Biomacromolecules 2008, 9, 2891–2897
2891
End-Functionalized Phosphorylcholine Methacrylates and their Use in Protein Conjugation Debasis Samanta, Samantha McRae, Beth Cooper, Yunxia Hu, and Todd Emrick* Polymer Science and Engineering Department, University of Massachusetts, 120 Governors Drive, Amherst, Massachusetts 01003
Jeanne Pratt and Stephen A. Charles Oligasis Corporation, Huntsville, Alabama 35806 Received June 20, 2008; Revised Manuscript Received August 1, 2008
Polymer-protein conjugation was performed using N-hydroxysuccinimide and aldehyde-terminated zwitterionic polymers, and the resulting polymer-protein conjugates were characterized by gel electrophoresis and fast protein liquid chromatography. Methacryloyloxyethyl phosphorylcholine (MPC) polymers were prepared by atom transfer radical polymerization in which the requisite functional end-groups for protein conjugation were embedded within the polymerization initiators. These phosphorylcholine polymers were conjugated to lysozyme as a model protein, as well as two therapeutic proteins, granulocyte colony stimulating factor (G-CSF) and erythropoietin (EPO). These MPC polymer-protein conjugates represent alternatives to PEGylated proteins, with the potential to provide improved efficacy in a therapeutic treatment relative to the protein itself.
Introduction Polymer-protein conjugation offers new possibilities in protein therapeutics.1-3 In particular, the conjugation of poly(ethylene oxide) (PEO) or poly(ethylene glycol) (PEG) to proteins, termed PEGylation, can improve the efficacy and efficiency of therapeutic proteins, leading to new therapeutic drugs for diseases such as hepatitis C and anemia. The advantages of PEGylation for both protein and small molecule drugs are numerous, including increased drug hydrodynamic volume over the protein alone, and decreased clearance rate, immunogenicity, aggregation in vivo, and associated deleterious side-effects. Permanent protein modification, by PEGylation or other methods, alters biological activity. While PEGylated therapeutic proteins typically exhibit less than 100% of the native protein activity, this characteristic is offset by the enhanced stability, circulation time, and therapeutic convenience offered by PEGylation. Polymer-drug conjugation is not limited to PEG, and in recent years various examples of polymer-drug conjugates have emerged, including those using N-hydroxypropylmethacrylamide (HPMA),4 as well as dendritic scaffolds that offer many end-groups for bioconjugation.5 For protein modification, PEGylation is central to current efforts, with fewer examples of non-PEG protein conjugates.6,7 However, conventional linear PEG, despite its numerous functional and heterobifunctional derivatives, cannot cover all the potential parameter space envisaged for protein therapeutics. In this context, modern polymerization methods are emerging for protein therapeutics. One such method is atom transfer radical polymerization (ATRP) using, for example, PEG methacrylate to produce comblike “polyPEG” structures. By appropriate selection of ATRPinitiator, polyPEG can be conjugated to proteins through conventional reactive chemistries or by grafting-from techniques using functionalized proteins.8 * To whom correspondence should be addressed. E-mail: tsemrick@ mail.pse.umass.edu.
