biodegradable poly(disulfide)s derived from raft polymerization: monomer scope, glutathione...

Biodegradable Poly(disulfide)s Derived from RAFT Polymerization:Monomer Scope, Glutathione Degradation, and Tunable ThermalResponsesDaniel J. Phillips and Matthew I. Gibson*

Department of Chemistry, University of Warwick, Coventry, CV4 7AL, United Kingdom

*S Supporting Information

ABSTRACT: Telechelic, RAFT (reversible addition−frag-mentation chain transfer)-derived macromonomers with apyridyl disulfide end-group were converted into highmolecular weight, disulfide-linked polymers using a poly-condensation, step-growth procedure. The applicability of themethod to polycondense a library of macromonomers withdifferent functionalities including (meth)acrylates and acryl-amides was investigated. Side-chain sterics were found to beimportant as nonlinear poly(ethylene glycol) analogues, whichproved incompatible with this synthetic methodology, as were methacrylates due to their pendant methyl group. This methodwas used to incorporate disulfide bonds into poly(N-isopropylacrylamide), pNIPAM, precursors to give dual-responsive(thermo- and redox) materials. These polymers were shown to selectively degrade in the presence of intracellular concentrationsof glutathione but be stable at low concentrations. Due to the molecular weight-dependent cloud point of pNIPAM, the lowercritical solution temperature behavior could be switched off by a glutathione gradient without a temperature change: anisothermal transition.

■ INTRODUCTIONResponsive or “smart” materials are capable of undergoing aphysical response upon application of an external stimulus.These materials are finding application in a diverse range offields including switching surfaces and adhesives, artificialmuscles, sensors,1 and may find use in biomedical fields such asdrug delivery,2 gene delivery,3 and tissue engineering.4

Synthetic polymers exhibiting a lower critical solution temper-ature (LCST) have been extensively investigated as smart,thermoresponsive materials. Upon increasing the solutiontemperature above the cloud point (the measurable propertyof an LCST), an aqueous polymer solution becomes insolubleand aggregates/precipitates. This property can be exploited foreither drug release5 or hyperthermia-triggered cellular uptakedue to increased lipophilicity above the LCST.5−7 Examples ofpolymers display ing this behavior include poly-[(oligoethyleneglycol)methyl ether methacrylate] (pOEG-MA),8,9 poly(N-isopropylacrylamide) (pNIPAM),10 and poly-(N-vinylpiperidone).11 While temperature changes are useful,some applications may require a change in polymer solubility(i.e., to switch between “active” and “inactive” states) withoutapplying a thermal gradient. “Isothermal” transitions havepreviously been demonstrated by Alexander et al. based on saltconcentration gradients,12,13 Moeller et al. following aminolysisof RAFT-derived polymer chains14 and by Rimmer et al. due tobacterial binding.15 We have demonstrated that selectivecleavage of a single polymer end-group results in a shift inpNIPAM cloud point, allowing an isothermal transition basedon bioreduction.16

Polymer degradation is a critical consideration for most invivo applications such as drug or gene delivery where cytotoxicpolycations are frequently employed. Degradation is typicallyachieved by the incorporation of one or more cleavable linkersinto either the polymer backbone, side-chain, or as a cross-linker.17,18 Common degradation triggers include hydrolysis,thermolysis, or enzymatic action.19−22 A major challengeassociated with the development of biodegradable polymers isthe introduction of functional groups onto the polymerbackbone. For example, ring-opening polymerization of N-carboxyanhydrides or cyclic esters, which give degradablepoly(amides) or poly(esters) respectively, are incompatiblewith most functional groups.23−25 Conversely, controlledradical polymerization processes enable a vast range offunctional groups to be incorporated into a polymer structure,but give rise to an all-carbon backbone that cannot degrade.26

To address the above paradox, we have previously reported asynthetic route toward degradable, main-chain disulfide bond-containing pNIPAM. This was achieved by polymerizingNIPAM using a RAFT (reversible-addition fragmentationchain transfer) agent containing a pyridyl disulfide moiety atthe α-terminus. Following aminolysis of the ω-terminaldithioester, a polycondensation-type, step-growth polymer-ization with release of pyridine thione was possible. Wedemonstrated the degradability of the material by the addition

Received: June 28, 2012Revised: September 7, 2012Published: September 10, 2012

Article

pubs.acs.org/Biomac

© 2012 American Chemical Society 3200 dx.doi.org/10.1021/bm300989s | Biomacromolecules 2012, 13, 3200−3208

pubs.acs.org/Biomac

of the reducing agent tributylphosphine, which producedpolymer chains with a far higher LCST than its disulfide-linkedcounterpart. Hence, a novel method to “switch off” LCSTbehavior using a secondary chemical stimulus was illustrated.27

A similar procedure combining aminolysis and thiol-disulfideexchange has also been recently applied to alkyl disulfides.28

The redox-sensitive nature of disulfide-containing polymersis of particular interest for triggered cellular deliveryapplications. The main reducing agent (antioxidant) insidecells is glutathione (GSH) where it is present at mMconcentrations. Conversely, the extracellular GSH concen-tration is only μM, hence, this differential provides a uniqueand selective trigger to promote intracellular degradation whileensuring stability in the circulation.29,30 The use of disulfide-containing polymers has seen a particular focus on genedelivery therapies to date31−36 though there are also reportsutilizing redox-sensitive materials for targeted, drug deliveryapplications.37−43

This study explores the scope of our synthetic methodologyas a means of preparing controlled radical polymerization-derived, highly functionalized disulfide-linked polymers. Thebiodegradability of these materials using biorelevant glutathioneconcentrations is investigated to ensure selectivity for intra-cellular conditions. Finally, thermally responsive, degradablepolymers are tested for their isothermal response to glutathionegradients to enable their LCST behavior to be “switched off”once inside a cell.

■ EXPERIMENTAL SECTIONMaterials. All chemicals were used as supplied unless stated.

