measured with a Rigaku Geigerflex type DF-3 x-ray generator, equipped with an automatic diffractometer, where Ni-filtered Cu-Ka radiation was set to radiate onto a flat surface of the film parallel to the reflecting surface, Thin films of block copolymers formed on copper mesh were stained with osmium tetroxide. The domain structures of A-B-A triblock copolymers were observed by a Hitachi High Resolution Transmission Electron Microscope H-500.

Permeability measurements

The membrane permeability was cast from a 2% solution of the polymer dissolved in a 4: 1 mixture of chloroform (CF) and trifluoroethanol (TFE) onto glass plate, then dried at room temperature. The membrane thickness ranged from 20 to 30 pm. The degree of swelling Qw of the membrane was expressed as the weight fraction of water content in wet membrane. That is, Qw = (Wwet - Wdry)/Wwet, where W,,, and w d r y are the weights of wet and

dry membranes, respectively. The weight of wet membrane was measured by putting it in a weighing bottle with a microbalance manufactured by Chyo Balance Co., Kyoto, after blotting lightly excess water on the surface of water swollen membrane at equilibrium with filter paper.

The hydraulic permeability of 12.57 cm2 membranes placed on porous supporters was measured with a low-pressure ultrafiltration apparatus (Bio- Engineering Co., Model MC 11) at temperatures from 25°C to 50°C and pressures from 1 to 4 atm. The flux of effluent was determined by weighing the effluent.

Test of biocompatibility

Polymer samples coated on polyester fiber mesh cloth were implanted subcutaneously in mongrel dogs for 4 weeks, and then tissue compatibility was observed by the same method6 as that for polypeptide block copolymer membranes containing poly( y-benzyl L-glutamate), poly( y-methyl L-glutamate), or poly( e-N-carbobenzyloxy L-lysine).

The polyester mesh 1 X 3 cm in size was soaked in ethyl alcohol for 24 h to remove the lubricant and plasticizer, then washed in pure water, and dried. Such polyester mesh served as the supporter of polymer film of ca. 20 pm in thickness, as a marker of polymer samples implanted in the tissue, and as a reference of polymers for tissue reaction. The 1.6% polymer solutions of EBE in a mixture of chloroform and trifluoroethanol(4: 1 v/v) were cast on the polyester mesh and air dried.

After being washed in pure water, polymer samples coated on the polyester mesh were sterilized by immersion in 70% ethyl alcohol aqueous solution for l week, then soaked in a physiologic salt solution before implantation. The samples, implanted in the muscle tissue on the back of dogs for 4 weeks, were taken out together with the tissue surrounding the

1138 S A T 0 ET AL.

polymer samples. The specimens obtained after fixation in 10% formalin were incubated in paraffin. The sliced specimens, stained with hematoxylin and eosin, were examined microscopically.

RESULTS AND DISCUSSION

Wide angle x-ray diffraction

The orientation of a-helical polypeptide molecules in unstretched PELG and EBE films, cast from solutions in a mixture of chloroform and trifluoroethanol (2: 1 v/v), was investigated with x-ray diffraction patterns on the (1 1 0) plane. The composition of the polypeptides, in mole %, in those polymers is listed in Table I, together with PA and XH, where PA and XH mean the degree of polymerization of the A component and the helical contentr4 respectively.

Figure 1 shows the wide-angle x-ray diffraction (WAXD) patterns for PELG and EBE block copolymer films. The diffraction peaks at 28 of ca. 7.5" become broader on increasing the polybutadiene content in the block copolymer, indicating less crystallinity and a more random orientation of the a-helix in EBE films with increasing polybutadiene content.

