Introduction

Numerous environmental problems associated with synthetic polymers, including greenhouse gas emissions and marine microplastics [1,2,3,4], can potentially be solved by shifting to recyclable and sustainable alternatives [5,6,7]. Consequently, considerable attention has been given to the replacement of nondegradable functional polymers with degradable polymers and the further development of the latter, as exemplified by the use of aliphatic polyesters (e.g., polylactides and poly(hydroxyalkanoate)s) for single-use and short-term applications [8,9,10,11]. In particular, functional degradable polymers with tunable degradability, i.e., controllable degradation rate and timing, have been actively explored.

Several degradable polymers with functional side chains have been developed for biomedical devices used in nano- and regenerative medicine [12,13,14]. In the last two decades, side-chain-functionalized aliphatic polycarbonates (APCs) have emerged as a widespread platform for realizing functional degradable biomaterials [15,16,17,18]. Inspired by the commercially available blood-compatible polymer poly(2-methoxyethyl acrylate) (PMEA) and its analogs [19,20,21,22], we developed ether-functionalized APCs for potential use as stent coatings, surface layers of resorbable artificial blood vessels, and other components of blood-compatible biomedical devices [23,24,25,26]. The ether side groups enhance hydration, helping to control the adsorption and denaturation of serum proteins and thus inhibiting platelet adhesion [27, 28]. This mechanism was also exemplified by a recently reported poly(p-dioxanone) derivative [29]. These studies yield important insights into the hydrolytic degradation of polymers, suggesting that hydration due to interactions with ether side groups facilitates the hydrolysis of polymer backbones composed of aliphatic ester and carbonate moieties [21, 29].

Poly(trimethylene carbonate) (PTMC) is a representative biodegradable APC widely used in commercialized biomaterials such as sutures [30] yet fairly stable for nonenzymatic hydrolysis [31]. Unlike its analogs that include oligo(ethylene glycol) moieties, PTMC is minimally degraded in alkaline solutions, suggesting that the degradability of the APC main chain is affected by the appended side chains [32]. The introduction of appropriate functional groups and structures renders APCs suitable for the fabrication of sustainable functional materials [33,34,35]. Accordingly, the hydrolyzability of such APCs in the absence of enzymes (or at low enzyme levels) is important for estimating the environmental impact of their disposal. Furthermore, the correlation of hydration properties with hydrolyzability may enable the control of the lifetime and stability of APC in aqueous environments and thus expand the application scope of these polymers.

Considering the above discussion, we investigated the relationship between the hydration and hydrolysis of APCs more deeply and enhanced these properties through structural optimization (Fig. 1) to help advance the controllable degradation of aliphatic polyesters and APCs. The extent of hydration induced by ether side groups varies depending on the carbon–oxygen ratio and the configuration of ether oxygen (e.g., distance from the main chain), as demonstrated by PMEA analogs [20, 22]. However, a similar systematic study has never been performed on APCs because of the difficulty in synthesizing analogs of equal quality and molecular weight. The investigated APCs were derived from 2,2-bis(hydroxymethyl)propionic acid (bis-MPA), as the resulting scaffolds can be efficiently grafted with side-chain functionalities using the postpolymerization protocol developed by Engler et al. [36]. This approach enables the preparation of APCs with different types of ether side groups at the same chain length, thus avoiding tedious individual monomer synthesis. Six bis-MPA-based APCs with ether group–containing side chains linked to the main chain by an amide group (P1–P6) were synthesized to elucidate the effects of hydration on degradability and biological functions (Fig. 2). Moreover, we investigated the effects of the amide linker on hydration by comparison with those of a previously reported ether-functionalized APC bearing an ester linker (P7) [23], an APC with a methoxy side chain (P8) [25], and PTMC (P9).