Methanol, hexane, ethyl acetate, dichloromethane, toluene, acetone,40−60 °C petroleum ether, tetrahydrofuran, diethyl ether, glacialacetic acid (analytical reagent grade), and 1,4-dioxane (analyticalreagent grade) were all purchased from Fisher Scientific at laboratoryreagent grade unless otherwise stated. Deuterated chloroform (99.9atom % D), aldrithiol-2 (98.0%), 2-mercaptoethanol (≥99.0%), 4-cyano-4-(phenylcarbonothioylthio)pentanoic acid (≥98.0%), dodec-ane thiol (≥98.0%), tribasic potassium phosphate (reagent grade,≥98.0%), 2-bromo-2-methylpropionic acid (98.0%), 4-dimethylami-nopyridine (≥99.0%), N,N′-diisopropylcarbodiimide (99.0%), N-isopropylacrylamide (97.0%), ethanolamine (≥99.0%), tributylphos-phine (97.0%), triethylamine (≥99.0%), tris(2-carboxyethyl)-phosphine hydrochloride (≥98.0%), glutathione (99.0% reduced),and mesitylene (analytical standard) were all purchased from Sigma-Aldrich. Poly(oligoethylene glycol), average Mn = 300 g·mol−1, methylether methacrylate, di(ethylene glycol) methyl ether methacrylate(95.0%), poly(oligoethylene glycol), average Mn = 480 g·mol−1,methyl ether acrylate, tert-butyl acrylate (98.0%), methyl methacrylate(99.0%), and N,N-dimethylacrylamide (99.0%) were also purchasedfrom Sigma-Aldrich, but inhibitors were removed by passing through acolumn of basic alumina before polymerization.Physical and Analytical Methods. NMR spectroscopy (1H, 13C)

was conducted on a Bruker DPX-300, Bruker DRX-500, or a BrukerAV III 600 spectrometer using deuterated chloroform as solvent. Allchemical shifts are reported in ppm (δ) relative to tetramethylsilane(TMS). High resolution mass spectra were recorded on a Brukerelectrospray ultra-high resolution tandem TOF mass spectrometerusing electrospray ionization (ESI) in positive mode on samplesprepared in methanol. Degraded sample mass spectral analysis wascarried using a Bruker MaXis UHR-Q-TOF mass spectrometer usingESI in positive mode over a scan range of 500−7000 m/z. Sampleswere prepared in methanol, diluted 20-fold in 50:50 methanol/waterand introduced by direct infusion at 90 μL·hr−1. Source conditionswere as follows: end plate offset at −500 V; capillary at −4500 V;nebulizer gas (N2) at 1.6 bar; dry gas (N2) at 8 L·min−1; drytemperature at 180 °C. Ion transfer conditions were as follows: ion

funnel RF at 400 Vpp; multiple RF at 400 Vpp; quadruple low massset at 455 m/z; collision energy at 5.0 eV; collision RF at 1200 Vpp;ion cooler RF at 250−600 Vpp; transfer time set at 121 μs; prepulsestorage time set at 15 μs. Calibration was completed with sodiumformate (10 mM). FTIR spectra were acquired using a Bruker Vector22 FTIR spectrometer with a Golden Gate diamond attenuated totalreflection cell. A total of 64 scans were collected on samples in theirnative (dry) state. SEC analysis was performed on one of two systems.(i) Dimethylformamide: Varian 390-LC MDS system equipped with aPL-AS RT/MT autosampler, a PL-gel 3 μm (50 × 7.5 mm) guardcolumn, two PL-gel 5 μm (300 × 7.5 mm) mixed-D columns usingDMF with 5 mM NH3BF4 at 50 °C as the eluent at a flow rate of 1.0mL·min−1. The GPC system was equipped with ultraviolet (UV) (setat 280 nm) and differential refractive index (DRI) detectors. Narrowmolecular weight PMMA standards (200−1.0 × 106 g·mol−1) wereused for calibration using a second order polynomial fit. (ii)Tetrahydrofuran: Varian 390-LC MDS system equipped with a PL-AS RT/MT autosampler, a PL-gel 3 μm (50 × 7.5 mm) guard column,two PL-gel 5 μm (300 × 7.5 mm) mixed-D columns equipped with adifferential refractive index and a Shimadzu SPD-M20A diode arraydetector, using THF (including 2% TEA) as the eluent at a flow rate of1.0 mL·min−1. Narrow molecular weight PMMA standards (200−1.0× 106 g·mol−1) were used for calibration using a second orderpolynomial fit.

The cloud points were measured using an Optimelt MPA100system (Stanford Research Systems). The recorded turbidimetry curvewas normalized between values of 0 and 1. The cloud point wasdefined as the temperature corresponding to a normalized absorbanceof 0.5. A polymer concentration of 5 mg·mL−1 and a constant heatingrate of 1 °C·min−1 were used in all experiments. UV−vis measure-ments were undertaken on a Biotech Synergy HT and processed usingthe Gen5 software package.

Procedures. Synthesis of 2-(Pyridyldisulfanyl) Ethyl 2-(Dode-cylthiocarbonothioylthio)-2-methylpropanoate. 2-(Pyridyldisulfan-yl) ethyl 2-(dodecylthiocarbonothioylthio)-2-methylpropanoate wasprepared using a method similar to that already reported.44

