Hyperbranched PEG by Random ... - Wiley Online Library

5 downloads 0 Views 315KB Size Report
Aug 5, 2010 - Daniel Wilms, Martina Schцmer, Frederik Wurm, M. Iris Hermanns,. C. James ... exhibits a maximum of two functional end groups, limiting.
Communication

Hyperbranched PEG by Random Copolymerization of Ethylene Oxide and Glycidola Daniel Wilms, Martina Scho ¨mer, Frederik Wurm, M. Iris Hermanns, C. James Kirkpatrick, Holger Frey*

The synthesis of hyperbranched poly(ethylene glycol) (hbPEG) in one step was realized by random copolymerization of ethylene oxide and glycidol, leading to a biocompatible, amorphous material with multiple hydroxyl functionalities. A series of copolymers with moderate polydispersity (Mw =Mn < 1.8) was obtained with varying glycidol content (3–40 mol-%) and molecular weights up to 49 800 g mol1. The randomly branched structure of the copolymers was confirmed by 1H and 13C NMR spectroscopy and thermal analysis (differential scanning calorimetry). MTS assay demonstrated low cell toxicity of the hyperbranched PEG, comparable to the highly established linear PEG.

Introduction Conspicuous by its remarkable biocompatibility, chemical inertness, and excellent solubility in both organic and aqueous media, poly(ethylene glycol) (PEG) and its derivatives have found a tremendous variety of applications in food production, cosmetics, and pharmaceutics.[1,2] PEG is D. Wilms, M. Scho ¨mer, F. Wurm, H. Frey Institute of Organic Chemistry, Johannes Gutenberg-University, Duesbergweg 10-14, D-55099 Mainz, Germany Fax: (þ)49 6131 39 26106; E-mail: [email protected] M. I. Hermanns, C. J. Kirkpatrick Institute of Pathology, Johannes Gutenberg-University, Langenbeckstrasse 1, D-55101 Mainz, Germany y Daniel Wilms and Martina Scho ¨mer contributed equally to this work. a

: Supporting information for this article is available at the bottom of the article’s abstract page, which can be accessed from the journal’s homepage at http://www.mrc-journal.de, or from the author.

Macromol. Rapid Commun. 2010, 31, 1811–1815 ß 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

also used to enhance the circulation lifetime of biologically active molecules such as proteins or low molecular weight drugs (PEGylation).[3,4] However, PEG is associated with a high degree of crystallization due to efficient packing of the linear polyether chains, and this has represented a drawback for a number of applications. In addition, linear PEG exhibits a maximum of two functional end groups, limiting its potential for further chemical modification and, for example, the loading capacity when used as soluble support for organic synthesis.[5] Hawker et al.[6] described a multi-step approach to branched PEG structures by incorporation of an AB2 macromonomer containing an aromatic branching unit. Gnanou and coworkers[7] synthesized dendrimer-like PEGs by an elegant, divergent approach, based on repeated allylation, multihydroxylation, and polymerization of ethylene oxide (EO). On the other hand, hyperbranched polyglycerol,[8,9] obtained by controlled anionic ring-opening polymerization of glycidol under slow monomer addition conditions,[10,11] has been discussed as an alternative to PEG, since it combines excellent biocompatibility[12] with high functionality and

wileyonlinelibrary.com

DOI: 10.1002/marc.201000329

1811

D. Wilms, M. Scho ¨mer, F. Wurm, M. I. Hermanns, C. J. Kirkpatrick, H. Frey

amorphous nature, albeit obvious limitations with respect to industrial use due to the costly glycidol monomer. Very recently, Gnanou’s group developed a sequential procedure for the polymerization of propylene oxide, interrupted by glycerol units as branching points, onto a PEG macroinitiator.[13] However, the few pathways to (hyper)branched PEG derivatives reported to date clearly do not fulfill the requirements for a simple, one-step procedure allowing control over degree of crystallization, terminal functionalities, and solubilityies. Direct copolymerization of EO with glycidol as branching agent is a self-evident, but synthetically challenging approach to hyperbranched PEG copolymers. Vastly different boiling points of the monomers (11 and 161 8C, respectively) and low reactivity of glycidol at low temperature have hitherto limited the obtained copolymers to broadly distributed materials without control over the macromolecular properties and less than 3% incorporation of glycidol.[14] Here we present a simple one-step approach to random copolymers of EO and glycidol by anionic ring-opening copolymerization, allowing for the first time the synthesis of well-defined branch-on-branch PEG copolymers by random incorporation of glycerol units into the poly(ethylene oxide) backbone. The resulting random poly(ethylene glycol)-co-poly(glycerol) copolymers, herein termed P(EG)-co-P(G), consist exclusively of ethylene glycol and glycerol repeat units, assuring maximum biocompatibility.