Fig. 1
figure 1

Schematics of (A) hydration and (B) hydrolysis of the employed aliphatic polycarbonates (APCs)

Full size image
Fig. 2
figure 2

Molecular structures of the employed APCs (P1–P9). PTMC = poly(trimethylene carbonate)

Full size image

Results and discussion

APC synthesis and characterization

The triflic or methanesulfonic acid–catalyzed ring-opening polymerization of M1 yielded a precursor polymer (P0) with a sufficiently high number-average molecular weight (Mn = 18.3 × 103 g mol–1) and low molar-mass dispersity (ÐM = 1.14), and subsequent side-chain (postpolymerization) amidation yielded P1–P6 (Scheme 1) [36]. Figure 3 shows the 1H nuclear magnetic resonance (NMR) spectrum of P2 as a representative example. All amidation reactions were shown to be quantitative by 1H NMR spectroscopy (Figs. S1–S6), although the reaction completion time depended on the side-chain structure, equaling 1 h for P1–P3 and 3 h for P4–P6, possibly because of steric effects. This steric effect is associated with the nucleophilicity of the amine and the accessibility of the amine to unreacted sites in the polymer. The former is applicable to P4 formation, where the amino group is sterically hindered by a tetrahydrofurfuryl group. A similar steric effect was found in our previous study to examine esterification between carboxylate anion and tetrahydrofurfuryl bromide [26]. The latter effect is explained by the formation of P5 and P6, where the side groups attached to bulky gallate derivatives hinder the subsequent access of the amines to the neighboring pentafluorophenyl ester side groups of P0. These two mechanisms may induce the longer reaction times associated with the corresponding processes. The ÐM values of P1–P6 were not significantly different from that of P0; i.e., amidation did not induce main-chain decomposition (Table 1 and Fig. S7).

Scheme 1
scheme 1

Syntheses of precursor polymers (P0) and ether-functionalized APCs (P1–P6)

Full size image
Fig. 3
figure 3

1H nuclear magnetic resonance (NMR) spectrum (400 MHz) of P2 in deuterated chloroform (CDCl3; internal standard = tetramethylsilane (TMS))

Full size image
Table 1 Characteristics of ether-functionalized APCs
Full size table

P1–P6 are amorphous substances with side-chain-dependent glass transition temperatures (Tg, 3–56 °C). The relatively high Tg values of P4–P6 ( > 40 °C) were attributed to their bulky and rigid side chains (Table 1). The Tg of amide-linked P1 was 40 °C greater than that of its ester-linked analog (P7), which was attributed to the amide group engaging in interchain hydrogen bonding.

Hydration properties of APCs

The hydration properties of the APCs were evaluated by differential scanning calorimetry (DSC) analysis and water contact angle measurements [27, 28]. Static water contact angles were measured for spin-coated samples using sessile-drop (θw) and captive-bubble (θair) methods. The θw of P1 (53°) was 10° lower than that of P7 (63°) (Table 1); i.e., the surface of the former polymer was more hydrophilic. For P1–P6, θw was closely correlated with the side-chain structure and increased with the number of carbon atoms and the carbon-to-oxygen ratio. The differences among the θair values of P2–P6 were minor compared to those among the corresponding θw values. P2–P6 had higher θair values (i.e., more hydrophilic surfaces in water) than did P7–P9. The above results reflect the effects of the amide linker on hydration. Considering that in the ideal scenario, θw = 180°– θair [37], the difference between the measured θw and the value calculated as 180°– θair (= θw-calc.) represents deviation from the theoretical value for hydrophilicity of a given surface in air or water. In fact, the θair values of P1 and P2 are lower than expected, while the θw values of P5 and P6 are higher than expected. These results are likely related to the mobility of the amide-linked side-chain ether groups in contact with water. In particular, P1 and P2 may engage in distinctive interactions with water because of the configuration of their amide and methoxy groups, which restricts side-chain motion in water. Based on the difference between P1 and P7, the hydrogen-bond donor function of the amide N–H significantly contributes to the enhanced hydration of P1 and P2, as does the methoxy group (Fig. 4A, D). Although the amide C = O also interacts with water, this interaction is omitted from the discussion on the effects of the amide N–H. Furthermore, the methoxy oxygen serves as a hydrogen-bond acceptor, which may lock the water molecule bound to the amide N–H (Fig. 4B, E). These mechanisms motivate the proposed intramolecular hydrogen-bonding modes of the side chains of P1 and P2 (Fig. 4C, F), which may be associated with their glass transition and response upon contact with water. The side chain of P1 should theoretically be more hydrophilic than that of P2 based on the carbon-to-oxygen ratio. However, because the θw value of P2 was lower than that of P1 (Table 1), we suspect that the populations of the double (Fig. 4B, E) and intramolecular (Fig. 4C, F) hydrogen-bonding modes for P1 are greater than those for P2, which decreases the wettability and hydration.