Hydroxyethyl pyridyl disulfide (1.61 g, 8.57 mmol), 2-(dodecylth-iocarbonothioylthio)-2-methylpropanoic acid (2.50 g, 6.86 mmol), and4-dimethylaminopyridine (0.25 g, 2.06 mmol) were dissolved indichloromethane (50 mL). N,N′-Diisopropylcarbodiimide (1.08 g,8.57 mmol) was added to the solution under ice and dropwise over aperiod of 20 min. The mixture was left to warm to room temperatureand stirred for 24 h. The observed precipitate was removed by gravityfiltration and the solvent concentrated in vacuo. The resulting yellowresidue was purified by column chromatography using silica gel as thestationary phase and 49:1 hexane/ethyl acetate mixture as eluent toyield a yellow/orange oil (1.07 g, 29% yield). 1H NMR (600 MHz,CDCl3) δppm: 8.47 (1H, d, J19−18 = 4.14 Hz, H19), 7.73 (1H, d, J16−17 =7.86 Hz, H16), 7.66 (1H, t, J17−16, 17−18 = 7.71 Hz, H17), 7.10 (1H, t,J18−17, 18−19 = 6.24 Hz, H18), 4.36 (2H, t, J14−15 = 6.36 Hz, H14), 3.27(2H, t, J12−11 = 7.56 Hz, H12), 3.03 (2H, t, J15−14 = 6.36 Hz, H15), 1.70(3H, s, H13), 1.65 (2H, p, J11−12, 11−10 = 7.50 Hz, H11), 1.37 (2H, p,J10−11, 10−9 = 7.50 Hz, H10), 1.23−1.31 (16 H, m, H2−9), 0.88 (3H, t,J1−2 = 6.96 Hz, H1). 13C NMR (600 MHz, CDCl3) δppm: 221.48 (C

13),177.80 (C16), 159.90 (C19), 149.68 (C23), 137.09 (C21), 120.77 (C22),119.76 (C20), 63.36 (C17), 55.79 (C14), 37.21 (C18), 37.00 (C12),31.91, 29.71, 29.62, 29.55, 29.44, 29.34, 29.10, 22.69 (C2−9), 28.93(C10), 27.86 (C11), 25.34 (C15), 14.13 (C1). IR cm−1: 3049 (aryl-Hstretch); 2923 (alkyl-H stretch); 1735 (CO stretch); 1063 (S-(CS)-S stretch). HRMS (ESI+) m/z (C24H40NO2S5 [M + H]+): Calcd,534.1657; found, 534.1661.

Polymerization of N-Isopropylacrylamide. In a typical procedure,N-isopropylacrylamide (0.25 g, 2.21 mmol), 2-(pyridyldisulfanyl) ethyl2-(dodecylthiocarbonothioylthio)-2-methylpropanoate (59.0 mg,11.05 μmol), and 4,4′-azobis(4-cyanovaleric acid) (6.19 mg, 2.21μmol) were dissolved in methanol/toluene (1:1; 4 mL) in a glass vialcontaining a stir bar. Mesitylene (100 μL) was added as an internalreference and the mixture was stirred (5 min). An aliquot of thisstarting mixture was removed for 1H NMR analysis. The vial was fittedwith a rubber septum and degassed by bubbling with nitrogen gas (30

Biomacromolecules Article

dx.doi.org/10.1021/bm300989s | Biomacromolecules 2012, 13, 3200−32083201

min). The vial was then placed in an oil bath thermostatted at 70 °C.After 4 h, the reaction mixture was opened to air and quenched inliquid nitrogen. An aliquot was removed and conversion determinedby 1H NMR. The product was purified three times by precipitationfrom toluene into diethyl ether, isolated by centrifugation, and driedunder vacuum overnight to give a yellow solid. The overall monomerconversion was determined from the 1H NMR spectrum by measuringthe decrease in intensity of the vinyl peaks associated with themonomer relative to mesitylene. Conversion (NMR): 91.7%; Mn(theoretical), 2100 g·mol−1; Mn (SEC), 2000 g·mol−1; Mw/Mn(SEC), 1.18.Polymerization details of other monomers used in this study are

included in the Supporting Information.General Polycondensation Procedure. Polymer with pyridyl

disulfide end-group (170 mg) was dissolved in tetrahydrofuran (2 mL)in a Schlenk tube equipped with a stir bar and rubber septum. Thissolution was degassed by a minimum of three freeze−pump−thawcycles. A separate solution containing ethanolamine (5 equiv) andtriethylamine (2 equiv) was degassed by a minimum of three freeze−pump−thaw cycles. The amine solution was added to the polymersolution by a syringe purged with dry nitrogen, and the solution left tostir at room temperature for five days. After this time, the mixture hadturned bright yellow. The polymer was purified by multipleprecipitations from tetrahydrofuran into diethyl ether and the productisolated by centrifugation. After drying overnight under vacuum, awhite solid was achieved. This product was then analyzed by SEC. Thesame sample was also treated with tributyl phosphine (2 drops) andreanalyzed by SEC to confirm disulfide-bond presence.General Procedure for Turbidimetry-Monitored, Glutathione-

Mediated Degradation of Disulfide-Containing Polymers. A 5mg·mL−1 stock solution of pNIPAM (theoretical Mn = 4000 g·mol−1)in water was prepared. A total of 3 × 180 μL aliquots of this stock weretransferred into three individual wells of a 96-well plate. The plate wasleft to incubate at 37 °C in the plate reader until all samples hadprecipitated out of solution. Once this had been observed, the platewas removed and the well volume made to 200 μL with distilled wateror glutathione solution to give final glutathione concentrations of 0, 1μM, and 1 mM. The plate was reincubated at 37 °C and theabsorbance at 650 nm was recorded for an additional 2.5 h.General Procedure for SEC Evaluation of Glutathione-Mediated

Degradation. Polycondensed pNIPAM (theoretical Mn = 40000g·mol−1, 16 mg) was dissolved in distilled water (3.96 mL) in a glassvial (vial A). A total of 1.98 mL of this stock solution was transferredto a separate vial (vial B). To vial A, 20 μL of a 100 μM glutathionesolution was added, and to vial B 20 μL of a 100 mM glutathionesolution to give final concentrations of 1 μM and 1 mM, respectively.Both vials were left to stir at ambient conditions with a 4 mg aliquotremoved after 24 and 72 h. Aliquots were freeze-dried andsubsequently analyzed by SEC (DMF).

