可交联且可生物降解的高分子膜

陈帅 ，秦江雷 ，杜建忠 ,＊

1同济大学附属上海市第十人民医院骨科，同济大学医学院，上海 200072

2同济大学材料科学与工程学院高分子材料系，上海 201804

3河北大学化学与环境科学学院，河北 保定 071002

4北京化工大学有机无机复合材料国家重点实验室，北京 100029

1 Introduction

Polymer films are versatile materials that can be used in tissue engineering, artificial organs, and biomedical devices1,2. They can also play very important roles in electronic devices, food packaging, and gas separation technologies3–5. Unfortunately,natural polymer films usually lack chemical diversity and are difficult to process6,7. Synthetic polymer films can be obtained from polymers with complex architecture such as polymer brushes, which are widely used as anti-fouling surfaces8–10.Although synthetic polymers are easily prepared, and suitable for processing, they can be potentially toxic11,12. Besides,synthetic polymers are generally not biodegradable; however, a low molecular weight supports this desired characteristic.Additionally, this can be achieved by incorporating biodegradable bonds (ester bond, peptide bond, disulfide bond,etc.)13. A common example of biodegradable synthetic polymers is polycaprolactone (PCL). This polymer is biocompatible and biodegradable, and is therefore widely used as tissue engineering scaffolds14, drug delivery vehicles15,16, and biomedical films17,18.Unfortunately,in vivodegradation of PCL lasts several days, or even months19,20. A faster and adjustable polymer degradability would allow the preparation of more diversified materials.Therefore, the effective regulation of the degradability of PCL-based materials is an important challenge.

Polymer materials can be cross-linked by various methods21.For example, poly(2-cinnamoyloxyethyl methacrylate)(PCEMA) can be photocross-linked under ultraviolet (UV)light22, and the same approach applies to coumarin-incorporated polymers23–25. Other well-known examples include polybutadiene, which is cross-linked by radical polymerization26,and poly(3-(trimethoxysilyl)propyl methacrylate) or poly(3-(triethoxysilyl)propyl methacrylate) that can be cross-linked by hydrolysis27,28. The undegradable nature of these cross-linked polymer materials prevents disassociation29. Cross-linking polymer films can effectively increase their strength and stability. This makes them abrasion-resistant and resilient to internal stresses3,6. However, cross-linking usually leads to a reduction in biodegradability and structural flexibility30,31.These disadvantages restrict usage of these materials in biomedical applications32.

A variety of cross-linking strategies were developed for the preparation of biodegradable (nano)materials. For example,Matyjaszewski and coworkers33,34prepared degradable corecross-linked star polymers and gels. These materials were prepared from disulfide-containing initiators and disulfide crosslinkers. Furthermore, they studied their redox-responsive degradation behaviors. Jianget al.35synthesized stepwise cleavable star polymers and polymeric gels by atom transfer radical polymerization and atom transfer radical coupling. The degradation of these two materials is based on redoxresponsiveness of disulfide bond and base-catalyzed hydrolysis of ester bond. Qiao and Wiltshire36,37reported a variety of degradable core-cross-linked star polymers by polymerizing bifunctional 4,4’-bioxepanyl-7,7’-dione and 2,2-bis(ɛcaprolactone-4-yl) propane as cores. The polymers have similar structures to PCL, thus the cores of the star polymer can be degraded under acidic conditions. Denget al.38prepared a selfhealing hydrogel. The introduction of dynamic acylhydrazone linkages allowed a reversible sol-gel transition. This innovation attracted great interest in reversible cross-linking strategies39,40.Other degradable gels with cross-linked structures were also reported41,42. These examples utilize an uncontrolled crosslinking reaction that occurred simultaneously with the polymerization (or coupling reaction). Therefore, this characteristic limits the strategies to the formation of core-crosslinked star polymers and gels. And the strategies are not suitable for the preparation of polymer films, which require postpolymerization cross-linking to guarantee the morphology.

Various PCL cross-linking strategies have been reported. Most literature examples describe un-controlled cross-linking strategies43,44. However, functionalization of PCLviaclick chemistry was reported45. The same strategy can be used for the preparation of PCL-based polymer brushes46. Based on these methods, cross-linkable groups can be introduced to achieve controlled cross-linking of PCL. As a typical photocrosslinkable functional group, cinnamate can be cross-linked using UV irradiation, which causes a [2 + 2] cycloaddition47,48. Besides,cinnamate-derived compounds are generally nontoxic49and therefore ideal for usage in biomedical applications. Our previous report demonstrated that UV irradiation can effectively cross-link dispersed cinnamate-containing polymersomes50.These cross-linked membranes can be degraded in an acidic environment, or with the usage of lipase. Moreover, the introduction of side groups attenuates the crystallization of PCL backbone. This makes the polymersome more susceptible to degradation. Varying the degree of cross-linking allows fine control over the degradability of membranes. This characteristic is also potentially applicable to polymer films.