Fig. 4
figure 4

Schematics of possible hydrogen-bonding modes of the side chains of P1 (A–C) and P2 (D–F). A, D Amide N–H and methoxy oxygen interact with water separately. B, E The amide N–H and methoxy oxygen lock a water molecule. C, F Intramolecular hydrogen bonding in the side chain

Full size image

DSC analysis of the hydrated polymers provided further insights into the polymer–water interactions. DSC peaks corresponding to the fusion and crystallization of water can appear at temperatures below 0 °C, depending on the above interactions, e.g., strong polymer–water interactions retard the crystallization of water, similar to cryoscopy. Figure 5 shows the DSC thermograms of hydrated APCs (P1–P3, P7) with near-equilibrium water contents. Hydrated P4–P6 could be prepared from only the overhydrated state, and the corresponding DSC profiles are shown in Figure S8. The equilibrium water contents of P1–P3 (~17 wt%) exceeded those of P7 (5 wt%) [23] and previously developed ether-functionalized polymers [24,25,26], which demonstrated the beneficial effect of the amide linker on polymer hydration (Fig. 4). Furthermore, the heating DSC profiles of hydrated P1–P4 exhibited exothermic peaks corresponding to cold crystallization of water at −40 to −10 °C (Figs. 5A–C and S8A), which is typically observed for hydrated PMEA and its derivatives [19, 22]. This cold crystallization is associated with biocompatibility, along with the crystallization of water at approximately −40 °C upon cooling observed for ether-functionalized APCs reported previously, including P7 (Fig. 5D) [23,24,25,26]. Hydrated P1–P4 are the first APCs revealed to exhibit cold crystallization of water by DSC analysis. The formation of water that underwent cold crystallization was mainly attributed to the combination of amide and side-chain aliphatic ether groups (Fig. 4). Considering the cooling and heating of hydrated polymers, water crystallization during heating is more affected by interactions with the polymer structure (which mitigates freezing) than crystallization during cooling is. Thus, we concluded that the amide linker of P1–P4 binds water more strongly than the ester linker of P7; therefore, the enhanced hydration yields cold crystallization of water. The extent of this hydration is reflected by the enthalpies of the cold crystallization of water, which are 17, 26, 13, and 6 J g–1 for hydrated P1, P2, P3, and P4, respectively (Figs. 5A–C and S8A). Given that the enthalpy of crystallization of water during cooling for hydrated P7 is 4 J g–1 (Fig. 5D), the enhancement of the hydration properties of P1 and P2 by the amide linker is shown to be significantly high. Despite not explicitly appearing in the related thermograms, the glass transitions of hydrated P1–P4 is expected to occur at temperatures below those of cold crystallization, which are considerably lower than those observed for dry states (Table 1). Accordingly, the amide linker in P1–P4 was concluded to substantially improve the hydration properties.