The main diffraction peak of PELG at 28 of 7.25" corresponds to an intermolecular spacing 12.19 A of the PELG chains in homopolypeptide film, while the 28 of the main diffraction peak and the intermolecular spacing for EBE films are 7.75" and 11.41 A, respectively. The phenomenon that the existence of polybutadiene in block copolymers results in shorter intermolecular spacing of a-helix than that in homopolypeptide film was observed with the MBM' and EBE copolymer, but not with GBG copolymer.* The intermolecular spacing of the a-helices in polypeptide films is influenced by the casting solvent in the film preparation',' because the solvent affects the packing of side chains in polypeptide molecules. The intermolecular spacing is less by 0.78 A than in the PELG homopolymer film, perhaps due to better compatibility of polybutadiene with aliphatic part of side chains such as ethyl and methy1 residues. On the assumption that PELG adopts hexagonal packing in a manner similar to the hexagonal unit cell of PMLG,' the packing

TABLE I Molecular Parameters of PELG and EBE

~~

Mole fraction of Degree of poly- Helical polypeptide (A) merization of A content

Code (mole %) P A XU

PELG EBE-05 EBE-10 EBE-20 EBE-40 EBE-70

100 94.5 89.5 80.6 60.5 31.5

588 1 495 0.95 260 0.89 127 0.80 47 0.60 14


P E L G

€BE-10

€BE-20

5 10 15 20 2 5

2 e(degree1

Figure 1. solid films.

Wide-angle x-ray diffraction profiles of unoriented PELG and EBE

helical diameter of PELG was calculated to be 14.08 A. Valle et al.," however, found from x-ray analysis of PELG cast from dichloroethane and dimethyl formamide that the unit cell of the PELG was pseudo-hexagonal, i.e., monoclinic, with a packing helical diameter of 12.5 A. Apparently interhelical interactions of the ethyl, as opposed to the methyl group, caused PELG to pack in the monoclinic rather than the hexagonal form.

Formation of microheterophase structure

As can be seen in Table I, the A-block chain portion of EBE copolymers assumes an a-helical conformation in a helicogenic solvent. Polybutadiene exists in the random coil conformation in solvent. The molecular conformation of the EBE copolymer in solution is represented in Fig. 2, a is the radius of cross section of an a-helix, and h is the residue translation in the

I I I I I I

I I 1 I ,

1 A - b l o c k 1s-block I A - b l o c k I I

Figure 2. Molecular conformation of EBE in helicogenic solvent.

1140 SAT0 ET AL.

Sphere Cy I i nder Lamella

Micelles formed from A-B-A block copolymer solution at critical Figure 3. concentration.

a-helix. The B-chain is divided into two unit chains connected at the mid- point M of the B-chain, and the root mean square end-to-end distance and the root mean square radius of gyration of the unit chain having a degree of polymerization of PB/2 are denoted by ( Y & ~ ) ~ ’ ~ , and ( s & ~ ) ” ~ , respectively. Thus, an A-B-A triblock chain is assumed to be composed of two A-(B/2) diblock chains connected at the point M.

At the critical micelle concentration, the A-B-A block copolymer undergoes microphase separation and aggregates into micelles, such as the spherical, cylindrical, and lamella-like micelles, as illustrated in Fig. 3, in accordance with the copolymer composition, dimension of blocks, and environmen- tal conditions.

The Gibbs free energy AG per unit volume for the micelle formation is given by

A G = r A W - TAS

where r is the area of A/B interface per unit volume of micelle, AW (erg/cm’) is the interfacial free energy per unit area of A/B interface and is equal to the interfacial tension YAB (dyne/cm), and AS is the entropy change accom- panied by the micelle formation. According to Nakajima et al.,’ the equilibrium micelle dimensions D , eq, D , eq, and Leq, respectively, for spherical, cylindrical, and lamella-like micelles, are given by the following equations.

(1)

in which +B is the volume fraction of the domain occupied by B block to the total volume occupied by the copolymer, both in solution, given by Eq. (5).


N, the number of junction points between A and B block per unit volume of micelle, i.e., twice the number of block copolymer per unit volume of micelle, is given by

N = 1/[(4/3)~(~28/2)~'~ + nu2P~h] (6)

Accordingly, if the numerical values of (&2), (s&~), u, and AW are known, we can estimate the dimension of micelles from Eqs. (2), (3), and (4).