In this paper, we fabricated two cross-linkable, yet biodegradable, polymer films which were prepared from poly[α-(cinnamoyloxymethyl)-1,2,3-triazol caprolactone] (PCTCL133)and the copolymer of P(CL156-stat-CTCL28) (Scheme 1). UV irradiation on the polymer films efficiently cross-linked them and makes them undissolvable in tetrahydrofuran (THF, a good solvent for the un-cross-linked polymer film). Despite the robust cross-linked structure, the network can be completely degraded by acid-catalyzed degradation of ester linkages within PCL backbones and the side groups. The degradable characteristics,and the film transparency, can be conveniently adjusted by copolymerization with caprolactone, which results in different crystallinity of the polymer. In addition, the degradation rate of the polymer films can also be adjusted by varying the crosslinking density. This strategy is based on the following principles: primarily, the densely-grafted side groups in PCTCL impede the crystallization of the PCL backbone. This leads to an upper limit for the biodegradation rate. Additionally, an increased fraction of copolymerized CL lowers the biodegradation rate, and the transparency of the film. The weight percentage of the PCL segments in the polymer film, and the photocross-linking density, can be accurately tuned to control the biodegradable character and the stability of this material. In this way, these degradable PCTCL-based films show great potential for biomedical applications such as drug delivery, tissue engineering, or artificial organs.

Scheme 1 Schematic illustration for photocross-linkable yet biodegradable polymer films made of (a) PCTCL133 and (b) P(CL156-stat-CTCL28).Films were made from both polymers by solution casting, and they can be photocross-linked to increase their stability and mechanical strength.The copolymer film appeared translucent, while the homopolymer one was transparent because of its impeded crystallinity. Both polymer films were able to fully degrade under acidic conditions.

2 Results and discussion

2.1 Syntheses and Characterizations of PCTCL133 and P(CL156-stat-CTCL28)

A PCTCL133homopolymer was prepared according to a previously reported protocol (see Scheme S1 (Supporting Information (SI)))50. Briefly,α-chloro-caprolactone (αClεCL)monomer was preparedviaBaeyer-Villiger oxidation ofαchloro-cyclohexanone45,51. Then the ring-opening polymerization (ROP) ofαClεCL was initiated by isopropanol for the preparation of P(αClCL)133. This polymer was allowed to react with sodium azide to afford P(N3CL)133. (The1H NMR spectra ofαClεCL, P(αClCL)133and P(N3CL)133are shown in Fig. S1 (SI).) After azide substitution, the peak at 4.29 ppm,representing the proton adjacent to carbonyl groups on P(αClCL), shifted to 3.85 ppm. This indicates that all chlorine atoms were substituted by azide groups. These groups were then reacted with propargyl cinnamate to afford the final homopolymer PCTCL133using click chemistry45. (Propargyl cinnamate was synthesized from propargyl alcohol and cinnamoyl chloride.) The1H NMR spectrum of propargyl cinnamate is shown in Fig. 1c. By comparing the peak area of m(4.85 ppm) derived from propargyl alcohol and n (6.51 ppm)derived from cinnamoyl chloride, we can conclude that the product was of high purity.

The chemical composition of the final PCTCL133homopolymer, was confirmed by1H NMR spectroscopy. The1H NMR spectra of PCTCL133 and its precursors are shown in Fig.1. After click reaction with propargyl cinnamate, the proton a’ (t,OCH2) on P(N3CL) shifted from 4.20 to 4.08 ppm (a’’) (see Fig.1a–b). Proton f’ (t, CHN3CO) shifted from 3.85 to 5.36 ppm (f’’).The complete shift indicated a high efficiency of the click reaction. Additionally, the introduction of the propargyl cinnamate to PCTCL caused the proton m (s, OCH) to shift from 4.85 to 5.28 ppm (m’), and the proton k (s, CCH) shifted from 2.53 to 7.88 ppm (k’) (Fig. 1a and c). These1H NMR analyses suggested that the final PCTCL133was well-defined, and that the synthetic procedures were well-controlled.