Fig. 5
figure 5

Differential scanning calorimetry profiles of hydrated (A) P1, (B) P2, (C) P3, and (D) P7. The areas of exothermic peaks on the heating (A–C) and the cooling (D) are shaded. The water contents are 17 wt% for P1–P3 and 5 wt% for P7

Full size image

Hydrolytic degradation of APCs

Polymer degradation is often conducted at ≤25 °C and neutral pH, especially when marine and environmental degradability is simulated [38]. However, previously developed APCs exhibit low degradability under mild conditions [25]. Given that we aimed to understand the relationship between side-chain structure and hydrolytic properties rather than to verify marine and environmental degradability, APC hydrolyzability was evaluated under moderate heating (37 °C) in an alkaline solution and in phosphate-buffered saline (PBS) formulated in deuterium oxide (D2O) (Fig. 6A). The former environment corresponded to accelerated degradation conditions and was used for comparison with PTMC, which is stable under such conditions [32], while the latter environment better represented physiological conditions relevant to biomaterials applications. Because none of the tested polymers were water soluble, their degradation proceeded inhomogeneously. These degradation processes were examined by visual inspection (Figs. 6B–D and S9) and 1H NMR spectroscopy using internal standards to estimate the amounts of soluble degradates (Figs. S10–S27).

Fig. 6
figure 6

Hydrolytic degradation of APCs in D2O. A Schematic degradation of the amide-linked APCs Px, affording degradates D(Px) (x = 1–6). B–D Typical appearances of degradation solutions at 30 days. B P1 was completely hydrolyzed in NaOD–D2O. C P7 was partially degraded in NaOD–D2O with residues attached to the stir bar. D P4 was partly degraded in phosphate-buffered saline–D2O with insoluble unreacted particles

Full size image

Alkaline hydrolysis was conducted in 0.1 M NaOD in D2O. The degradation behavior of amide-linked APCs (P1–P6) was different from that of nonamide-linked APCs (P7–P9). The alkaline hydrolyses of P1–P6 were complete (i.e., the solution became transparent) after 25 days (Fig. 6B and Table 2). The 1H NMR spectra of the related degradates revealed that amide-linked bis-MPA derivatives were the major products (m = 0 for D(Px) in Figs. 6A, 7A and S10–S15), and the peaks of the minor products were tentatively assigned to short-chain oligomers. The size-exclusion chromatography (SEC) profiles of the degradates showed single peaks appearing considerably later (at 10–11 min) than those of the original polymers (Fig. S28A–F). Thus, no polymers were contained in the degradates of P1–P6 produced in 0.1 M NaOD–D2O. In contrast, the alkaline hydrolyses of P7–P9 resulted in the formation of aggregates around stir bars (Figs. 6C and S9A). The water-soluble degradates in the corresponding supernatants were identified as monomer-derived diols, e.g., bis-MPA and 2-methoxyethanol, in the case of P7, which indicated that the ester linker was cleaved by alkaline hydrolysis (Figs. S16–S18). The degradation percentages of P7–P9 were calculated to be 28%, 30%, and 2.5%, respectively, based on the integrated NMR signals and dimethyl sulfone (Me2SO2) as the internal standard (Table 2). The percentages of residual P7–P9 calculated from the weights of the related dry residues were almost complementary (e.g., the values added to be 100%) to those of the soluble degradates (Table 2). These results are consistent with those of our previous studies, in which ether-functionalized APCs were degraded to a greater extent than PTMC was [25].

Table 2 Percentages of soluble degradates and residual weights of APCs after hydrolysis at 30 days
Full size table
Fig. 7
figure 7

1H NMR spectra of hydrolysates of P2 in (A) NaOD–D2O and (B) PBS–D2O (internal standards = dimethyl sulfone (Me2SO2) and sodium 3-(trimetylsilyl)-1-propane-1,1,2,2,3,3-d6-sulfonate (DSS-d6))

