Stimuli-Responsive Thiomorpholine Oxide-Derived Polymers with Tailored Hydrophilicity and Hemocompatible Properties

Thermo-responsive hydrophilic polymers, including those showing tuneable lower critical solution temperature (LCST), represent a continuous subject of exploration for a variety of applications, but particularly in nanomedicine. Since biological pH changes can inform the organism about the presence of disequilibrium or diseases, the development of dual LCST/pH-responsive hydrophilic polymers with biological potential is an attractive subject in polymer science. Here, we present a novel polymer featuring LCST/pH double responsiveness. The monomer ethylthiomorpholine oxide methacrylate (THOXMA) can be polymerised via the RAFT process to obtain well-defined polymers. Copolymers with hydroxyethyl methacrylate (HEMA) were prepared, which allowed the tuning of the LCST behaviour of the polymers. Both, the LCST behaviour and pH responsiveness of hydrophilic PTHOXMA were tested by following the evolution of particle size by dynamic light scattering (DLS). In weak and strong alkaline conditions, cloud points ranged between 40–60 °C, while in acidic medium no LCST was found due to the protonation of the amine of the THOX moieties. Additional cytotoxicity assays confirmed a high biocompatibility of PTHOXMA and haemolysis and aggregation assays proved that the thiomorpholine oxide-derived polymers did not cause aggregation or lysis of red blood cells. These preliminary results bode well for the use of PTHOXMA as smart material in biological applications.


Introduction
Smart polymers which react upon external stimuli are attractive materials in biomedical applications and promise for example the selective release of encapsulated drugs only under specific conditions or in targeted tissue or cells [1]. LCST behaviour is one unique feature polymers may demonstrate, which can be exploited in this regard [2]. Lower critical solution temperature (LCST) in water is an important property for polymers destined to biological applications. Below the LCST, the polymer is soluble due to the H-bonds formed with water molecules [3]. Above the LCST, these H-bonds are disrupted and polymer chain aggregation takes place as a result of non-covalent interactions between hydrophobic moieties of polymer [4]. In addition, above the cloud point, the entropy is a dominating factor which results in the release of water molecules and turns into the collapse and coagulation of the polymer chains [4].

Instrumentation
1 H-NMR spectra were recorded on NMR Bruker Avance 400-MHz or III HD-400 MHz spectrometers using CDCl 3 or DMSO-d 6 as deuterated solvent. The chemical shifts of protons were relative to tetramethylsilane (TMS) at δ = 0.
Size exclusion chromatography (SEC) data were obtained in DMF containing 0.1 wt% LiCl, with a flow rate of 0.8 mL/min at 40 • C. Samples were filtered using TE36 Whatman PTFE-supported membrane filter paper (0.45 µm, 47 mm diameter) before the injection. The data were calibrated using poly(methyl methacrylate) (PMMA) narrow standards.
Fourier transform infrared spectroscopy (FTIR) analysis was achieved with a Perkin Elmer Spectrum 100 spectrometer. The spectral data were acquired in the 3500-500 cm −1 range.
Dynamic light scattering (DLS) analyses were performed on a Malvern Zetasizer NanoZS instrument at a detection angle of 173 • (back scattering), in the 10-60 • C temperature range.
Cytotoxicity and hemolysis assays were performed using a Tecan plate reader (Tecan, Männedorf, Switzerland). The fluorescence measurements used to determine the cell viability were assessed using an Infinite M200 PRO microplate reader from Tecan, Germany.
Following the previous protocol, a range of statistical copolymers of 2-thiomorpholine oxide ethyl methacrylate (THOXMA) and 2-hydroxyethyl methacrylate (HEMA) P(THOXMA nstat-HEMA m ) were prepared. For example, P(THOXMA 50 -stat-HEMA 50 ) was prepared as follows. THOXMA (2 g, 8.65 mmol, 50 eq.), HEMA (1.12 g, 8.65 mmol, 50 eq.), CPDB (38 mg, 1 eq.), AIBN (7 mg, 0.25 eq.) and 1.5 mL of DMSO were mixed in a 10 mL ampoule. The mixture was degassed via three freeze pump thaw cycles, the ampoule was filled with nitrogen and then immersed in an oil bath at 75 • C. An aliquot was taken every hour for 1 H-NMR and SEC-HPLC analyses. After 4 h (conversion above 99% for both monomers), the reaction was quenched by the exposure of the mixture to air. At the end, the mixture was dialysed against water (with a 1 kDa MWCO membrane) for 2 days, followed by lyophilisation for 1 day. The pink polymer (yield: 78%) was analysed by 1 H NMR and where M n,chain trans f er agent = 221.34 g/mol, M n,HEMA = 130.14 g/mol, M n,THOXMA = 231 g/mol. DP (the degree of polymerisation) was calculated by 1 H-NMR according to Equation (2). The conversion of HEMA and THOXMA co-monomers (Conv. HEMA, Conv. THOXMA) was calculated by 1 H-NMR according to Equation (1).