Ionic polymerization methodology for preparing PEG is effective but has limited functional group tolerance, requiring postpolymerization modification of the PEG chain-ends to introduce functionality useful for protein conjugation. In contrast, free radical polymerization, with its inherently better functional group tolerance, has been used for many years to polymerize acrylates and methacrylates, including methacryloyloxyethyl phosphorylcholine (MPC). MPC polymers are hydrophilic and biocompatible, like poly(ethylene oxide), but can be tailored for applications such as biocompatible coatings and implants by copolymerization with other methacrylates.9 ATRP of MPC opens opportunities for phosphorylcholine polymers, due to the end-group fidelity and well-defined molecular weight and molecular weight distribution achievable through this chemistry. By incorporating the conjugating functional group into the ATRP initiator, the need for postpolymerization functionalization is precluded, which benefits the extension of MPC polymers to bioconjugation area. Here we report the preparation of N-hydroxysuccinimide (NHS) ester and aldehyde-terminated MPC polymers by Cu(I)catalyzed ATRP, and the use of these polymers in protein conjugation by amidation and imine formation, respectively. These hydrophilic, end-functionalized PC polymers exhibit low polydispersity indices (PDIs), as desired for polymer therapeutics, and effectively conjugate to proteins. Lysozyme was used as a model protein, and granulocyte colony stimulating factor (G-CSF) and erythropoietin (EPO) as therapeutic examples. The PC polymers were characterized by spectroscopic and chromatographic methods, while the conjugation reactions were monitored by high performance liquid chromatography (HPLC) and sodium dodecylsulfate polyacrylamide gel electrophoresis (SDS-PAGE).
Results NHS-functionalized ATRP initiators 1 and 2 were prepared according to literature procedures.7 Our initial attempts to use
10.1021/bm8006715 CCC: $40.75 2008 American Chemical Society Published on Web 09/25/2008
2892
Biomacromolecules, Vol. 9, No. 10, 2008
Samanta et al.
Scheme 1
these initiators for MPC polymerization encountered difficulties due to the narrow range of solvents useful for PC polymers relative to amphiphilic PEG-derivatives. While free radical polymerization of MPC in aqueous environments is generally effective hydrolysis of the NHS end-groups was observed under aqueous conditions, precluding subsequent conjugation. As depicted in Scheme 1, NHS-terminated polyMPC 3 and 4 were synthesized successfully in polar solvent mixtures, such as DMSO and MeOH. Polymer 3 was prepared from initiator 1 using the CuBr/2,2′-bipyridine catalyst/ligand system for ATRP10 in a 0.9 M monomer solution in 3:1 DMSO/MeOH. These conditions gave nearly quantitative monomer conversion
in 12 h. The polymer product was purified by elution through a short plug of silica gel to give a white solid. 1H NMR spectroscopy (Figure 1) was used to monitor monomer conversion, by the disappearance of monomeric olefin signals at 5.70 and 6.13 ppm, and the appearance of broad signals centered at 1.0 and 2.0 ppm for the methyl and methylene groups of the polymer backbone, respectively. The signal at 2.8 ppm, arising from the NHS methylenes, is useful for number-average molecular weight estimation by end-group analysis. Table 1 lists NHS-terminated MPC polymers of type 4 prepared using various monomer-to-initiator ratios, with molecular weights estimated by aqueous gel permeation chromatography (GPC) and 1H NMR
Figure 1. 1H NMR spectrum of polymer 3 in D2O. Table 1. MPC Polymerization Using Initiator 1 sample
target molecular weight (g/mol)
initiator/monomer ratio
% conversion by 1 H NMR
molecular weight by 1 H NMR (g/mol)
aq. GPC: Mn (g/mol)
PDI (Mw/Mn)
1 2 3 4
3200 4500 7500 11000
1:10 1:15 1:25 1:37
complete complete complete complete
2300 4000 5900 8200
3500 5200 7700 9600
1.5 1.2 1.5 1.3
Phosphorylcholine Methacrylates in Protein Conjugation
Figure 2. Aqueous GPC traces: (A) polymer 3 prepared using monomer-to-initiator ratios of (a) 25:1, (b) 15:1, and (c) 10:1; (B) polymer 10 prepared using a monomer-to-initiator ratio of 27:1.