The solvent used for casting of membranes was a mixture of chloroform (CF) and trifluoroethanol (TFE) (4:l v/v) in which TFE is more volatile than CF. Furthermore, the polybutadiene used was rich in trans-1,4- configuration. At the critical micelle concentration, the content of TFE in the solvent mixture may be regarded as very low. So the (&) and (s&~) were estimated from literature data11J2 on poly(trans-1,4 butadiene)-chloroform, as 41.34 A and 16.88 A, respectively. The radius, u, of the cross-section of PELG helix was estimated as 6.58 8, using the intermolecular spacing 11.41 A obtained from x-ray data.

With respect to interfacial tension YAB of PELG/poly(trans-1,4 butadiene) system, nothing has been reported. We used the value of 25.7 x erg/A for AW, the value obtained for PMLG/poly(trans-1,4 butadiene) system. l3

Table I1 summarizes the data on microheterophase structures for EBE block copolymers calculated from Eqs. (2), (3), and (4), together with the dimensions DEM estimated from electron micrographs.

Electron microscopic observation on microheterophase structures

Figure 4 shows the electron micrographs of EBE-10, EBE-20, EBE-40, and EBE-70 membranes cast from a mixture of CF-TFE (4: 1 v/v). The domains of polybutadiene, stained with osmium tetroxide, correspond to the dark portions. Various microheterophase structures are observed depending on the content of polybutadiene: EBE-10 adopts a structure intermediate between sphere and cylinder (a) as shown in Fig. 4, EBE-20 adopts a cylindrical structure (b), EBE-40 adopts a lamella-like structure (c), and EBE-70 adopts a lamella-like structure (d), in which domains of PELG block exist in the matrix of polybutadiene.

TABLE I1 Micelle Dimensions of EBE Block Polymers

EBE-05 0.17 470 480 (sphere) EBE-10 0.28 429 420 (cylinder) EBE-20 0.44 341 380 (cylinder) EBE-40 0.68 264 350 (lamella) EBE-70 0.88 204 360 (lamella)

1142 SAT0 ET AL.

( c.) LEI,-30 0.1 )4 ( d l EBE-70 U

Figure 4. (4: 1 v/v) at 25°C. Dark portions correspond to B domain.

Electron micrographs of EBE membranes cast from CF:TFE

The formation of microheterophase structures were r ep~r t ed '~ , '~ to be dependent on compatibility with the casting solvent, and examined on EBE- 10 cast from the mixtures of CF and TFE (see Fig. 5). As seen in the electron micrographs, light portions corresponding to the PELG block increase with a decreasing ratio of CF to TFE in the mixtures. Because TFE is good solvent for PELG block in comparison with CF,I6 the end-to-end distance of the PELG block in the membrane appears to be enlarged with increasing ratio of TFE to CF in casting solvent.

Membrane permeability

The permeation of water through bIock copolymer membranes may be closely related to the microheterophase structure of the membranes. The

A-B-A TRIBLOCK COPOLYMER MEMBRANES

(a ) CF : T F E = 4 : 1

( C ) CF:TFE = 1:4

1143

Figure 5. mixing ratios (v/v) at 25°C.

Electron micrographs of EBE-10 cast from CF-TFE of various

permeability coefficient, K , of a membrane of thickness AX under hydraulic pressure difference Ap across the membrane is expressed by the equa t i~n '~

K =""("), At Ap (7)

where V, q, A, and t designate the volume of water, the viscosity of water, the area of membrane, and the elapsed time, respectively. If the water flux, If, per unit area of membrane is measured in unit time, Eq. (7) sim- plifies to18-'l

AX K = If(-&J (7')

1144 SAT0 ET AL.

“0 1 2 3 4

A P (atrn) Figure 6. temperatures for EBE-20 membrane cast from CF:TFE (2: 1 v/v) mixture.

Water flux Jf plotted against pressure difference Ap at various

The relation between the water flux and the pressure difference at various temperatures is illustrated in Fig. 6. In the measured range of pressure difference is proportional to Ap, i.e., the K values are constant because the membrane compression is negligibly small under the given condition, just as was observed in the GBG block copolymer membrane. The permeability coefficient at a given temperature was calculated from the plotted slope.