Full size image

P1 and P2 were susceptible to hydrolysis in PBS–D2O, yielding transparent solutions with low residual polymer contents after 30 days (Table 2 and Fig. S9B). For P3–P9, hydrolysis was considerably slower than that for P1 and P2. After 30 days, the residues of P3 and P7–P9 were present as aggregates around stir bars (Figs. 6C and S9B), while those of P4–P6 were dispersed in the solution as small particles (Figs. 6D and S9B). The supernatants (separated by centrifugation for P4–P6) were characterized by 1H NMR spectroscopy, which revealed that the soluble degraded compounds resembled those formed under alkaline conditions (Figs. 7B and S19–S27). The degradation percentages of P1–P6 depended on the side-chain structure (Table 2), i.e., exceeded 80% for P1 and P2 and were as low as ~20% for P3 and P4. Like P7–P9, P5 and P6 were difficult to hydrolyze under neutral conditions. The residual polymer percentages showed the opposite trend, i.e., those of P7–P9 were close to 100%, and those of P3 and P4 were 77% and 73%, respectively. The P5 and P6 residues were not fully recovered by centrifugation because of the very small particle size, and their contents (67% and 78%, respectively) were therefore underestimated. The SEC profiles of the P3 and P4 residues exhibited decreased molecular weights (Fig. S28C, D), which was indicative of bulk erosion–type degradation, whereas the molecular weights of the residual P5–P8 were unchanged (Fig. S28E–H). This degradation trend was closely associated with polymer hydration properties, such as water contact angles (θw in Table 1); that is, more hydrated polymers were degraded faster. The rapid degradation of hydrophilic P1 and P2 was attributed to the formation of a surface-localized swollen phase where the polymer chains underwent bulk erosion-type random scission by water.

A bis-MPA-based APC with ester-linked ammonium side chains was previously reported to undergo complete hydrolysis in PBS–D2O within 24 h because the ammonium side chains serve as nucleophiles and catalysts [39]. However, the high hydrolyzability of P1 and P2 could not be rationalized in the same way, as their hydrolysis was predominantly facilitated by polymer hydration enhanced by the amide linker and methoxy groups (Fig. 4). The hydrolysis of the examined APCs in PBS–D2O may be divided into two groups according to their hydration properties: P1–P4 and P5–P9. P1–P4 were hydrolyzed to some extent (>16%) and exhibited cold crystallization of water in the hydrated state, whereas P5–P9 showed no such crystallization and remained almost intact after hydrolysis (~2.7%). Preliminarily, we obtained a tendency for the degradation percentages to increase with increasing enthalpy during the crystallization of water (Fig. S29). However, further investigations should be performed to reach a quantitative conclusion by adopting spectroscopic analyses of the adsorption behavior of water to polymers [27]. Consequently, we clarified the impact of hydration (synergistically enhanced by amide and aliphatic ether groups) on APC hydrolysis.

Platelet adhesion properties of APCs

The platelet adhesion properties of the APCs were evaluated for spin-coated substrates by a method previously used for other ether-functionalized polymers [19,20,21,22,23,24,25,26]. A platelet suspension was placed on the polymer surface, and the adherent platelets were observed by scanning electron microscopy (SEM; Fig. S30) and quantified (Fig. 8). The amide-linked APCs (excluding P2) demonstrated higher adherent platelet counts than did P7 and PMEA, which were previously reported as blood-compatible polymers [19, 23]. For P1–P7, the number of adherent platelets increased with increasing number of side-chain carbons. P2 showed minimally lower platelet adhesion than did P1, which is in line with a previous report on acrylate analogs with 2-methoxyethyl and 3-methoxypropyl groups [22]. These results indicated that the amide linker had no adverse effect on platelet adhesion despite the potential influence of hydrogen bonding between the amide linker and biomolecules such as serum proteins. P3 and P4 exhibited stronger platelet adhesion than did P1, P2, and P7, likely because of the greater hydrophobicity of the side chains in the former case. The alkyl aryl ether structures in P5 and P6, which have never been assessed for platelet adhesion, did not efficiently lower platelet adhesion despite these molecules containing two and three methoxy groups, respectively. These results were also attributed to the hydration properties of the APCs (Table 1 and Figs. 5 and S8).