Fig. 1 1H NMR spectra of (a) PCTCL133, (b) P(N3CL)133 and(c) propargyl cinnamate in CDCl3. Herein δ represents chemical shift of protons.

The statistical copolymer, P(CL156-stat-CTCL28), was synthesized according to the same protocol as was used for the PCTCL133synthesis, as described above (see Scheme S1). The1H NMR spectra of P[CL156-stat-(N3CL)28] and P(CL156-stat-CTCL28) are shown in Fig. 2. Functionalization caused a change in the chemical shift of the protons adjacent to oxygen (h) from 4.20 to 4.05 ppm (Fig. 2b), causing these protons to overlap with the existing peak of this chemical shift (a). Similar to PCTCL133,the P[CL156-stat-(N3CL)28] became completely functionalized and formed P(CL156-stat-CTCL28), as shown in Fig. 2a (also see the magnification of the spectrum in Fig. 2a). The degree of polymerization for CTCL monomers was calculated from this analysis and was 28. After functionalization, the molecular weight ratio of PCL segment in the copolymer decreased from 80.4% to 64.9% and the PCTCL segment occupied 35.1% of the total molecular weight in P(CL156-stat-CTCL28).

Fig. 2 1H NMR spectra of (a) P(CL156-stat-CTCL28) and(b) P[CL156-stat-(N3CL)28] in CDCl3.

SEC analysis of P(αClCL)133and P(N3CL) homopolymers showed similar number-average molecular weights (Mn) and dispersities (Ð), 19700 (Ð= 1.30) and 20300 (Ð= 1.31),respectively (see Fig. S2 (SI)). Compared to the precursors, the molecular weight of the functionalized PCTCL133significantly increased, while maintaining a relatively low dispersity (Mn =45400,Ð= 1.36). This was because the overall molecular weight of PCTCL133is 120% larger than the polymer precursor,P(N3CL)133. These SEC results also confirm that the PCL chains are stable during substitution, and that the ester bonds do not react with azide ions. Similar results were obtained for P(CL156-stat-CTCL28). Details of the final homopolymer and statistical copolymer are shown in Table 1.

Table 1 Characterizations of PCTCL133 and P(CL156-stat-CL28).

The functionalization of the homopolymer and the statistical copolymer was also characterized using FT-IR spectroscopy (see Fig. S3 (SI)). Propargyl cinnamate exhibited an absorbance at 1632 cm−1which corresponds to the double bond that is conjugated with the carbonyl group. The sharp peak at 2127 cm−1was ascribed to the alkynyl group. The azide side group P(N3CL) exhibited a strong absorbance at 2113 cm−1(see Fig.S3b)52,53. The absorbance of the carbonyl group in this polymer was 35 cm−1higher than that of propargyl cinnamate (1747 cm−1in comparison to 1712 cm−1, respectively). This is because the carbonyl group on propargyl cinnamate is conjugated with double bond and the benzene ring. After the click reaction, both the absorbance of the azide group on P(N3CL) and the propargyl group on propargyl cinnamate disappeared. As expected, a peak at 1632 cm−1remained. This peak corresponds to the double bond introduced from propargyl cinnamate to PCTCL (see Fig.S3c). This indicated that the efficiency of the click reaction was high and all azide groups in P(N3CL) were consumed. However,the absorbance of carbonyl groups was the same after click reaction because the reaction did not change the chemical environment of carbonyl groups from propargyl cinnamate.

Differential scanning calorimetry (DSC) was used to assess the thermal properties of the final PCTCL133and P(CL156-stat-CTCL28)50. P(CL156-stat-CTCL28) exhibits a melting point of 59.2 °C. PCTCL133is an amorphous polymer with a glass transition temperature (Tg) of 53.7 °C under a heating rate of 10 °C·min−1. In contrast to crystallizable PCL, the amorphous PCTCL with a moderateTgis desired for further processing of the polymer.

In summary, well-defined PCTCL133and P(CL156-stat-CTCL28) were synthesized and characterized. In the next section,films of these polymer were subsequently prepared and successfully cross-linked using UV irradiation.