Figure 7 shows the Arrhenius plot: log K plotted against 1/T. The activation energies, calculated from the slopes, were 7.9 and 8.0 kcal/mole, respectively, for PELG and EBE-40. The activation energy of membrane permeation of MBM-11 was 8.2 kcal/mole, a little larger than that of PMLG, 7.6 kcal/mole. The activation energy of water permeation of GBG block copolymer membranes was, accordingly, two to three times as large as those of EBE and MBM block copolymer membranes. The water permeation region through GBG membrane seems to be different from that of EBE and MBM membranes.

According to Chang et al.,” the permeability constant K for steady-state diffusion is given by Ficks first law as

where Dw is the diffusion coefficient of water inside the membrane, V, is the volume fraction of water in the membrane, and V is the partial molar volume of water. Figure 8 shows the relation between K and V, for EBE and MBM block copolymer membranes. The K values change stepwise with increasing


- 1 0 E -18

x . w i -19 " - C x -

-20

I I 1 I I I I I I I

MBM-11 -18 - -%: - EBE-40 !-\I MBM-14

- EEE-05 v

X c - -

-20 - -

- PMLG-12

" -19 - -

- PELG -

- -

I I I I I I I I I I

V,, which means that these block copolymer membranes should be regarded as heterogeneous membranes in permeation behavior. Such a change of K values corresponds to the microheterogeneity of block copolymer membranes. In Fig. 9, the volume fraction of water V, in the membrane is plotted against the volume fraction +B of polybutadiene in nonswollen membrane. In the region of Fig. 9 labeled B, EBE-05 membrane exhibits the spherical

- 6 E + 2 3 E " 5 h X

Y

v

4

3 1 2 3 4 5 " B

vw ["I*) Figure 8. Relation between K and V, for EBE and MBM block copolymer membranes cast from CF:TFE (2 : l v/v) mixture. Solid circle and triangle correspond to data of PELG and PMLG, respectively.

1146 SAT0 ET AL.

6

2

0 0.2 Ok 0.6 0.8 1.0

$ Figure 9. same membranes as in Fig. 8.

V, plotted against volume fraction, $B, of polybutadiene in the

domain structure. In the C region, EBE-20 and MBM-13 membranes have a structure intermediate between the cylindrical and lamella forms. In the D region, both EBE-70 and MBM-11 membranes have a lamellar structure in which the @-helical polypeptide domains are distributed in a polybutadiene matrix because of the high content of polybutadiene. In the A region, no microphase separation occurs.

Relatively swollen membranes from 3% to 8% in V, in the D region, however, yield a lower permeability than permeability extrapolated from membranes of 1-2% in V, to high V, values, as seen in Fig. 8. In a homogeneous membrane, the permeability should increase initially in a linear relationship with the volume fraction of waterz1 and then exponentially in sufficiently high swollen membrane.u In anisotropic membranes made of styrene and 4-vinyl pyridine copolymer^,^^ the relation between permeability and membrane water content was observed to be exponential. In EBE and MBM membranes of large &,, the water content of which is 3-8%, there are some regions that do not serve as channels for water permeation. Therefore, it is presumed that water trapping by the structural distortion, generated by a closely linked and bulky a-helical structure, of the hydrophobic polybutadiene matrix works as an immobile container of water, because PELG and PMLG are hydrophilic materials.

The effect of casting solvent on the degree of swelling in water is examined, as shown in Table 111. A mixture of CF and TFE or TFA (trifluoroacetic acid) was used as casting solvent for membrane preparation. A larger CF content in the casting solvent caused a higher degree of water swelling. In comparison with CF, TFE, which is a good solvent for PELG,l6 disperses the polypeptide component monomolecularly and causes closer polypeptide packing in the membrane, but it also reduces the size of the polybutadiene domain, as observed in electron micrographs in Fig. 5. In the membrane preparation of MBM, a small amount of TFA in CF increases the volume fraction of water in the swollen membranes. Since TFA changes a-helical


TABLE I11 Effect of Casting Solvents on Membrane Preparation

Butadiene Casting solvent QW

Sample code (mole %) (V/V) (%I EBE-10 10.5 CF:TFE, 4: 1 2.82

CF:TFE, 2 : 1 1.11 CF:TFE, 1:4 0.95

MBM-13 19.7 CF:TFA, 8:1 3.96 CF:TEE, 2:1 0.88

MBM-12 31.9 CF:TFA, 8:1 4.71 CF:TFE, 2 : l 1.01

conformation of polypeptides into a partly random coil c~nformation,’~ MBM membranes cast from a mixture of CF and TFA should have some random coil conformation, which seems to result in high degree of swelling of MBM block copolymer membrane. Thus the degree of water swelling of EBE and MBM block copolymer membranes cast from TFA, CF, and TFE are in the following order: TFA S CF > TFE.