■ RESULTS AND DISCUSSIONThe aim of this work was to exemplify our previous report onthe synthesis of disulfide-linked pNIPAM, obtained by apolycondensation-type reaction of pyridyl disulfide-terminated,RAFT-derived polymers, Scheme 1.27 Previously we used adithioester chain transfer agent (Scheme 2, compound 1)which, although an excellent mediator for the polymerization ofa wide range of monomers, is less synthetically accessible thantrithiocarbonate RAFT agents. We therefore prepared thetrithiocarbonate 4, propanoic acid 2-[[(dodecylthio)-thioxomethyl]thio]-2-methyl-2-(2-pyridinyldithio)ethyl ester(PADE) using a method modified from that described byO’Reilly et al.45 The acid functionalized RAFT agent 2 wascoupled to alcohol 3 in the presence of N,N-diisopropylcarbo-diimide and 4-dimethylaminopyridine to give PADE (Scheme2, compound 4) in reasonable yield. Structure and purity wasconfirmed by 1H NMR, 13C NMR, and high resolution massspectrometry (Supporting Information).

Our strategy for the preparation of disulfide-linked polymersfrom RAFT-derived macromonomers requires aminolysis of theRAFT agent to yield a thiol-terminated polymer. This can thenreact in situ with pyridyl disulfide groups on neighboring chains(Scheme 1) giving a polycondensation type, step-growthpolymerization, with elimination of pyridine thione. To verifythat our method could be extended to a trithiocarbonate CTA,pNIPAM-1 was prepared using PADE (Table 1, entry 1).Attempts to polycondense in the presence of oxygen failed toproduce a substantially higher molecular weight polymer due tooxidative side reactions giving macromonomer- and dimer-derived species only (Figure 1A). When degassed (Figure 1A),the same bimodal trace was observed, albeit with a small highmolecular weight tail indicating either a slow rate of reaction orthe presence of residual oxygen. To circumvent the formerissue, triethylamine (TEA) was added to the degassed solution.In this case, a higher molecular weight polymer was formedwith only small amounts of residual macromonomer observed.Inclusion of base was therefore crucial, presumably byincreasing thiol nucleophilicity by deprotonation. To verifythat the resulting polycondensed product contained the desireddisulfide linkages, it was incubated with the strong reducingagent tributylphosphine and analyzed by SEC (Figure 1B). Atrace with a molecular weight slightly lower than the startingmaterial (Figure 1B) was observed, concurrent with disulfidebond cleavage. The loss of end-groups and the potentiallydifferent elution behavior of thiol-terminated chains couldexplain the difference in observed molecular weight betweenstarting and cleaved materials. These results confirm thattrithiocarbonates provide suitable end-groups for our newpolycondensation-type polymerization.Our study subsequently focused on extending the poly-

condensation methodology to a wider range of polymericprecursors. A representative library of polymers was synthesizedusing PADE covering methacrylate, acrylate, and acrylamidefunctionalities. All polymers were characterized by SEC and 1HNMR and found to be well-defined with narrow polydispersityindices (Table 1). These polymers were subjected to theoptimized polycondensation conditions (vide supra) and theresults are summarized in Table 2.

Scheme 1. Synthetic Strategy for the Preparation of aDisulfide-Linked Polymer from Trithiocarbonate StartingMaterial: (i) Polymerization of Monomer; (ii)Polycondensation-Type Reaction, Employing Nucleophileand Base, With Liberation of Pyridine Thione



Polycondensation of poly(N,N-dimethylacrylamide)(pDMA) proceeded to give a noticeable increase in both Mw

(weight-average molecular weight) and Mp (peak maximummolecular weight; Table 2, entry 9; Figure 2). Identical resultswere observed for pDMA obtained from the precursor preparedwith dithioester 1 as a control experiment (SupportingInformation), indicating that this strategy can be applied tomonomer types other than NIPAM; a key aim of this study.The average number of “macro” repeat units per chain wasestimated from SEC and shown in Table 2. All successfulpolymerizations gave 2−3 disulfides (3−4 repeat units)whereas the unsuccessful (vide infra) gave numbers close to 0.

A particularly interesting class of polymers is based onoligo(ethylene glycol) methacrylates (OEGMAs). Thesepolymers are indicated to be biocompatible, similar to theirparent linear PEG, but are synthetically accessible by controlledradical polymerization. Furthermore, they also display tunableLCST behavior which can be exploited for a range ofbiotechnological applications.8,46 It is important to note thatthe pOEG(M)As prepared in Table 1 have notably highermolecular weights than the other polymers due to theirrelatively long side chains. They were designed to have similardegrees of polymerization to allow direct comparison withother polymers used in this study. Initial attempts topolycondense pOEGMA300 proved unsuccessful, indicating

Scheme 2. RAFT Agents Used in This Study: (A) Chemical Structure of Pyridyl Disulfide-Containing Dithiobenzoate (1); (B)Synthetic Strategy for the Preparation of Pyridyl Disulfide-Containing Trithiocarbonate (PADE, 4)

Table 1. Polymers Prepared in This Study Using PADEa

entry polymer [M]/[CTA] conversionb (%) Mn(th)b (g·mol−1) Mn(SEC)

c (g·mol−1) Mw/Mnc

1 pNIPAM-1 20 91.7 2100 2000 1.182 pNIPAM-2 40 91.4 4100 3100 1.163 pNIPAM-3 488 73.3 40500 25400d 1.24d

4 pOEGMA300 15 89.9 5400 18300 1.125 pOEGA480 22 85.1 10600 9600 1.16 pDEGMA 20 80.4 3000 23000 1.247 pMMA 30 58.1 1700 6400 1.388 ptBuA 20 91.0 2300 2500 1.169 pDMA 34 93.1 3000 2800 1.1

apNIPAM = poly(N-isopropylacrylamide); pOEGMA300 = poly[oligo(ethylene glycol, average Mn = 300 g·mol−1) methyl ether methacrylate];pOEGA480 = poly[oligo(ethylene glycol, average Mn = 480 g·mol−1) methyl ether acrylate]; pDEGMA = poly(diethylene glycol methyl ethermethacrylate); pMMA = poly(methyl methacrylate); ptBuA = poly(tert-butyl acrylate); pDMA = poly(N,N-dimethylacrylamide). bDetermined by1H NMR relative to an internal standard (mesitylene). cDetermined by SEC (THF inc. 2% TEA) relative to PMMA standards unless otherwisestated. dDetermined by SEC (DMF inc. 5 mM NH3BF4) relative to PMMA standards.