Experimental Part Typical Procedure for the Preparation of Random Copolymers of Ethylene Oxide and Glycidol A two-necked glass flask equipped with a septum, teflon seal, and a magnetic stirrer was connected to a vacuum line. The desired amount of 1,1,1-tris(hydroxymethyl)propane (TMP) was introduced to the flask and suspended in benzene. After stirring the

suspension for 30 min, the flask was evacuated for at least 3 h to remove traces of water azeotropically as well as other volatiles. The flask was then filled with argon before freshly distilled diglyme was introduced to dissolve the TMP, followed by dropwise addition of freshly prepared potassium naphthalide solution in THF to deprotonate 10% of the hydroxyl groups. The solution was stirred for 1 h under argon. Subsequently, the initiator solution was cooled to 80 8C and the flask was evacuated. EO was transferred to an ampoule, dried over calcium hydride and subsequently transferred to the reaction flask in vacuo. The flask was sealed and freshly distilled glycidol was introduced through the septum via cannula. The reaction mixture was then immediately heated to 80 8C and stirred for 18 h. After addition of excess methanol, the solution was dialyzed against water for 3 d. Removal of water under reduced pressure and drying in vacuo for 24 h at 80 8C afforded the respective P(EG)-co-P(G) copolymer in ca. 90% yield.

Results and Discussion Synthesis A well-established trifunctional initiator for anionic epoxide polymerization, TMP, has been employed in our studies, and the glycidol comonomer fraction was varied from 3 to 50%. The straightforward synthetic protocol to multifunctional hyperbranched PEGs (cf. Scheme 1) can be summarized as follows: (i) simultaneous addition of the monomer mixture (EO and glycidol) to the dissolved, deprotonated initiator in vacuo, (ii) random copolymerization at elevated temperature (80 8C) in an aprotic aliphatic ether solvent (diglyme), and (iii) protic termination to release the hydroxyl groups. Combination of the monomers and initiator in high vacuum at low temperature, followed by fast heating and concurrent pressurization of the reaction mixture, represents an important difference in comparison to a previous approach, as well as the use of potassium naphthalide for deprotonation of the TMP-initiator.[14]

Scheme 1. Synthetic approach to P(EG)-co-P(G) random copolymers with TMP core and multiple hydroxyl functionalities. The average linear segment length n depends on the comonomer ratio.

1812

Macromol. Rapid Commun. 2010, 31, 1811–1815 ß 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

DOI: 10.1002/marc.201000329

Hyperbranched PEG by Random Copolymerization of Ethylene

employed to control the anionic hyperbranching polymerization of glycidol.[8,9] Size exclusion chromatography (SEC) yields molecular weight distributions with Mw =Mn between 1.1 and 1.8 (Table 1) after either dialysis or precipitation to remove a small low molecular weight fraction (5%) that is observed in some cases. Polymer yields are generally above 90%. Evidence for Branched Structure 1

H NMR analysis of the copolymers in DMSO-d6 permits to calculate the ratio of comonomers incorporated by determination of the fraction of glycerol units Figure 1. 13C NMR spectrum of P(EG)-co-P(G) with a glycidol content of 41% (sample #1), from the intensity of the hydroxyl proton inset: schematic structure of different repeating units present in P(EG)-co-P(G). signals. However, unequivocal evidence for random comonomer incorporation can only be obtained by IG 13C NMR spectroscopy. Besides The end group functionality of the resulting copolymer depends on the initiator-core functionality and the number the remarkable difference of the physical properties of of incorporated glycerol branching units and can thus be both comonomers there is a certain bias that the reactivity directly adjusted by the comonomer ratio. Incorporated of EO is much higher than of substituted oxiranes, rendering glycidol monomer may result in branching points (if both random copolymerization impossible. However, recent functional groups propagate) as well as linear or terminal works on the linear copolymerization of EO with various units. The assignment in Figure 1 is based on literature glycidyl ethers clearly demonstrate similar reactivity ratios data[8] as well as two-dimensional NMR spectroscopy for copolymerization[16] and compatible copolymerization 13 1 that allows a correlation between C and H shifts kinetics,[17,18] resulting in random linear copolyethers. (cf. Supporting Information). The degree of branching Thus, we assumed that also glycidol should exhibit similar (DB, cf. Table 1) can be calculated from Inverse Gated (IG) reactivity as glycidyl ethers, acting, however, as a cyclic 13 latent AB2-comonomer due to proton transfer, leading to C NMR spectra.[15] It is remarkable that hyperbranched PEG copolymers of moderate polydispersity index an unprecedented hyperbranched PEG topology upon (PDI < 1.8) with high molecular weights and functionality copolymerization with EO. (adjusted by the glycidol feed) can be obtained, despite As can be seen from Figure 2, the incorporation of glycerol of deviating from a slow addition protocol, as is typically units into the PEG scaffold leads to the appearance of Table 1. Characterization data of the hyperbranched P(EG)-co-P(G) copolymers.