Determination of the pK a of the Monomers and Polymers
The pK a of the monomers and corresponding polymers were evaluated by titration with a solution of HCl. Aqueous solutions of THOXMA (0.1 M) and PTHOXMA (0.1 M) were titrated with a solution of HCl 0.1 M. Each measurement of the pH (at different volumes of HCl added) was carried out in triplicate. The pK a values were derived from the values of pK b which were measured at the midpoint of the titration curves. The pK a was calculated according to Equation (4).
Additional experiments were performed in an NaCl aqueous solution (0.9 wt%) in order to evaluate the acido-basic properties of PTHOXMA in conditions similar with those of physiological medium.
2.8. Evaluation of the LCST of P(THOXMA 100 ) and P(THOXMA n -stat-HEMA m ) Statistical Copolymers DLS measurements were performed at different temperatures ranging between 10 • C and 60 • C. The temperature corresponding to a sharp increase in the particle size was associated with the LCST of the statistical P(THOXMA n -stat-HEMA m ) copolymers. The LCST of P(THOXMA) 100 was determined by extrapolation (to % HEMA = 0) on the plot of the evolution of the LCST value as a function of m. For each measurement, solutions of 1 g/L of the corresponding copolymers in three distinct buffers (pH = 4, 7.4 and 10) were analysed by DLS. The measurements were carried out in triplicate.

Cytotoxicity Assays
Cytotoxicity studies were performed using the mouse fibroblast cell line L929 (400620, CLS), as recommended by ISO10993-5. L929 cells were routinely cultured in Dulbecco's Modified Eagle Medium with 2 mM L-glutamine supplemented with 10% fetal calf serum, 100 U/mL penicillin and 100 µg/mL streptomycin at 37 • C under a humidified 5% (v/v) CO 2 atmosphere. In detail, cells were seeded at 10 3 cells/mL (10 4 cells per well) in a 96-well plate and incubated for 24 h. No cells were seeded in the outer wells. The medium was changed to fresh cell culture medium 1 h prior to treatment. Afterward, the cold polymer solution in 20 mM HEPES (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid) was added to the cells at the indicated concentrations (from 5 to 700 µg/mL), and the plates were incubated for 24 h. The control cells were incubated with fresh culture medium containing the same amount of HEPES as the treated cells. Subsequently, the medium was replaced by a mixture of a fresh culture medium and the resazurin-based solution PrestoBlue (prepared according to the manufacturer's instructions). After further incubation for 45 min at 37 • C under a humidified 5% (v/v) CO 2 atmosphere, the fluorescence was measured at λ ex = 560 nm/λ em = 590 nm with gain set to optimal, with untreated cells on the same well plate serving as negative controls. The negative control was standardised as 0% of metabolism inhibition and referred to as 100% viability. Cell viability below 70% was considered to be indicative of cytotoxicity. Experiments were conducted in six technical replicates. All experiments were conducted including blanks and negative controls.