Figure 1. SEC data showing the polycondensation of pNIPAM-1: (A) Optimization of the polycondensation conditions using an inert atmosphereand a base; (B) Verifying the redox-responsive nature of the polycondensed species.



dimers and no higher order polymers could be formed (Table2, entry 4; Figure 3A). Xu et al. observed that thiolactones canform between a terminal thiol, generated by aminolysis of aRAFT end-group, and the penultimate repeat unit of a methylmethacrylate backbone.47 While this explanation for thioldeactivation fits our observations, there are other reports of thesuccessful reaction of RAFT-derived thiols with pyridyldisulfide groups in acrylate and methacrylate polymers.48 Analternative explanation therefore may be steric hindrance of theterminal thiol by the oligo(ethylene glycol) side chains. Thistheory is supported by visual examination of the reactionmixture: when pyridyl disulfide is successfully displaced by athiol, the released pyridine thione has a strong yellowcoloration. In the case of pOEGMAs, however, no colorchange was observed, suggesting no thiol-pyridyl disulfidereaction had occurred. To further probe this, poly-(diethyleneglycol methacrylate), pDEGMA, which has a shorterPEG chain (2 units, as opposed to 7) was tested but yieldedonly a minor increase in Mp (Table 2, entry 6; Figure 3B). Theincrease in Mw of approximately 50% suggested some dimer-formation was occurring, while the solution remained colorlessindicating limited or no pyridine thione elimination. Toconfirm that the reaction was not kinetically limited, it wasleft for an extended time period (7 days). Analysis after thistime showed no additional change to that seen after 48 h,suggesting that pOEGMAs cannot be polycondensed in thisfashion. Another reason for the successful polycondensation ofpDMA and pNIPAM compared to pOEGMA could be the lack

of a backbone methyl group. The resulting secondary thiolwould be expected to be more nucleophilic than a tertiary thiol.To address this, poly[oligo(ethyleneglycol) methyl etheracrylate] (pOEGA) was also synthesized using PADE.However, subsequent polycondensation gave only a highermolecular weight shoulder of approximately double the startingmaterial molecular weight (Table 2, entry 5; Figure 3C),confirming that that the PEG side chain can limit thepolycondensation process. To elucidate the role of the pendantmethyl group, polycondensation of a poly(methyl methacry-late) macromonomer, pMMA, which should have no sterichindrance was attempted (Table 2, entry 7). SEC analysisshowed no evidence for multi disulfide-linked polymers evenwith extended reaction times (4 days, see SupportingInformation). We propose the tertiary thiol present at thechain end of poly(methacrylate)s is not sufficiently nucleophilicto react with the pyridyl disulfide moiety and thereforemethacrylates are not compatible with this methodology.To further investigate steric effects at the thiol terminus, tert-

butyl acrylate was investigated given its branched structure issimilar to NIPAM and DMA. It also lacks the long side-chainpresent in OEGMA, DEGMA, and OEGA and has no pendantmethyl group. Polycondensation saw a 4-fold increase in Mwand Mp, indicative of a disulfide-linked polymer (Table 2, entry8; Figure 4). This demonstrates that acrylates are compatiblewith our technique to generate disulfide-linked polymers, butthat sterics must be considered. Postpolymerization modifica-tion may offer an approach to overcome this.

Table 2. SEC Characterization of Poly(disulfides)a

entry poly (disulfide) Mwb (SM) Mw

b (Pc) Mpb (SM) Mp

b (Pc) −S−S links/chainc Mw/Mnb (Pc)

1 ss-pNIPAM-1 2400 8900 2400 7600 2.2 1.812 ss-pNIPAM-2 3600 20300 3700 15300 3.1 1.783 ss-pNIPAM-3 31200d 128900d 35900d 140700d 2.9 2.01d

4 ss-pOEGMA300 22100 59600 20000 38500 0.9 1.385 ss-pOEGA480 28500 44200 26600 32600 0.2 1.56 ss-pDEGMA 10500 15700 10400 10900 0.0 1.277 ss-pMMA 8900 8800 9000 8900 0.0 1.618 ss-ptBuA 2900 12200 3100 11700 2.8 2.359 ss-pDMA 2800 8300 3100 11100 2.6 1.45

ass-pNIPAM = disulfide-linked poly(N-isopropylacrylamide); ss-pOEGMA300 = disulfide-linked poly[oligo(ethylene glycol, average Mn = 300g·mol−1) methyl ether methacrylate]; ss-pOEGA480 = disulfide-linked poly[oligo(ethylene glycol, average Mn = 480 g·mol−1) methyl ether acrylate];ss-pDEGMA = disulfide-linked poly(diethylene glycol methyl ether methacrylate); ss-pMMA = disulfide-linked poly(methyl methacrylate); ss-ptBuA= disulfide-linked poly(tert-butyl acrylate); ss-pDMA = disulfide-linked poly(N,N-dimethylacrylamide). bWeight-average molecular weight (Mw),peak maximum molecular weight (Mp), or polydispersity index (Mw/Mn) of starting material (SM) or polycondensed product (Pc) determined bySEC (THF inc. 2% TEA) relative to PMMA standards unless otherwise stated. cAverage number of disulfide links per chain followingpolycondensation {1 − [Mp(Pc)/Mp(SM)}. dDetermined by SEC (DMF inc. 5 mM NH3BF4) relative to PMMA standards.

Figure 2. Polycondensation of pDMA: (A) SEC data; (B) chemical structure of polymer precursor.