No.

[G] Fraction monomer feed

[G] Fractiona)

[G] Fractionb)

DBc)

Mn d)

Mw =M n d)

g mol1

Tge)

Tme)

DHe)

-C

-C

J g1





1

0.50

0.41

0.46

0.43

38 700

1.83

61

2

0.30

0.31

0.30

0.35

25 000

1.44

60





3

0.20

0.16

0.12

0.20

30 100

1.40

62

10

9.2

4

0.15

0.15

0.10

0.17

49 800

1.10

61

8

35.2

5

0.10

0.12

0.09

0.13

47 300

1.28

58

13

45.4

6

0.05

0.09

0.07

0.11

36 000

1.24

53

22

69.8

7

0.03

0.07

0.04

0.07

23 400

1.41

60

37

102.1

a) Fraction of incorporated glycerol units calculated from inverse gated (IG) 13C NMR spectra; b) Fraction of incorporated glycerol units calculated from 1H NMR; c)Degree of branching, calculated from IG 13C NMR spectra; d)Determined by SEC in DMF versus PEG standards; e) Glass transition Tg and melting points Tm determined by differential scanning calorimetry (second heating run, 10 8C min1).

Macromol. Rapid Commun. 2010, 31, 1811–1815 ß 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.mrc-journal.de

1813

D. Wilms, M. Scho ¨mer, F. Wurm, M. I. Hermanns, C. J. Kirkpatrick, H. Frey

at around 15% glycerol units. The gradual change of the melting point and melting enthalpy with branched comonomer fraction once again supports random incorporation of the glycerol units, thus confirming the NMR results. Structural characteristics, such as the degree of crystallization, functionality (up to 300 hydroxyl end groups), solubility, and degree of branching can be controlled by adjusting the comonomer feed ratio [G]/[EO]. Glycidol fractions as low as 5% already result in honey-like materials Figure 2. Inverse gated 13C NMR spectrum of P(EG)-co-P(G) with increasing fractions of with low Tg at 60 8C, completely differincorporated glycidol comonomer. ent from the highly crystalline linear PEG. This behavior is also in clear contrast to PEG star triad sequences due to the presence of neighboring polymers,[21] which show a pronounced tendency to comonomer units, as is well known from the random crystallize at elevated arm lengths exceeding 10 EO units. copolymerization of EO with propylene oxide or protected glycidyl ethers.[16–20] Higher [G]/[EO] feeds gradually lead to a reduction of the EO triad intensity at 70–72 ppm Biocompatibility and concurrently to the separation of the respective Cytotoxicity of P(EG)-co-P(G) #4 against L-929 mouse glycerol signals, as it becomes increasingly probable that fibroblasts and primary human endothelial cells has been two glycerol units are located adjacent to each other. studied using an MTS assay and compared to linear PEG Assignment of all peaks to either one of the comonomer (Figure 3). Both cell lines showed a high degree of units permits calculation of the comonomer incorporation biocompatibility for the hyperbranched PEG samples, from IG 13C NMR spectra and the degree of branching (DB) demonstrating high potential of the new hyperbranched by referencing to the methine carbon of the glycerol units at PEG analogs for biomedical application. a chemical shift around 78 ppm (Table 1). Thermal Behavior Thermal analysis by differential scanning calorimetry (DSC) reveals a continuous shift of the melting point towards lower temperatures with an increasing fraction of glycerol branch points. Remarkably, a [G] fraction as low as 4 mol-% is sufficient to decrease the melting point by 30 8C compared to a linear PEG homopolymer of comparable molecular weight. A further increase of the glycerol fraction leads to a continuous decrease of melting temperature and enthalpy, until the melting endotherm eventually vanishes

Conclusion In summary, the hyperbranched PEG materials obtained from the direct random copolymerization of EO and glycidol are promising for a plethora of applications in biomedicine and cosmetics as well as for novel Li-ion conductors, combining unusual structural features with straightforward synthetic accessibility. Even low amounts of glycerol branching points have been demonstrated to possess a dramatic effect on the properties of the resulting materials, when compared to linear PEG.