Haemolysis Tests
The membrane damaging properties of the polymer were quantified by analysing the release of haemoglobin from erythrocytes. Sheep blood was provided by the Institute for Experimental Animal Science and Animal Welfare, Jena University Hospital. Briefly, sheep blood was centrifuged at 4500× g for 5 min. The pellet was washed three times with PBS (pH 7.4) by centrifugation at 4500× g for 5 min. Erythrocytes were suspended in PBS at pH 7.4 to resemble physiological conditions in blood/cytoplasm or in PBS at pH 6 to mimic the slightly acidic environment in the early endosome. The polymer was dissolved in cold 20 mM HEPES (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid) at a concentration of 7 mg/mL and diluted to 1 mg/mL. Cold polymer solutions of different concentrations in PBS with the respective pH were mixed 1:1 with cold erythrocyte suspensions and were incubated at 37 • C for 1 h. Erythrocyte suspensions were centrifuged at 2400× g for 5 min. The release of haemoglobin in the supernatant was determined at 544 nm. The absorbance was measured using a plate reader. Concurrently, determinations were conducted with washed erythrocytes either lysed with 1% Triton X-100 or suspended in PBS at the respective pH as a reference. The haemolytic activity of the polymer was calculated as follows (Equation (5)): Here, A (sample) , A (PBS) and A (Triton X-100) are the absorbance of erythrocytes incubated with a respective sample, suspended in PBS and erythrocytes lysed with Triton X-100, respectively. The analysis was repeated with blood from three different animals.

Erythrocyte Aggregation
To investigate the behaviour of the pH-sensitive polymer towards cellular membranes at different pH values, red blood cells were treated with the polymer under physiological conditions in human blood (pH 7.4) and in a slightly acidic environment representing the pH of the early endosome (pH 6). Erythrocyte suspensions in PBS at different pH values were prepared and mixed 1:1 with polymer solutions as described above. After incubation at 37 • C for 2 h, erythrocyte aggregation was measured at 645 nm. As positive and negative assay controls, erythrocytes were treated with 50 µg/mL 25 kDa branched poly(ethylene imine) (bPEI) solution or PBS buffer at a respective pH. The aggregation activity of the polymer at different concentrations is given as an aggregation rate calculated as follows (Equation (6)): Here, A (sample) is the mean absorbance of a given sample.

Synthesis of Monomers and Hydrophobic/Hydrophilic Polymers
Two new methacrylates derived from thiomorpholine (named THMA) or thiomorpholineoxide (named THOXMA) were prepared as presented in Figure 1. THMA was prepared by nucleophilic substitution of 2-bromoethylmethacrylate with thiomorpholine in the presence of K 2 CO 3 . The one-step reaction afforded THMA with a yield of 75% ( Figure 1). The mild oxidation of THMA with H 2 O 2 (aq., 30%) produced THOXMA with a yield of 80% ( Figure 1). Monomers were obtained in sufficient purity without a need for elaborate purification methods such as distillation or column chromatography. 1 H-NMR analysis confirmed the formation of both pure monomers ( Figure S1). FTIR spectroscopy confirmed the presence of sulfoxide group (S=O) in THOXMA and the success of the oxidation reaction ( Figure S2).

Erythrocyte Aggregation
To investigate the behaviour of the pH-sensitive polymer towards cellular membranes at different pH values, red blood cells were treated with the polymer under physiological conditions in human blood (pH 7.4) and in a slightly acidic environment representing the pH of the early endosome (pH 6). Erythrocyte suspensions in PBS at different pH values were prepared and mixed 1:1 with polymer solutions as described above. After incubation at 37 °C for 2 h, erythrocyte aggregation was measured at 645 nm. As positive and negative assay controls, erythrocytes were treated with 50 μg/mL 25 kDa branched poly(ethylene imine) (bPEI) solution or PBS buffer at a respective pH. The aggregation activity of the polymer at different concentrations is given as an aggregation rate calculated as follows (Equation (6)): Here, A(sample) is the mean absorbance of a given sample.