The next discussion leads to the mechanism of water permeation in microheterophase structure membranes of block copolymers composed of polypeptides and polybutadiene. Such microheterophase membranes are assumed to be composed of several regions such as (i) a relatively hydrophilic polypeptide domain region composed of a-helical chains, (ii) a hydrophobic polybutadiene domain region, (iii) a hydrophilic interfacial region3 existing between polypeptide domain and polybutadiene domain, and (iv) the space between micelles, i.e., a kind of microvoid. These regions should have different water permeabilities. It was pointed outz that these different mechanisms of water permeation give additive contributions to the permeability coefficient of membrane. The permeability coefficient K of microheterophase structure membrane is given by

K = K A ~ A + Kntf4intf + K B ~ B + K v + v (9) where the subscripts A, intf, B, and v denote polypeptide, interfacial, polybutadiene, and microvoid regions, respectively.

In an A-B-A block copolymer membrane, the polybutadiene domain is considered to take no part in water permeation because of its hydro- phobicity. This means that KB can be negligible. As shown in Fig. 8, permeability coefficients are on the order of lo-’ cm’/s - atm for EBE and MBM block copolymer membranes, from which it is anticipated that water perme- ates by means of diffusion flow in these so-called dense membranes. The radiusz6 r of a pore in a swollen membrane can be estimated by Eq. (10) from Poiseuille’s law and the method of Maneg~ld,’~ assuming that water channel of membrane is cylindrical in the direction of permeation, as

r = (3.8$2K/Wg)1’2 (10) where K is the permeability constant in c.g.s. unit, 1 is the thickness of swollen membrane, W is the water fraction in membrane per unit membrane

1148 SAT0 ET AL.

Figure 10. membranes.

K,ntf+intr plotted against +:'3 for EBE and MBM block copolymer

area, and g is the acceleration of gravity. The pore radius in EBE-05 membrane was calculated to be 2.74 A, which supports the prediction that the water permeation of the membrane is driven by the diffusion flow, and that there exists no microvoid surrounded by micelles in EBE and MBM block copolymer membranes. Thus the term KV+" is omitted, and Eq. (9) is written as Eq. (9') for EBE and MBM block copolymer membranes.

K = KA+A + K n t t +mtf

Because c,blntf is small compared with +A and (bB, it may be assumed in the first approximation that +A + GB = 1. Thus, K,ntf+,ntf values can be calculated.

As is obvious in Fig. 10, when the values Klntf &,tf are plotted against +%3,

good linearity between Klntr and +$3 is obtained for both EBE and MBM block copolymer membranes. An interfacial region contributing to +lntf

should exist along the outside of polybutadiene domain. The area of polypeptide/polybutadiene interface per unit volume of micelle, calculated from respective model for spherical, cylindrical, and lamella-like micelles, could be derived to be proportional to 4g3 from Eqs. (7), (8), (9), (13), (14), and (15) in a paper reported by us.' Finally, Klntf increases with +g3, which is in accordance with the dimension of area of the interfacial region. In other words, the thickness of the interfacial region is considered to be constant, regardless of structural change of micelles.