Acknowledgements: H. F., D. W., and M. S. acknowledge financial support by the Fonds der Chemischen Industrie as well as by the Max Planck Graduate Center (MPGC) with the Johannes Gutenberg-University Mainz.

Received: June 3, 2010; Published online: August 5, 2010; DOI: 10.1002/ marc.201000329 Figure 3. Cytotoxicity of linear PEG and hyperbranched copolymers (sample #4) against HUVEC cells (left) and L-929 cells (right) at different concentrations after 48 h of incubation.

1814

Macromol. Rapid Commun. 2010, 31, 1811–1815 ß 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Keywords: anionic ring-opening polymerization; biocompatible; hyperbranched; polyethers; poly(ethylene glycol)

DOI: 10.1002/marc.201000329

Hyperbranched PEG by Random Copolymerization of Ethylene

[1] J. M. Harris, S. Zalipsky, Poly(ethylene glycol)–-Chemistry and Biological Applications, American Chemical Society, Washington, DC 1997. [2] R. Gref, Y. Minamitake, M. Peracchia, V. Trubetskoy, V. Torchilin, R. Langer, Science 1994, 263, 1600. [3] S. Zalipsky, Adv. Drug Delivery Rev. 1995, 16, 157. [4] M. E. Fox, F. C. Szoka, J. M. J. Fre´chet, Acc. Chem. Res. 2009, 42, 1141. [5] T. J. Dickerson, N. N. Reed, K. D. Janda, Chem. Rev. 2002, 102, 3325. [6] C. J. Hawker, F. Chu, P. J. Pomery, D. J. T. Hill, Macromolecules 1996, 29, 3831. [7] X.-S. Feng, D. Taton, E. L. Chaikof, Y. Gnanou, J. Am. Chem. Soc. 2005, 127, 10956. [8] A. Sunder, R. Hanselmann, H. Frey, R. Mu ¨ lhaupt, Macromolecules 1999, 32, 4240. [9] D. Wilms, F. Wurm, J. Nieberle, P. Bo ¨hm, U. Kemmer-Jonas, H. Frey, Macromolecules 2009, 42, 3230. [10] D. Wilms, S.-E. Stiriba, H. Frey, Acc. Chem. Res. 2010, 43, 129.

Macromol. Rapid Commun. 2010, 31, 1811–1815 ß 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

[11] M. Caldero´n, M. A. Quadir, S. K. Sharma, R. Haag, Adv. Mater. 2010, 22, 190. [12] R. K. Kainthan, J. Janzen, E. Levin, D. V. Devine, D. E. Brooks, Biomacromolecules 2006, 7, 703. [13] X. Feng, D. Taton, E. L. Chaikof, Y. Gnanou, Macromolecules 2009, 42, 7292. [14] P. Dimitrov, E. Hasan, S. Rangelov, B. Trzebicka, A. Dworak, C. B. Tsvetanov, Polymer 2002, 43, 7171. [15] H. Frey, D. Ho ¨lter, Acta Polym. 1999, 50, 67. [16] X. Pang, R. Jing, J. Huang, Polymer 2008, 49, 893. [17] C. Mangold, F. Wurm, B. Obermeier, H. Frey, Macromol. Rapid Commun. 2010, 31, 258. [18] B. Obermeier, F. Wurm, H. Frey, Macromolecules 2010, 43, 2244. [19] F. Heatley, G.-E. Yu, C. Booth, T. G. Blease, Eur. Polym. J. 1991, 27, 573. [20] V. Rejsek, D. Sauvanier, C. Billouard, P. Desbois, A. Deffieux, S. Carlotti, Macromolecules 2007, 40, 6510. [21] G. Lapienis, Prog. Polym. Sci. 2009, 34, 852.

www.mrc-journal.de

1815