The work described above utilized relatively low molecularweight polymers, as it was anticipated that these would showfaster rates of polycondensation. It was therefore prudent toattempt to apply this methodology to larger precursors, whichin turn may give rise to higher molecular weight, disulfide-linked polymers. This is particularly important for drug deliveryapplications where the polymeric delivery device should ideallybe larger than the renal filtration limit (∼40 kDa) but havedegradation products below this.17 In the case of pNIPAM,tuning the molecular weight of the polymeric precursors alsoallows control over their lower critical solution temperature.27

Two further pNIPAM homopolymers were therefore preparedwith PADE to give molecular weights (by SEC) of 3100(pNIPAM-2) and 25400 (pNIPAM-3) g·mol−1, respectively(Table 1, entries 2 and 3). The molecular weight dependentcloud point (CP) of these materials was measured with thesmall and large precursors, giving CPs of 22 and 31 °C,

respectively (Figure 5C,D). Both polymers were successfullypolycondensed into disulfide-linked materials (Table 2, entries2 and 3; Figure 5). It should be noted that these reactions wereperformed on a larger scale (∼100 mg) than those shown in theabove section (5−10 mg) and therefore required longerreaction times to achieve high conversions (5−7 days). ss-pNIPAM-3 gave an Mw above the renal filtration limit (Mw =128900 g·mol−1), while the precursor Mw of 31200 g·mol

−1 wasbelow this. Turbidity measurements indicated a higher CP forboth disulfide-linked polymers relative to their precursors. ss-pNIPAM-2 had a CP of 30 °C (Figure 5C), while ss-pNIPAM-3 had a CP of 32 °C (Figure 5D), an increase of 8 and 1 °C,respectively. Although pNIPAM generally exhibits an inverselyproportional relationship between cloud point and molecularweight, the increase observed here can be attributed to a changein end-group functionality, namely, loss of the highlyhydrophobic dodecane thiol and pyridyl disulfide moieties.

Figure 3. SEC data showing the polycondensation reactions of (A) pOEGMA300, (B) pDEGMA, and (C) pOEGA480. Bottom-right: Chemicalstructure of polymer precursors.

Figure 4. Polycondensation of pt-BuA: (A) SEC data; (B) chemical structure of polymer precursor.



We have recently reported the importance of end-groups onthe cloud point of pNIPAMs.16

The redox-sensitivity of these disulfide-linked pNIPAMs wasagain demonstrated by the addition of tributylphosphine(Figure 5A,B), resulting in a reduction in molecular weightback to the original polymer. To assess the effect of disulfidedegradation on the CP, the samples were treated with TCEP, awater-soluble phosphine, and analyzed by turbidimetry (Figure5C,D). The CP of ss-pNIPAM-2 following TCEP reductionincreased compared to the disulfide-linked starting material, asexpected given the inversely proportional relationship betweenCP and pNIPAM molecular weight. The CP of ss-pNIPAM-3following TCEP reduction did not change as pNIPAMs CP isrelatively constant above 10000 g·mol−1.27 Analysis of ss-pNIPAM-2 by mass spectrometry following the addition oftributylphosphine confirmed that degradation was via disulfidebond reduction. Lower molecular weight, thiol-terminatedpolymers were observed (see Supporting Information), thus,

ruling out alternative modes of degradation such as hydrolysisor reduction of the ester units.To test degradation under more biorelevant conditions,

glutathione (GSH) was used as the reducing agent. This in vivoreducing agent is present at μM concentrations in theextracellular environment but mM concentrations in theintracellular environment. Previous work has shown thatpoly(disulfides) are preferentially cleaved at higher GSHconcentrations, providing a selective, intracellular trigger forpolymer degradation.34,35,37 Both ss-pNIPAM-2 and ss-pNIPAM-3 were stirred in aqueous glutathione solutions at 1μM and 1 mM concentrations for 24 h. The samples were thenanalyzed by turbidimetry (Figure 5C,D). An increase in CP wasagain observed, with the change larger than that obtained withTCEP for both samples. This is possibly due to a difference inthe mechanism of degradation: TCEP works by simple disulfidebond cleavage, while GSH, a tripeptide containing a cysteineresidue, is more likely to degrade by thiol-disulfide exchange.There is hence a difference in hydrophilicity of the generated

Figure 5. SEC data showing polycondensation of (a) pNIPAM-2 and (b) pNIPAM-3. Cloud point data (polymer concentration =5 mg·mL−1) of (c)pNIPAM-2/ss-pNIPAM-2 and (d) pNIPAM-3/ss-pNIPAM-3. Note: GSH concentration = 1 mM.

Figure 6. SEC data showing degradation of (a) ss-pNIPAM-2 and (b) ss-pNIPAM-3 at different GSH concentrations.



end-groups. After stirring in aqueous GSH solution thepolymers were freeze-dried and analyzed by SEC. After 72 hat mM concentrations, both polymers had undergonesignificantly more degradation than those incubated at μMGSH concentration (Figure 6), indicating the specific responseto intracellular-GSH.Given the large change in CP following incubation of ss-

pNIPAM-2 with GSH (shift of 9 °C), we investigated whetherthis could be used as a switch to trigger isothermal polymerresolubilization. This would be useful, for example, followingthermally triggered uptake into hyperthermic cells7 where theincreased GSH concentration should lead to degradation andan increase in cloud point. This will prompt polymerresolubilization, which is preferable for eventual excretion,and removes the issue of insoluble polymer aggregates insidecells. Three aqueous polymer solutions of ss-pNIPAM-2 wereincubated at 37 °C and their absorbance at 650 nm wasobserved. Each sample aggregated in solution as expected giventhe CP of the polymer was 30 °C. Glutathione was then addedto give total GSH concentrations of 0 (control), 1 μM, and 1mM. While the polymer solutions with 0 and 1 μM GSHconcentrations remained turbid, resolubilization of the polymerheld at 1 mM GSH concentration occurred rapidly withinminutes as the CP exceeded that of the temperature at which itwas being held (Figure 7). This was confirmed by the rapid

decrease in absorbance and was also observed visually (Figure7, inset). The same experiment using ss-pNIPAM-3 yielded thesame results (see Supporting Information). It was not possibleto estimate the kinetics of degradation from this experimentdue to the cooperative behavior of pNIPAM. We have recentlydemonstrated that mixtures of pNIPAMs with different chainlengths (and hence CPs) show a single transition, controlled bythe mass fraction of each chain length.49 Therefore, completeresolubilization will occur when sufficient numbers of shortchains are present to increase the CP above the localtemperature. This is a desirable feature allowing for morerapid responses without being limited by chain degradationkinetics.These experiments demonstrate that disulfide-linked poly-

mers obtained by RAFT polymerization can be used as doublyresponsive polymeric carriers with potential application in drugdelivery or biosensing fields.