From the results mentioned above, the following conclusions on the water permeability of polypeptide membranes were reached: (1) Water permeation in both PMLG and PELG membranes takes place in the side chain region of the polypeptide a-helices, and K A of PMLG is larger than that of PELG, as shown in Fig. 8. Such conclusion is in accord with the results of Takizawa et al.28,29 and Minoura and N a k a g a ~ a ~ ~ that water vapor diffuses and per- meates through the side chain regions of the a-helices of PMLG membrane. (2) From the slopes of two straight lines in Fig. 10, Klntr of EBE is observed to

(9')

A-8-A TRIBLOCK COPOLYMER MEMBRANES 1149

be larger than that of MBM, which is attributed to the longer side chain of PELG than that of PMLG and to the better compatibility of ethyl group with polybutadiene than that of methyl group having better crystallinity in block copolymer membranes. (3) Taking into consideration that K value of GBG membranes was on the order of low6 cm'/s * atm, the GBG membranes should have microvoids among micelles. The benzyl group is less compatible with polybutadiene, and the crystallinity of the GBG membranes is better than that of EBE. Accordingly, the micelles formed in GBG membranes are better defined than those of EBE membranes. The aggregation of GBG molecules causes no structural disorder of micelles and results in the formation of well-defined micelles, but better orientation of rigid rod-like a-helices and bulky benzyl groups may result in microvoid. (4) Finally, the permeability coefficients KA, Kintf, KB, and K, for polypeptide block copolymer membranes are in the following order:

K , Kinif > K A > KB 2( 0 Takizawa et al.30 reported that a series of alcohols permeate through the

side chain region among helices of PMLG, and the diffusion coefficient decreased systematically with increasing molecular size of penetrants. A similar tendency is observed for the permeation of alcohol penetrants in EBE-10 membrane, as illustrated in Fig. 11. Table IV shows the permeability coefficient of EBE membranes for alcohols, which increases with polybutadiene content. Alcohol molecules appear to permeate not only through

I I I I I I I ]

0 1 2 3 4 5

Carbon Number d Penetrants

Figure 11. Permeability coefficients of EBE-10 membrane plotted against carbon number of alcohol penetrants: 0, H20; 1, methyl alcohol; 2, ethyl alcohol; 3, n-propyl alcohol; 4, n-butyl alcohol; 5 , n-amyl alcohol.

1150 SAT0 ET AL.

TABLE IV K of EBE Membrane Cast from Mixture of CF:TFE ( 4 : l v/v)

Permeant PELG EBE-05 EBE-10 EBE-20

CHSOH 26.4 39.4 CIHSOH 8.04 12.9 18.4 23.9

side chain region of a-helices but also through the interfacial region between polybutadiene and polypeptide domains.

Biocompatibility

The tissue compatibility of EBE block copolymer materials was examined, and the results in comparison with those' of MBM block copolymer materials are listed in Table V, where the levels to evaluate the biocompatibility of materials is as follows:6

(-) cells do not recognize the entity as a foreign body; (k) borderline recognition between (-) and (+); (+) cells recognize it as a foreign body, giant cells, and fibroblasts

gather close to the test materials, and 10 or fewer cell layers surround the foreign body, which corresponds to a light foreign body reaction;

(+ +) borderline recognition between (+) and (+ + +); (+ + +) muItiple layers of cells lump together close to the test material in

a strong foreign body reaction; (++++) the test material poisons to cells, causing necrosis.

The degree of absorbance by living body is evaluated according to the following indices?

(-) no low absorbance is observed; (+) about 25% of test material is absorbed;

(+ +) about 50% of test material is absorbed; (+ + +) about 75% of test material is absorbed;

(+ + + +) more than ca. 95% of test material is absorbed.

TABLE V Biocompatibility of EBE and MBM6 Materials

Sample Foreign body Absorbance by

4B reaction living body

PELG EBE-05 EBE-10 EBE-20 EBE-40 PMLG MBM-14 MBM-12

0 0.17 0.28 0.44 0.68 0 0.28 0.65

+ t + * + +.

+++ ?

-

- or + - or +

-

- + or ft t -


(a) PELG

[b) EBE-35

Figure 12. Photomicrographs of polymer samples and surrounding tissue: P, polyester fiber; G, giant cell; CT, connective tissue; arrow, sample material. Bar indicates 100 pm.