end-group analysis (integrating the signals at 2.8 and 1.0 ppm). Figure 2 shows GPC traces of NHS-terminated MPC polymers prepared for this study. NHS-terminated polyMPC samples 3 and 4 were used to conjugate lysozyme under different aqueous buffer conditions (from pH 6 to 9.4) and at different functional group stoichiometry and concentration. The conjugation reactions were monitored by HPLC with size exclusion columns, as well as by SDS-PAGE. In general, NHS-MPC polymers of type 4 gave much better conjugation to lysozyme than 3, owing to a reduced steric hindrance in the vicinity of the reactive end-group (one methyl group vs two). Such steric sensitivity was not surprising in light of reports on protein PEGylation with polyPEG derivatives.8 Moreover, successful conjugation relied on working within an appropriate concentration range. Dilute conditions, using for example 10 mg of polymer 4 and 1 mg of lysozyme in 10 mL HEPES buffer at pH 9.0 gave only ∼25% protein conjugation in 24 h (conversion estimated from the relative areas of protein and conjugate in the HPLC trace). However, at 10fold higher concentration, 80% conversion was seen in 24 h, and nearly quantitative conversion was achieved following a second addition of polymer (Figure 3). Clearly, the higher concentration is preferred, with the dual benefit of increasing the reaction rate and reducing the impact of competitive hydrolysis of the NHS end-group. Further evidence in support of lysozyme conjugation with polymer 4 was obtained by SDSPAGE, which separates the PC-protein conjugate from unreacted lysozyme (Figure 3). The conjugates smear on the gel (as seen for PEGylated proteins7,8) due to inherent polydispersity of the attached chains and variability in conjugation sites. Lysozyme itself appears as a distinct band corresponding to 15 kDa. Polymer-lysozyme conjugates were purified by fast protein liquid chromatography (FPLC), using Superose 6 10/300
Biomacromolecules, Vol. 9, No. 10, 2008
2893
preparative columns, eluting with PBS buffer at a flow rate 0.5 mL/min. As seen in Figure 4, the conjugate elutes over the range of 14-19 mL, at significantly lower retention volume relative to lysozyme itself (20-21 mL), due to the larger size of the conjugate. The bimodal nature of the conjugate peak is expected for samples possessing a heterogeneity in the number of conjugated polymers per protein. Typically, the elution volume of conjugate was collected in 5 × 1 mL fractions, each then analyzed further by HPLC with size-exclusion columns. In addition to NHS-ester conjugation, we also considered conjugation of lysine residues to aldehyde-functionalized MPC polymers. We chose a benzaldehyde terminus, as benzaldehydes are convenient synthetically and well-tolerated in ATRP, and useful for conjugation with amines. Few examples of proteinpolymer conjugation with benzaldehydes have been reported. One example, from Kataoka and co-workers, describes the synthesis of a heterotelechelic PEG containing pyridyldithiol at one chain-end and benzaldehyde at the other.11 Reaction of benzaldehyde with free amines gives imine linkages, stable under physiological conditions (unlike the aliphatic versions which require reductive amination), and hydrolyzable under acidic conditions. Standard borohydride reduction to amines can also be performed following the initial conjugation. Benzaldehyde functionalized ATRP initiators 7 and 9 were prepared from readily available 4-hydroxybenzaldehyde. As shown in Scheme 2, initiator 7 was prepared by acylation of 4-hydroxybenzaldehyde, and 9 from the ethanol-extended derivative. Both 7 and 9 were purified by column chromatography on silica gel (eluting with ethyl acetate/hexane mixtures), and characterized by 1H and 13C NMR spectroscopy and high resolution mass spectrometry. Compound 7 is known from the reports on ATRP of poly(methyl methacrylate) (PMMA) by Haddleton and co-workers,12 while compound 9 is novel and chosen to provide an aliphatic hindered ester as an alternative to the phenyl ester of 7. ATRP of MPC was