■ CONCLUSIONSThis manuscript demonstrates the polycondensation of pyridyldisulfide-terminated, RAFT-derived polymers to obtain alibrary of disulfide-linked polymers. The use of trithiocarbonateRAFT agents widens the scope of this versatile method.Investigations into the polycondensation process indicated thatmacromonomer side-chain length limits the reactivity of thethiol end-groups and hence prevents the polymerization ofnonlinear PEG analogues. Moreover methacrylates, on accountof their pendant methyl group, were found to be incompatible.A range of acrylamides and acrylates were shown to beapplicable to this methodology. Dual-responsive materials werecreated by introducing disulfides into the main chain of thethermoresponsive pNIPAM. The disulfide unit showed a strongselectivity for glutathione degradation at concentrationsexpected in an intracellular environment. This, coupled withthe molecular-weight dependent cloud point behavior ofpNIPAM enabled a change in solubility to be triggered withoutthe need for a change in applied temperature, an isothermaltransition. This may be important in some biological situationswhere, for example, an LCST-type transition is desired whileavoiding protein denaturation. Ultimately this method providesa valuable way of incorporating redox-responsive linkages intopolymer structures en route to preparing increasingly complexmaterials with multiple responses.

■ ASSOCIATED CONTENT*S Supporting InformationAdditional details on synthesis, characterization, and isothermaltransition analysis are included. This material is available free ofcharge via the Internet at http://pubs.acs.org.

■ AUTHOR INFORMATIONCorresponding Author*Fax: +44(0)2476 524112. E-mail: [email protected].

NotesThe authors declare no competing financial interest.

■ ACKNOWLEDGMENTSEquipment used was supported by the Innovative Uses forAdvanced Materials in the Modern World (AM2), with supportfrom Advantage West Midlands (AWM) and partly funded bythe European Regional Development Fund (ERDF). M.I.G. is aBirmingham Science City Interdisciplinary Research Fellowfunded by the Higher Education Funding Council for England(HEFCE). D.J.P. acknowledges UoW for a postgraduatescholarship. Philip Aston and Dr. Lijiang Song are thankedfor their assistance with mass spectrometry analysis.

■ REFERENCES(1) Stuart, M. A. C.; Huck, W. T. S.; Genzer, J.; Muller, M.; Ober, C.;Stamm, M.; Sukhorukov, G. B.; Szleifer, I.; Tsukruk, V. V.; Urban, M.;Winnik, F.; Zauscher, S.; Luzinov, I.; Minko, S. Nat. Mater. 2010, 9(2), 101−113.(2) Priya, B.; Viness, P.; Yahya, E. C.; Lisa, C. d. T. Biomed. Mater.2009, 4 (2), 022001.(3) Shim, M. S.; Kwon, Y. J. Adv. Drug Delivery Rev. 2012, 64, 1046−1059.(4) Chan, G.; Mooney, D. J. Trends Biotechnol. 2008, 26 (7), 382−392.(5) Akimoto, J.; Nakayama, M.; Sakai, K.; Okano, T. Mol. Pharm.2010, 7 (4), 926−935.

Figure 7. Isothermal turbidimetry data for ss-pNIPAM-2 (polymerconcentration = 5 mg·mL−1). Temperature = 37 °C, GSH added attime indicated by asterisk. Black = control (polymer and water); red =1 μM GSH; blue = 1 mM GSH. Inset = photo at end of experiment(left = control; center = 1 μM GSH; right = 1 mM GSH).



http://pubs.acs.org

mailto:[email protected]