In comparison with MBM, GBG, and LBL, block copolymers composed of poly( e-N-carbobenzyloxy L-lysine) as the A component, it is surprising that the foreign body reaction and the degree of absorbance of EBE was low. As seen on microphotographs of polymer samples and surrounding tissue in Fig. 12, a mild foreign body reaction can be observed in the case of PELG, EBE, and PB. Inflammatory cell infiltration and granulomatous tissue was observed in the case of both LBL and MBM in the cylindrical structure, and GBG in the lamella-like structure. The ethyl ester apparently causes special

1152 SAT0 ET AL.

(2) ERE-.':)

Figure 12. (Continued from the previous page.)

biocompatibility, which may result from the worse ordering structure of EBE because of bad crystallinity, as mentioned about wide-angle x-ray diffraction. And as seen in Fig. 10, the permeability coefficients of the interfacial domain in EBE membranes are larger than those of MBM membrane, which may mean the existence of a more hydrophilic domain on the surface of EBE membranes than in the case of MBM membranes. Such hydrophilicity of EBE membranes may result in good biocompatibility.


Figure 12. (Continued from the previous page.)

In conclusion, EBE block copolymers look to be promising materials for biomedical application. An tithrombogenicity of EBE block copolymers will be reported somewhere so as to be compared with thrombogenesis of a triblock copolymer of similar structure. l5

The authors wish to thank Daikin Kogyo Co. for taking electron micrographs, Prof. Matsumoto in Kobe University for wide-angle x-ray diffraction measurement, and Prof. Kitamaru in Kyoto University for microbalance measurement.

1154 SAT0 ET AL.

References 1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

A. Nakajima, T. Hayashi, K. Kugo, and K. Shinoda, "Synthesis and structural study of the A-B-A type block copolymer consisting of poly( y-benzyl L-glutamate) as the A component and polybutadiene as the B component," Macromolecules, 12, 840-843 (1979). A. Nakajima, K. Kugo, and T. Hayashi, "Microheterophase structure of the A-B-A type block copolymer consisting of a-helical poly( y-benzyl L-glutamate) as the A component and polybutadiene as the B component,'' Macromolecules, 12, 844-848 (1979). A. Nakajima, K. Kugo, and T. Hayashi, "Mechanical properties and water permeability of A-B-A tri-block copolymer membranes consisting of poly(y-benzyl L-glutamate) as the A component and polybutadiene as the B component," Polym. J., 11, 995-1001 (1979). G. -W. Chen, T. Hayashi, and A. Nakajima, "Synthesis and molecular characterization of A-B-A tri-block copolymers composed of poly(y-ethyl L-glutamate) as the A component and polybutadiene as the B component, Polym. J., 13, 433-442 (1981). T. Hayashi, K. Kugo, and A. Nakajima, "Structure and properties of polypeptide block copolymer membranes," in Contemporary Topics in Polymer Science, Vol. 4, W. J . Bailey and T. Tsuruta, Eds., Plenum Press, New York, 1984, pp. 685-714. Y. Noishiki, Y. Nakahara, H. Sato, and A. Nakajima, "Study on the relationship between biocompatibility and chemical structure of amino acid homopolymers and copolymers," Artif. Organs, 9, 678-682 (1980). A. L. Lehninger, "Bioenergetics and metabolism," in Principles of Bio- chemistry, Worth, New York, 1982, p. 330. Y. Uematsu, "Viscoelastic properties of poly(y-benzyl glutamate)," Polym. J., 6, 431-437 (1974). C .H. Bamford, L. Brown, A. Elliot, W.E. Hanby, and I.F. Trotter, "Structure of synthetic polypeptides," Nature, 169, 357-360 (1952). G. Valle, A. Del Pra, P. Spadon, and M. Mami, "X-ray studies on poly(7-ethyl L-glutamate)," Biopolymers, 14, 437-439 (1975). F. T. Wall, "Statistical lengths of rubber-like hydrocarbon molecules," J. Cem. Pkys., 11, 67-71 (1943). M. Kurata, Y. Uchiyama, T. Koyama, and H. Fujita, "Properties of polybutadiene solution. 11. Transpolymer," Rep. 12th Polym. Symp. of SOC. Polym. Sci., Japan, 5, 129-131 (1963). A. Nakajima, K. Kugo, T. Hayashi, and H. Sato, "Microheterophase structure of A-B-A type block copolymers consisting of polypeptide- polybutadiene-polypeptide and their properties as biomaterials," &sen Koensku, 36, 125-138 (1979). A. Nakajima and T. Korenaga, "Micro-heterophase structure in membrane of vinylpyridine-styrene-vinylpyridine block copolymers and their quaternized products," Bull. lnst. Ckem. Res., Kyoto Univ., 52, 295-307 (1974). S. A. Barenberg, J . M. Anderson, and K. A. Mauritz, "Thrombosis: an epitaxial phenomenon. I," J. Bid . Mat. Res., 15, 231-245 (1981). M. Terbojevich, E. Piggion, A. Cosani, G. D'Este, and E. Scoffone, "Solution properties of synthetic polypeptides. Light scattering and viscosity of polyfy-ethyl L-glutamate) in dichloroacetic acid and trifluoroethanol," Eur. Polym. J., 3, 681-689 (1967). M. F. Refojo, "Polyelectrolyte complexes: Permeability to water and potential uses in ophthalmology," J. Appl. Polym. Sci., 11, 1991-2000 (1967). H. Yasuda, C.E. Lamaze, and A. Peterlin, "Diffusive and hydraulic permeabilities of water in water-swollen polymer membranes," J. Polym. Sci., A-2, 9, 1117-1131 (1971).