Synthesis of Monomers and Hydrophobic/Hydrophilic Polymers
Two new methacrylates derived from thiomorpholine (named THMA) or thiomorpholine-oxide (named THOXMA) were prepared as presented in Figure 1. THMA was prepared by nucleophilic substitution of 2-bromoethylmethacrylate with thiomorpholine in the presence of K2CO3. The one-step reaction afforded THMA with a yield of 75% (Figure 1). The mild oxidation of THMA with H2O2 (aq., 30%) produced THOXMA with a yield of 80% ( Figure 1). Monomers were obtained in sufficient purity without a need for elaborate purification methods such as distillation or column chromatography. 1 H-NMR analysis confirmed the formation of both pure monomers ( Figure S1). FTIR spectroscopy confirmed the presence of sulfoxide group (S=O) in THOXMA and the success of the oxidation reaction ( Figure S2). Since THOXMA is hydrophilic, one of the goals of this work was to study the homoand co-polymerisation of this monomer, in order to develop new hydrophilic polymers. The RAFT process was used to prepare P(THOXMA)100 homopolymer (Figure 2), as well as a series of P(THOXMAn-stat-HEMAm) statistical copolymers ( Figure 3). All polymerisations were performed in DMSO using CPDB as chain transfer agent, while the pH was maintained at 4 to prevent CPDB degradation via hydrolysis. The molar ratios of monomer(s):CPDB:AIBN was kept at 100:1:0.25 in all polymerisations (Table S1). Since THOXMA is hydrophilic, one of the goals of this work was to study the homoand co-polymerisation of this monomer, in order to develop new hydrophilic polymers. The RAFT process was used to prepare P(THOXMA) 100 homopolymer (Figure 2), as well as a series of P(THOXMA n -stat-HEMA m ) statistical copolymers ( Figure 3). All polymerisations were performed in DMSO using CPDB as chain transfer agent, while the pH was maintained at 4 to prevent CPDB degradation via hydrolysis. The molar ratios of monomer(s):CPDB:AIBN was kept at 100:1:0.25 in all polymerisations (Table S1).
The 1 H NMR spectrum of P(THOXMA) 100 , a new hydrophilic homopolymer, is shown in Figure 2. High conversion (i.e., 99.3%, by 1 H-NMR) of THOXMA monomer was achieved after 6 h (Table 1, Figure S3). The linear increase in M n over time ( Figure 2B.), the linear first-order kinetic plot ( Figure S3) as well as the low dispersity (~1.2) indicate that the homopolymerisation was well controlled.   THOXMA was then copolymerised with HEMA by RAFT. The structure and the characterisation results of all the (co)polymers are presented in Figure 3 and Table 1, respectively. Overall, the DPs of the resulting polymers (evaluated by 1 H NMR, Figure  S5) were above 80 and were close to the targeted DPs. High conversions (≥95% after 4 h, Table 1, Figure S4), determined by 1 H NMR, were attained for all (co)polymerisations. Again, the linear evolution of ln(M 0 /M) with time ( Figure S4), as well as the low dispersity (remaining relatively constant around 1.2) (Table 1) suggested that the polymerisations were controlled.  Table 1) evaluated by SEC (B); 1 H-NMR spectrum of P(THOXMA100) in DMSO d6 (C).

Acido-Basic Properties of THOXMA and PTHOXMA
The acido-basic properties of the monomer and polymer are summarised in Table S3 and shown in Figure S6. THOXMA displayed a pK a value around 5.42 (Table S3). This value corresponds to a slightly acidic character. The pK a of PTHOXMA is around 5.57 (Table S3). However, this value is lower than that of polymers presenting a tertiary amine function, such as poly(2-(dimethyl amino)ethyl methacrylate PDMAEMA (pK a = 7.5) [22] and could be due to the steric hindrance of the heterocycle which renders the tertiary amine less accessible to protonation by the acid [23]. Hausig et al. [13] reported low pK a values (between 6 and 7.1) for polymers comporting N-alkyl-piperazine units. Low pK a values were also observed for poly(2-methyl-acrylic acid 2-[(2-(dimethylamino)-ethyl)-methylamino]-ethyl ester), [24] a polymer used in cellular transfection. The tertiary amine situated in the beta position of the ester function presented a pK a around 5 [24]. In addition, low pK a values (around 4.9) were reported by Butun et al. for poly(morpholine ethyl methacrylate) (PMEMA) [18]. According to their study [18], these unexpectedly low values of pK a were a consequence of intra-molecular cyclisation (between the amino group of cyclic morpholine and the carbonyl groups of side chains) which decreased the overall basicity of the polymers. The same ring structure likely also explains the low pK a of PTHOXMA. The presence, in THOXMA, of a sulfonyl group which is less electron withdrawing than an oxygen atom may explain the higher pK a of PTHOXMA compared to PMEMA.
When the acido-basic properties of PTHOXMA were tested in a solution of NaCl (0.9 wt%), at a concentration specific to physiological serum, the pK a values shifted slightly from 5.57 to 5.65 (Table S3). This slight increase in the pK a values in NaCl solution is related to the shielding effect of the salt that minimised the charge repulsion between the protonated amino groups. This result was consistent with previous works reported by Douglas et al. [25]. It is important to underline that the pK a value of PTHOXMA around 5.6 is interesting for a potential use in biological applications such as cellular transfection [26].