2.2 Preparation and degradation of photocrosslinked polymer films

Polymer films were prepared from PCTCL133homopolymer and P(CL156-stat-CTCL28) by solution casting (see Scheme S2(SI)). First, a polymer solution (50 mg·mL−1) was prepared from either the homopolymer or the statistical copolymer in THF. This solution was spread onto a Teflon plate, and was allowed to evaporate until a polymer film of ～0.2 mm was formed.Subsequently, residual organic solvent was removed from the polymer filmin vacuoat 40 °C for 12 h. The obtained PCTCL133film appeared transparent, which is likely because of its amorphous character; and the P(CL156-stat-CTCL28) film appeared translucent, which is likely because of the crystalline PCL segments (see Fig. 3a–b). Finally, both polymer films were cross-linked. Irradiation from an 800 W UV lamp caused the cinnamate groups to react with each other. After 1 h of irradiation, the films were immersed in THF and shaken periodically to investigate if they became insoluble. In contrast to the un-cross-linked films, which dissolved after several minutes, the cross-linked films remained undissolved for more than 24 h (Fig. 3c–d). This indicated that both cross-linked films formed stable polymer networks. Furthermore, the shapes of the polymer films were perceived, indicating that the chemical networks were pretty tough. Additionally, both cross-linked films remained relatively unswollen, indicating a high crosslinking density.

Fig. 3 Photographs of cross-linked polymer films of (a) PCTCL133,(b) P(CL156-stat-CTCL28) and (c and d) the same polymer films immersed in THF for 24 h.

These observations from the cross-linked films suggest that cinnamate can efficiently be used as a photocross-linkable group. However, it was noticed that the volume of the P(CL156-stat-CTCL28) film reduced and that the color of the solvent turned light brown, as shown in Fig. 3d. SEC analysis confirmed that a significant amount of dissolved P(CL156-stat-CTCL28)copolymer was present in the THF solution (see Fig. S4 (SI)). In contrast, no dissolved homopolymer was observed in the corresponding solvent phase of the PCTCL133film in THF. This indicates that this homopolymer film was completely crosslinked. The lower degree of cross-linking of the P(CL156-stat-CTCL28) film likely relates to the crystallizable PCL segments.These segments were responsible for the translucent character of the copolymer film, resulting in less efficient UV-cross-linking reactions in the bulk of this dense film. This suggests that ultrathin P(CL156-stat-CTCL28) films, or nano-objects (such as polymersomes and micelles), can be more efficiently crosslinked than the assessed film which had a thickness of 0.2 mm.Furthermore, these experiments suggest that the cross-linking efficiency of dense films (or bulk materials) can be increased by increasing the fraction of PCTCL in this copolymer.

The degradable character of the cross-linked and un-crosslinked PCTCL133film was assessed. Both films were submerged in a 0.12 mol·L−1HCl/THF solution. The un-cross-linked film degraded within 2 h, and the cross-linked film degraded in 36 h.In this way, the degradation rate dropped by more than 94.4% after 1 h of UV irradiation. SEC analysis further confirmed the required degradation time of the PCTCL133 films, since no polymer peaks were visible at the original retention time of this polymer50. The appearance of low molecular weight peaks in the SEC chromatogram indicated that the degradation products were small molecules. Another indication regarding the size of the degraded material was that it could pass through a 0.22 μm filter without any resistance. These two aspects suggest that the cross-linked PCTCL film was degraded. The degradation rate of the cross-linked PCTCL film under acidic conditions is significantly decreased after photocross-linking. It is likely that this is related to the high cross-linking density of this material, which reduces the permeability of this film. Degradation likely occurs gradually from the surface to the center of this material. SEC analysis of the degraded polymer solution indicated solely the presence of small molecules. These data indicate that the PCTCL133film remained completely degradable, even after photocross-linking.A possible chemical structure of the cross-linked polymeric network and its small molecule degradation products are shown in Scheme S3 (SI).

TGA analysis of the un-cross-linked and cross-linked PCTCL133films indicated a similar thermal degradation profile(see Fig. S5 (SI)), despite the large differences observed in the previously-discussed acidic degradation experiments. The TGA data seems to imply that thermal degradation occurs as a twostage process with onset temperatures atca.350 and 520 °C respectively.

Finally, the cross-linking density of the homopolymer film was determined. This was assessed by determining the conversion of the double bond of the cinnamate group by1H NMR analysis, the signal of which appears at 6.45 ppm. More specifically, the degradation products from the acidic degradation of the cross-linked PCTCL133was examined and compared to un-cross-linked sample to evaluate the relative drop of this proton signal, which indicated 43% conversion of the cinnamate group after 1 h of UV irradiation, and 49% conversion after 2 h of irradiation. Here it is important to point out that this reduction in peak area arises from both intermolecular and intramolecular cycloaddition reactions. Fig. 4 shows the1H NMR spectrum of the degraded film that was cross-linked by UV irradiation for 2 h. However, the cinnamate functional group cannot be completely cross-linked since the [2 + 2] cycloaddition of cinnamamide, or cinnamate, is reversible48. Despite this limitation, the cinnamate conversion is sufficiently high for the preparation of a variety of cross-linked materials. Additionally,the emerging signal of the carboxylic acid protons (9.94 ppm) is also a strong evidence for catalyzed degradation. Eventually, the films are able to degrade to products with molecular weights no more than 300 g·mol−1(see Scheme S3 for the chemical structures of the most likely degradation products).