Fig. 8
figure 8

Number of adherent platelets on polymer substrates spin-coated on poly(ethylene terephthalate) sheets

Full size image

The desired degradation time of biomaterials in vivo depends on the device. Degradation of P1 and P2 is completed in a month, which may be shorter than the duration (over one month) needed for blood-contacting devices such as stent coatings and blood vessels [40]. In contrast, P7 is less susceptible to enzymatic hydrolysis than P9 is [23], which usually stays in the body for several months to a few years, depending on its molecular weight [30]. The present study yields P2, which exhibited high blood compatibility comparable to P7, as well as faster hydrolyzability. The combination of P2 and P7, including copolymerization, ultimately enables the production of highly blood-compatible polymers with fine-tuned degradation properties.

Conclusion

Six APCs with ether-functionalized side chains grafted via an amide linker were prepared by postpolymerization functionalization, which enabled the efficient grafting of ether-functionalized side chains without side reactions and main-chain cleavage. The amide linker provided several advantages stemming from hydrogen bonding. For example, compared to nonamide-linked analogs, amide-linked APCs showed higher Tg values and better hydration and hydrolytic properties. The introduction of the amide linker increased hydrolyzability by enhancing molecular interactions with water and promoting the hydrolysis of carbonate bonds. Thus, our results demonstrate the potential of side-chain engineering for preparing hydrolyzable APCs and pave the way for the purposeful design of polymers with tunable degradability.

Experimental section

Materials

All reagents used were purchased from Tokyo Chemical Industry (Tokyo, Japan), Sigma‒Aldrich Japan (Tokyo, Japan), Fujifilm Wako Pure Chemical (Tokyo, Japan), Kanto Chemical (Tokyo, Japan), and Nacalai Tesque (Kyoto, Japan) and were used as received unless otherwise stated. Dehydrated tetrahydrofuran (THF) and dichloromethane (CH2Cl2) were supplied by a solvent supply system (Kanto Chemical, water content <10 ppm). PBS (pH 7.4) was prepared by dissolving tablets of phosphate buffer salts (Takara Bio, Tokyo, Japan) in ultrapure water. A glutaraldehyde solution (1%) was prepared by diluting 25% aqueous glutaraldehyde (Wako Pure Chemical) with PBS.

The reference polymers, namely, P7 (Mn = 9 kg mol–1, Ð = 1.19), P8 (Mn = 10 kg mol–1, Ð = 1.23), P9 (Mn = 34 kg mol–1, Ð = 1.11), and PMEA (Mn = 21 kg mol–1, Ð = 2.66), were prepared as described elsewhere [19, 23, 25].

Measurements

1H, 13C, and 19F NMR spectra were acquired on a JEOL JNM-ECX400 spectrometer at 400, 100, and 376 MHz, respectively. Deuterated chloroform (CDCl3) and D2O were used as solvents, and tetramethylsilane and sodium 3-(trimethylsilyl)-1-propane-1,1,2,2,3,3-d6-sulfonate were used as internal standards. Mn and ĐM were estimated by SEC measurements, which were performed at 40 °C using THF (eluent) and an integrated SEC unit (Malvern Viscotek TDAmax) equipped with three TSK-gel (one G2000HHR and two GMHHR-H) columns connected in series and a refractive-index detector. Mn and ĐM were calibrated using polystyrene standards (580 to 2.2 × 106 g mol–1). SEC measurements were also performed at 40 °C in N,N-dimethylformamide (DMF) containing 10 mM LiBr using an integrated SEC unit (HLC-8420 EcoSEC Elite, Tosoh, Tokyo, Japan) with two TSK-gel columns (SuperAWM-H) connected in series and equipped with refractive index and ultraviolet (UV) detectors (UV-8420). In this case, poly(ethylene oxide) standards with Mn = 2.4 × 104–7.9 × 105 were used for calibration. DSC measurements (Hitachi High-Tech Science X-DSC7000) were performed at heating and cooling rates of 5 °C min−1 under nitrogen for specimens weighing 3–5 mg. Hydrated samples were prepared by immersion in ultrapure water at 20–25 °C for at least 24 h before the measurement. No weight loss occurred during the measurement.