Determination of LCST
LCST was determined using DLS. The cloud point for PTHOXMA was barely visible as the upper temperature limit was 60 • C. Therefore, the cloud point temperature of the copolymers with variable composition was used and the LCST was estimated via extrapolation to pure PTHOXMA. HEMA was chosen because it provides a structural ethylmethacrylate pattern similar to THOXMA and because it is hydrophilic and biocompatible. The statistical P(THOXMA n -stat-HEMA m ) copolymers were dissolved in a range of aqueous buffers (at pH 4, 7.4 and 10) and the particle size evolution vs. temperature was investigated at each pH, in order to determine the LCST of the various copolymers.
The particle size variation over temperature at physiological pH (pH 7.4) is presented in Figure 4A. All copolymers showed a sharp increase in the particle size in the temperature range examined (10-60 • C), which was considered as the cloud point temperature. As expected, the decrease in the THOXMA molar fraction (from 75% to 20%) shifted down the cloud point temperature from 56 • C to 42 • C. The decrease in the cloud point temperature could be explained by the intermolecular H-bonds induced by HEMA.

Determination of LCST
LCST was determined using DLS. The cloud point for PTHOXMA was barely visible as the upper temperature limit was 60 °C. Therefore, the cloud point temperature of the copolymers with variable composition was used and the LCST was estimated via extrapolation to pure PTHOXMA. HEMA was chosen because it provides a structural ethylmethacrylate pattern similar to THOXMA and because it is hydrophilic and biocompatible. The statistical P(THOXMAn-stat-HEMAm) copolymers were dissolved in a range of aqueous buffers (at pH 4, 7.4 and 10) and the particle size evolution vs. temperature was investigated at each pH, in order to determine the LCST of the various copolymers.
The particle size variation over temperature at physiological pH (pH 7.4) is presented in Figure 4A. All copolymers showed a sharp increase in the particle size in the temperature range examined (10-60 °C), which was considered as the cloud point temperature. As expected, the decrease in the THOXMA molar fraction (from 75% to 20%) shifted down the cloud point temperature from 56 °C to 42 °C. The decrease in the cloud point temperature could be explained by the intermolecular H-bonds induced by HEMA.  At pH 10, the polymers had a behaviour close to that of pH 7.4 ( Figure 4B). The polymer cloud point varied from 52 • C (for a content of 75% in THOXMA) to 36 • C (for 20% THOXMA content).
Then, by extrapolation of the cloud point temperatures measured at different contents in HEMA to 0% HEMA, the apparent cloud point temperature of PTHOXMA was determined in both physiological and alkaline environments ( Figure 5). At pH 7.4, the cloud point temperature was 65.4 • C, and at pH 10 the cloud point temperature was 57.9 • C.
In comparison, PMEMA of similar DP (around 100) and at a similar concentration (1% w/v) displayed cloud points around 36 • C at pH 7 and at lower temperatures at higher pH (8 and 10) [18]. The differences between PTHOXMA and PMEMA are thus subtle. PTHOXMA is more water soluble and shows higher cloud points than PMEMA at different pHs. This is likely a consequence of the slightly higher pK a of PTHOXMA. ymer cloud point varied from 52 °C (for a content of 75% in THOXMA) to 36 °C (for 20% THOXMA content).
Then, by extrapolation of the cloud point temperatures measured at different contents in HEMA to 0% HEMA, the apparent cloud point temperature of PTHOXMA was determined in both physiological and alkaline environments ( Figure 5). At pH 7.4, the cloud point temperature was 65.4°C, and at pH 10 the cloud point temperature was 57.9 °C. In comparison, PMEMA of similar DP (around 100) and at a similar concentration (1% w/v) displayed cloud points around 36 °C at pH 7 and at lower temperatures at higher pH (8 and 10) [18]. The differences between PTHOXMA and PMEMA are thus subtle. PTHOXMA is more water soluble and shows higher cloud points than PMEMA at different pHs. This is likely a consequence of the slightly higher pKa of PTHOXMA.
In contrast, at pH = 4, no cloud point was observed for PTHOXMA (or for PMEMA). In acid medium the particle size was constant (around 7 nm) over the complete range of temperatures scanned (10-60 °C) and for all copolymer compositions ( Figure S7). This result is a consequence of the complete protonation of polymer chains at such acidic pH, leading to hydrosoluble polymers which were not sensitive to temperature.