Fig. 4 1H NMR spectrum of the acid-catalyzed degradation product from the PCTCL133 film which was exposed for 2 h under UV irradiation.

We have previously demonstrated the enzymatic and acidic biodegradation of membrane-cross-linked PCTCL-based polymersomes in water50. These studies imply that the herein reported PCTCL-based films are also biodegradable.Additionally, the degradation rate of the polymer film can be effectively tuned by varying the cross-linking density.Furthermore, film transparency and other properties can be altered by copolymerizing various amounts of caprolactone.These aspects can be used as handles to tune the characteristics of the polymer film, hence guaranteeing the diversity of PCTCL-based films for a broad range of applications.

Mechanical strength is an important factor regarding the applicability of polymer films. Herein, we used molecular dynamics (MD) simulations to evaluate the influence of the polymer concentration of the initial solution where the films were casted from. Polymer concentration dictates chain entanglement. A low polymer concentration supports a coiled polymer chain arrangement (with predominantly intramolecular interactions), whereas an increased amount of polymer leads to chain entanglements (or intermolecular interactions). More entanglements should lead to an enhanced mechanical strength54.

The actual MD simulations were performed on different amorphous cells each consisting of 5 PCTCL10with various initial distances between the chains (corresponding to the packing density/concentration of the initial polymer solutions,see Fig. S6 (SI)). First, the average total energy in each system was obtained based on six possible cells corresponding to each density (see Fig. S7 (SI)). As expected, the total energy significantly increased with the polymer packing density/concentration; except for the samples with densities less than 0.1 g·cm−3. (These samples likely correspond to the lower concentration limit where intermolecular interactions can be neglected.) Secondly, the preparation of polymer films was simulated with stepwise dynamics simulations based on the ensembles ofNVT(constant volume, constant temperature) andNPT(constant pressure, constant temperature), respectively.This simulates the progressively increased polymer packing as a result of solvent evaporation. These simulations were conducted at room temperature (298 K). With an additionalNPTdynamics simulations at 101.3 kPa, the final densities of all evaluated cells reached an average value of 1.16 ± 0.01 g·cm−3. Thirdly, the Young’s moduli of the cells in their final state was investigated(see Fig. S8 (SI)). As expected, the average value of the Young’s moduli increased with the initial density of each amorphous cell that was used in dynamics simulations. This suggests that tougher films will be produced by solution casting when increasing the initial polymer concentration. Because the actual size and total number of chains in practice are much larger than in our simulations, the influence of initial concentration on the mechanical strength of the final films can be more pronounced.Indeed, this initial polymer concentration for solution casting can be used as another handle to conveniently tune the physical characteristics of the polymer film.

3 Conclusions

In summary, we prepared two photocross-linkable, yet biodegradable, polymer films from PCTCL133and P(CL156-stat-CTCL28). The main results are concluded as follows: (1) the polymers of PCTCL133 and P(CL156-stat-CTCL28) were synthesized and the chemical structures were confirmed; (2)polymer films were prepared from these two polymersviasolution casting. Cross-linking of these polymer films was achieved by photocross-linking of the cinnamate groups. The homopolymer film appeared transparent and the copolymer film was translucent, which is likely because of the impediment of its crystalline structure; (3) The cross-linked polymer films can be completely degraded with acid. Through control over the crosslinking density or by copolymerization with PCL, the degradation rate of the films can be altered. Therefore, these polymer films can be easily modified to meet the specific demands for future applications. Additionally, the crosslinkable, yet biodegradable, structure of these films could be extended for the preparation of hydrogels and functional coatings. Overall, the easily adjustable structure makes these PCTCL-based films promising for a range of (biomedical)applications.

Supporting Information:Materials, characterization,experimental procedures, additional schemes and figures (1H NMR spectra, SEC curves, FT-IR spectra, TGA studies) are available free of chargeviathe internet at http://www.whxb.pku.edu.cn.