(6) Saaka, Y.; Deller, R. C.; Rodger, A.; Gibson, M. I. Macromol.Rapid Commun. 2012, 33 (9), 779−784.(7) Chilkoti, A.; Dreher, M. R.; Meyer, D. E.; Raucher, D. Adv. DrugDelivery Rev. 2002, 54 (5), 613−630.(8) Lutz, J.-F. J. Polym. Sci., Part A: Polym. Chem. 2008, 46 (11),3459−3470.(9) Weber, C.; Hoogenboom, R.; Schubert, U. S. Prog. Polym. Sci.2012, 37 (5), 686−714.(10) Schild, H. G. Prog. Polym. Sci. 1992, 17 (2), 163−249.(11) Ieong, N. S.; Redhead, M.; Bosquillon, C.; Alexander, C.;Kelland, M.; O’Reilly, R. K. Macromolecules 2011, 44 (4), 886−893.(12) Heath, F.; Saeed, A. O.; Pennadam, S. S.; Thurecht, K. J.;Alexander, C. Polym. Chem. 2010, 1 (8), 1252−1262.(13) Magnusson, J. P.; Khan, A.; Pasparakis, G.; Saeed, A. O.; Wang,W.; Alexander, C. J. Am. Chem. Soc. 2008, 130 (33), 10852−10853.(14) Steinhauer, W.; Hoogenboom, R.; Keul, H.; Moeller, M.Macromolecules 2010, 43 (17), 7041−7047.(15) Shepherd, J.; Sarker, P.; Swindells, K.; Douglas, I.; MacNeil, S.;Swanson, L.; Rimmer, S. J. Am. Chem. Soc. 2010, 132, 1736−1737.(16) Summers, M. J.; Phillips, D. J.; Gibson, M. I. Chem. Commun.2012, DOI: 10.1039/C2CC34236G.(17) Fox, M. E.; Szoka, F. C.; Frechet, J. M. J. Acc. Chem. Res. 2009,42 (8), 1141−1151.(18) Han, S.-o.; Mahato, R. I.; Sung, Y. K.; Kim, S. W. Mol. Ther.2000, 2 (4), 302−317.(19) Ulery, B. D.; Nair, L. S.; Laurencin, C. T. J. Polym. Sci., Part B:Polym. Phys. 2011, 49 (12), 832−864.(20) Okada, M. Prog. Polym. Sci. 2002, 27 (1), 87−133.(21) Schafer, F. Q.; Buettner, G. R. Free Radical Biol. Med. 2001, 30(11), 1191−1212.(22) Albertsson, A. C.; Karlsson, S. Acta Polym. 1995, 46 (2), 114−123.(23) Pounder, R. J.; Dove, A. P. Polym. Chem. 2010, 1 (3), 260−271.(24) Gibson, M. I.; Cameron, N. R. J. Polym. Sci., Part A: Polym.Chem. 2009, 47 (11), 2882−2891.(25) Kim, T.-i.; Kim, S. W. React. Funct. Polym. 2011, 71 (3), 344−349.(26) Moad, G.; Rizzardo, E.; Thang, S. H. Acc. Chem. Res. 2008, 41(9), 1133−1142.(27) Phillips, D. J.; Gibson, M. I. Chem. Commun. 2012, 48 (7),1054−1056.(28) Zhang, Q.; Noh, S. M.; Nam, J. H.; Jung, H. W.; Park, J. M.; Oh,J. K. Macromol. Rapid Commun. 2012, DOI: 10.1002/marc.201200289.(29) Anderson, M. E. Chem.-Biol. Interact. 1998, 111−112, 1−14.(30) Jones, D. P.; Carlson, J. L.; Samiec, P. S.; Sternberg, P., Jr;Mody, V. C., Jr; Reed, R. L.; Brown, L. A. S. Clin. Chim. Acta 1998,275 (2), 175−184.(31) Ou, M.; Xu, R.; Kim, S. H.; Bull, D. A.; Kim, S. W. Biomaterials2009, 30 (29), 5804−5814.(32) Oupicky, D.; Parker, A. L.; Seymour, L. W. J. Am. Chem. Soc.2001, 124 (1), 8−9.(33) Vader, P.; van der Aa, L.; Engbersen, J.; Storm, G.; Schiffelers, R.Pharm. Res. 2012, 29 (2), 352−361.(34) Cho, H. Y.; Srinivasan, A.; Hong, J.; Hsu, E.; Liu, S.; Shrivats, A.;Kwak, D.; Bohaty, A. K.; Paik, H.-j.; Hollinger, J. O.; Matyjaszewski, K.Biomacromolecules 2011, 12 (10), 3478−3486.(35) Cai, X.; Dong, C.; Dong, H.; Wang, G.; Pauletti, G. M.; Pan, X.;Wen, H.; Mehl, I.; Li, Y.; Shi, D. Biomacromolecules 2012, 13 (4),1024−1034.(36) Li, S.; Wang, Y.; Wang, S.; Zhang, J.; Wu, S.-F.; Wang, B.-L.;Zhu, W.; Yu, X.-Q. Bioorg. Med. Chem. 2012, 20 (4), 1380−1387.(37) Cheng, R.; Feng, F.; Meng, F.; Deng, C.; Feijen, J.; Zhong, Z. J.Controlled Release 2011, 152 (1), 2−12.(38) Ryu, J.-H.; Roy, R.; Ventura, J.; Thayumanavan, S. Langmuir2010, 26 (10), 7086−7092.(39) Ren, T.-B.; Feng, Y.; Zhang, Z.-H.; Li, L.; Li, Y.-Y. Soft Matter2011, 7 (6), 2329−2331.(40) Wang, Y.-C.; Li, Y.; Sun, T.-M.; Xiong, M.-H.; Wu, J.; Yang, Y.-Y.; Wang, J. Macromol. Rapid Commun. 2010, 31 (13), 1201−1206.

(41) Koo, A. N.; Lee, H. J.; Kim, S. E.; Chang, J. H.; Park, C.; Kim,C.; Park, J. H.; Lee, S. C. Chem. Commun. 2008, 48, 6570−6572.(42) Li, Y.-L.; Zhu, L.; Liu, Z.; Cheng, R.; Meng, F.; Cui, J.-H.; Ji, S.-J.; Zhong, Z. Angew. Chem., Int. Ed. 2009, 48 (52), 9914−9918.(43) Jiang, X.; Li, L.; Liu, J.; Hennink, W. E.; Zhuo, R. Macromol.Biosci. 2012, 12 (5), 703−711.(44) Ryu, J.-H.; Park, S.; Kim, B.; Klaikherd, A.; Russell, T. P.;Thayumanavan, S. J. Am. Chem. Soc. 2009, 131 (29), 9870−9871.(45) Skey, J.; O’Reilly, R. K. Chem. Commun. 2008, 35, 4183−4185.(46) Luzon, M.; Boyer, C.; Peinado, C.; Corrales, T.; Whittaker, M.;Tao, L.; Davis, T. P. J. Polym. Sci., Part A: Polym. Chem. 2010, 48 (13),2783−2792.(47) Xu, J.; He, J.; Fan, D.; Wang, X.; Yang, Y. Macromolecules 2006,39 (25), 8616−8624.(48) Roth, P. J.; Boyer, C.; Lowe, A. B.; Davis, T. P. Macromol. RapidCommun. 2011, 32 (15), 1123−1143.(49) Ieong, N. S.; Hassan, M.; Phillips, D. J.; Saaka, Y.; O’Reilly, R.K.; Gibson, M. I. Polym. Chem. 2011, 3, 794−799.



biodegradable poly(disulfide)s derived from raft polymerization: monomer scope, glutathione...

Documents