19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

H. Yasuda and C. E. Lamaze, ”Salt rejection by polymer membranes in reverse osmosis. I. Nonionic polymers,” J. Polym. Sci., A-2, 9, 1537-1551 (1971). H. Yasuda, C. E. Lamaze, and A. Schindler, “Salt rejection by polymer membranes in reverse osmosis. 11. Ionic polymers,” J. Polym. Sci., A-2, 9, 1579-1590 (1971). A. Peterlin, H. Yasuda, and H. G. Olf, “Flow of water through highly swollen membranes,” J. Appl. Polyrn. Sci., 16, 865-870 (1972). Y. J. Chang, C. T. Chen, and A. V. Tobolsky, “Correlations between types of absorbed water molecules and water permeability in swollen polymer membranes,” J. Polyrn. Sci. Polym. Pkys. Ed . , 12, 1-6 (1974). H. Yasuda and L. D. Ikenberry, “Permeabilities of coagulated cellulose acetate dialysis membranes,” J. Appl. Polym. Sci., 15, 2841-2853 (1971). L. J. Williams, A. Schindler, and A. Peterlin, “The influence of irradia- tion and hydration on the transport properties of isotropic and anisotropic membranes,” Mukromol. Ckem., 147, 175-184 (1971). M. Minoura and T. Nakagawa, ”The effect of casting solvent on the sorption and diffusion of water vapor in poly( y-methyl L-glutamate),” J. Appl. Polym. Sci., 23, 2729-2737 (1979). I. Sakurada, A. Nakajima, and H. Takita, ”Permeation of low polymers through membrane!" Kobunski Kuguku, 10, 522-529 (1954). V. E. Manegold, ”Uber Kollodiummembranen 111. Die Dialyse durch Kollodiummembranen und der Zusammenhang zwischen Dialyse, Dif- fusion und Membranstructur,” Koll.-Z., 372-395 (1929). A. Takizawa, T. Hamada, and J. Shimozawa, ”Hygroscopicity and permeability of water vapor and carbon dioxide gas through poly(y-methyl D-glutamate) membrane,” Kobunski Kuguku, 28, 751-755 (1971). A. Takizawa, T. Hamada, H. Okada, S. Kadota, and H. Nonoyama, “Permeability of water vapor and carbon dioxide gas through poly(n-alkyl L-glutamate),” J. Appl. Polym. Sci., 18, 1443-1450 (1974). A. Takizawa, T. Hamada, H. Okada, S. Imai, and S. Okada, ”Perme- ability of a series of alcohols through poly( y-methyl D-glutamate),” Poly- mer, 15, 157-161 (1974).

Received April 26, 1984 Accepted December, 1984

microheterophase structure, permeability, and biocompatibility of a-b-a triblock copolymer membranes...

Documents