Synthesis of P1

The precursor polymer (P0) was prepared as described elsewhere [36]. P0 (300 mg, 0.92 mmol) was dissolved in THF (1.5 mL) containing Amberlite-IRA67 (161 mg). The dispersion was slowly supplemented with a solution of 2-methoxyethylamine (72.7 mg, 0.97 mmol) in THF (1.0 mL) at 0–5 °C and stirred at 20–25 °C. After the conversion exceeded 90% (1H NMR control), the solution was filtered to remove the resin, and the filtrate was evaporated. The residue was dissolved in CH2Cl2 (2 mL) and reprecipitated by the addition of hexane (15 mL). Pentafluorophenol-containing precipitates were redissolved in CH2Cl2 (10 mL), and the solutions were washed with a saturated aqueous solution of NaHCO3 (10 mL) or passed through Amberlite-IRA67. The organic phase was then concentrated under vacuum to yield P1 as a white solid (124 mg, 62%). 1H NMR (400 MHz, CDCl3): δ 6.56–6.40 (br, NH), 4.29 (s, CH2O), 3.45 (s, NHCH2CH2OCH3), 3.35 (s, OCH3), 1.28 (s, CH3). 13C NMR (100 MHz, CDCl3): δ 171.7, 154.3, 70.9, 69.9, 58.8, 46.3, 39.5, 17.4.

Synthesis of P2

The reaction was performed as described for P1 using 3-methoxypropylamine (84.6 mg, 0.95 mmol) instead of 2-methoxyethylamine. The product was obtained as a transparent liquid (107 mg, 50%). 1H NMR (400 MHz, CDCl3): δ 6.88–6.72 (m, NH), 4.38–4.16 (m, CH2O), 3.49 (t, J = 5.66 Hz, CH2OCH3), 3.41–3.31 (m, NHCH2, CH2OCH3), 1.77 (quin, J = 5.89 Hz, CH2), 1.25 (s, CH3). 13C NMR (100 MHz, CDCl3): δ 171.4, 154.4, 71.8, 69.9, 58.7, 46.2, 38.6, 28.5, 17.2.

Synthesis of P3

The reaction was performed as described for P1 using 3-ethoxypropylamine (99.1 mg, 0.96 mmol) instead of 2-methoxyethylamine. The product was obtained as a transparent viscous solid (147 mg, 65%). 1H NMR (400 MHz, CDCl3): δ 6.80 (t, J = 4.98 Hz, NH), 4.32–4.19 (m, CH2O), 3.58–3.43 (m, CH2OCH2CH3), 3.38 (q, J = 5.59 Hz, NHCH2), 1.87–1.71 (m, CH2CH2CH2), 1.26 (s, CH3), 1.21 (t, J = 7.02 Hz, CH2CH3). 13C NMR (100 MHz, CDCl3): δ 171.3, 154.4, 70.0, 66.6, 46.3, 39.0, 28.8, 17.2, 15.3.

Synthesis of P4

The reaction was performed as described for P1 using tetrahydrofurfurylamine (95.6 mg, 0.95 mmol) instead of 2-methoxyethylamine. The product was obtained as a white solid by precipitation from hexane/Et2O (1:1, v/v) (126 mg, 56%). 1H NMR (400 MHz, CDCl3): δ 6.63–6.41 (br, NH), 4.34–4.22 (m, CH2O), 4.02–3.91 (m, CH), 3.89–3.68 (m, OCH2CH2), 3.63–3.47 (m, NHCHaHb), 3.28–3.09 (m, NHCHaHb), 2.05–1.82 (m, CH2CH2, CH2CHaHb), 1.601.45 (m, CH2CHaHb), 1.28 (s, CH3). 13C NMR (100 MHz, CDCl3): δ 171.7, 154.3, 77.6, 69.9, 68.2, 46.4, 43.4, 28.6, 25.9, 17.4.