performed from these benzaldehyde-functionalized initiators used a CuBr/2,2′bipyridine catalyst/ligand system, as for the NHS case. Monomer conversion was followed by 1H NMR spectroscopy, by decreaseing intensity of monomer olefin signals, and a corresponding increasing intensity in broad signals for the aliphatic protons of the polymer. Benzaldehyde-terminated polyMPC 10 was purified by column chromatography on a short plug of silica gel. The isolated polymer displayed the expected MPC polymer signals in the 1H NMR spectrum in D2O (Figure 5), aromatic proton signals at 7.15 and 7.91 ppm, and the characteristic aldehyde proton resonance at 9.77 ppm. These aromatic signals are convenient for estimating molecular weight by 1H NMR end-group analysis, which provides a useful comparison to the data obtained by aqueous GPC. Table 2 lists examples of prepared benzaldehyde-terminated MPC polymer 10, and their estimated molecular weights using these methods. As was typical for these polymerizations, nearly quantitative monomer conversion was achieved when targeting molecular weights in the 8000 g/mol range; lower conversion was obtained when targeting higher molecular weights. Benzaldehyde-terminated MPC polymer 10 (Mn 9700 g/mol) was then tested in protein conjugation, again using lysozyme as a test case. Reaction of lysozyme with 10 equiv of benzaldehyde-terminated 10 in pH 6 buffer produced the desired conjugates by imine formation. In some cases, the conjugates were reduced with sodium cyanoborohydride. The conjugation reactions were monitored by size exclusion (SEC) and cation exchange (CE) high performance liquid chromatography (HPLC), as well as SDS-PAGE. Broad lysozyme-polymer conjugate
2894
Biomacromolecules, Vol. 9, No. 10, 2008
Samanta et al.
Figure 3. HPLC traces (top) and SDS-PAGE (bottom) for the reaction of polymer 4 with lysozyme: (A) conjugation at 1 mg/mL polymer, monitored by HPLC using a Shodex protein KW-803 column; (B) conjugation (at 10 mg/mL) and monitored by HPLC using Shodex protein KW-803 and KW-804 columns in series. SDS-PAGE, right-to-left: (a) protein standards (from top to bottom: 209.0, 124.0, 80.0, 48.1, 34.8, 28.9, 20.6, 7.1 kDa); (b) lysozyme; (c) reaction mixture from conjugation reaction at pH 9.0; (d) conjugate after purification by FPLC (fraction at 18 mL from FPLC).
Figure 4. SEC-FPLC chromatogram of lysozyme-polyMPC conjugation. Conjugate elutes first, between 14 and 19 mL. Unreacted lysozyme is seen at a retention volume of 20 mL.
signals were observed by HPLC, using both size exclusion and cation exchange columns, as shown in Figure 6. The conjugates were also analyzed by SDS-PAGE, using gradient 4-20% TrisHCl polyacrylamide precast gels with 25 mM Tris, 192 mM glycine, 0.1% (w/v) SDS buffer at pH 8.3. The conjugates appeared on the gel as higher molecular weight smears accompanied by a distinct unreacted lysozyme band, similar to the results obtained in the case of NHS-ester conjugation. To test the reactivity of these benzaldehyde-terminated MPC polymers on therapeutic proteins, conjugation reactions were performed with G-CSF and EPO, again analyzing for conjugation by size-exclusion HPLC. G-CSF, a four-helix bundle protein of ∼20 kDa, is a cytokine that regulates differentiation of hematopoietic progenitor cells toward mature neutrophils, and is used to protect chemotherapy patients against infections.13 The G-CSF protein drug filgrastim (Neupogen) aggregates under physiological conditions, diminishing its therapeutic benefit,
while PEGylated G-CSF (i.e., Neulasta) exhibits much longer serum half-life. Conjugation of 8 (Mn 7000 g/mol) to G-CSF was performed in 100 mM sodium acetate buffer (pH 4.6), containing 40 mM sodium cyanoborohydride. Aliquots from the reaction were analyzed periodically on a Shodex KW-803 size exclusion column using at a flow rate of 1 mL/min. EPO, a glycoprotein of ∼40 kDa, stimulates proliferation of erythrocytes into mature red blood cells, and is used clinically to treat conditions such as anemia stemming from chemotherapy.14 The short (