Synthesis of P5

The reaction was performed as described for P1 using 3,4-dimethoxybenzylamine (161 mg, 0.96 mmol) instead of 2-methoxyethylamine. The product was obtained as a white solid (224 mg, 79%). 1H NMR (400 MHz, CDCl3): δ 6.80–6.70 (m, Ar–H), 6.61–6.50 (br, NH), 4.32 (d, J = 4.98 Hz, NHCH2), 4.25 (s, CH2O), 3.79 (s, OCH3), 1.22 (s, CH3). 13C NMR (100 MHz, CDCl3): δ 171.5, 154.3, 149.1, 148.4, 130.7, 119.8, 111.4, 111.2, 69.7, 55.9, 46.4, 43.4, 17.5.

Synthesis of P6

The reaction was performed as described for P1 using 3,4,5-trimethoxybenzylamine (192 mg, 0.98 mmol) instead of 2-methoxyethylamine. The product was obtained as a white solid (214 mg, 69%). 1H NMR (400 MHz, CDCl3): δ 6.79–6.61 (br, NH), 6.44 (s, Ar–H), 4.41–4.17 (m, NHCH2, CH2O), 3.82–3.72 (m, OCH3), 1.23 (s, CH3). 13C NMR (100 MHz, CDCl3): δ 171.7, 154.4, 153.3, 137.0, 134.1, 104.5, 69.7, 60.8, 56.1, 46.4, 43.7, 17.4.

Measurement of static contact angles

Contact angles were measured by sessile-drop and captive-bubble methods using a contact angle meter (Drop Master DM-501, Kyowa Interface Science, Japan) and polymer-coated substrates prepared as reported previously [23, 25, 26]. The contact angles were read 30 s after the deposition of a water droplet (2 µL) or air bubble (2 µL). Ultrapure water was used for both measurements, which were performed at three different points per substrate using five substrates per polymer.

Hydrolytic degradation test

Hydrolytic degradation was monitored by 1H NMR spectroscopy using PBS–D2O and NaOD–D2O (Sigma–Aldrich, Tokyo, Japan). The polymer of choice (10 mg) was dissolved in chloroform in a 5-mL scintillation vial, and the solution was evaporated and dried under vacuum to form a filmy sample. The vial was charged with 1 mL of PBS–D2O or 0.1 M NaOD–D2O, and the mixture was stirred at 37 °C. The water-soluble degradates in the supernatants were detected via 1H NMR at predetermined times and quantified using dimethyl sulfoxide and dimethyl sulfone (Sigma‒Aldrich, Tokyo, Japan) as internal standards. The completion of degradation was confirmed by visual inspection. At the completion of degradation or after 30 days, the reactions were quenched by the addition of deuterated trifluoroacetic acid. The solutions containing no residues were directly characterized by NMR spectroscopy, whereas those containing residues were separated into supernatants and residues. The supernatants were characterized by 1H NMR spectroscopy, and the residues were analyzed by SEC using 0.1 mM LiBr in DMF as the eluent. The tests were conducted twice for each sample to ensure reproducibility.

Platelet adhesion test

Platelet adhesion was quantified using the number of adherent platelets on APC-coated substrates according to a previously established procedure [23,24,25,26]. Briefly, a platelet suspension with a concentration of 4.0 × 107 cells cm–2 was placed on the polymer substrate, incubated for 1 h at 37 °C, and removed. The substrate was washed, and the adherent platelets were fixed with PBS containing 1% glutaraldehyde for SEM imaging (VE-9800, KEYENCE, Tokyo, Japan) at an acceleration voltage of 5 keV and magnification of 1500×. Observations were made for five randomly selected points per substrate, and three substrates were evaluated for each polymer. The number of adherent platelets was determined as the average of those manually counted in the SEM images (five points × three substrates).