Biocompatibility of PTHOXMA
In biological applications such as cellular transfection, the polymer that complex the genetic material must survive to the endosomal passage (where the pH is ~5-6) and thus prevent the degradation of the genes. Cationic polymers are currently used for this purpose, but they are toxic and damage the cell surface [19]. Thus, polymers acting as a proton-sponge system are highly investigated to overcome the endosomal barrier. Proton-sponge polymers are neutral at physiological pH (so before entering the endosome) but become charged in an acid environment (specific to the endosome). As it was presented before, PTHOXMA is neutral at physiological conditions, but become charged at a pH around 5.6, so it could emphasise an emerging potential for this application. Herein, the biocompatibility of PTHOXMA was assessed, by evaluating in vitro the cytotoxicity and hemocompatibility of this polymer. Copolymers of THOXMA and HEMA were not assessed since PHEMA is known to be biocompatible [27]. If both PHEMA and PTHOXMA are biocompatible, copolymers of HEMA and THOXMA will very likely be biocompatible too. In contrast, at pH = 4, no cloud point was observed for PTHOXMA (or for PMEMA). In acid medium the particle size was constant (around 7 nm) over the complete range of temperatures scanned (10-60 • C) and for all copolymer compositions ( Figure S7). This result is a consequence of the complete protonation of polymer chains at such acidic pH, leading to hydrosoluble polymers which were not sensitive to temperature.

Biocompatibility of PTHOXMA
In biological applications such as cellular transfection, the polymer that complex the genetic material must survive to the endosomal passage (where the pH is~5-6) and thus prevent the degradation of the genes. Cationic polymers are currently used for this purpose, but they are toxic and damage the cell surface [19]. Thus, polymers acting as a proton-sponge system are highly investigated to overcome the endosomal barrier. Protonsponge polymers are neutral at physiological pH (so before entering the endosome) but become charged in an acid environment (specific to the endosome). As it was presented before, PTHOXMA is neutral at physiological conditions, but become charged at a pH around 5.6, so it could emphasise an emerging potential for this application. Herein, the biocompatibility of PTHOXMA was assessed, by evaluating in vitro the cytotoxicity and hemocompatibility of this polymer. Copolymers of THOXMA and HEMA were not assessed since PHEMA is known to be biocompatible [27]. If both PHEMA and PTHOXMA are biocompatible, copolymers of HEMA and THOXMA will very likely be biocompatible too.
Firstly, the cytotoxicity of the PTHOXMA was studied on the mouse fibroblast cell line L929 using PrestoBlue assay, at pH 7.4 ( Figure 6A). This assay works as a cell health indicator which uses the reducing ability of living cells in order to measure the cellular viability [28]. A cellular viability below 0.7 (or 70%) indicates cytotoxic behaviour [29]. These assays showed that PTHOXMA presented no cytotoxicity on L929 cells for concentrations below 400 µg/mL, while the cellular viability was above 0.9 (90%). These promising results further proved the interest to investigate PTHOXMA for blood compatibility.
Then, further investigations were conducted to study the red blood cell aggregation activity of PTHOXMA in sheep blood (i.e., pH 7.4) and in a slightly acidic environment representing the pH of the early endosome (i.e., pH 6). The results are summarised in Figure 6B, where the aggregation activity was expressed by the aggregation rate and compared to a polycationic commercial polymer (i.e., polyethyleneimine PEI) which is known for its high aggregation rate [30]. For both pH values tested and all PTHOXMA concentrations, an aggregation rate equal to 1 was observed. These results confirm that PTHOXMA did not cause the undesired cell aggregation.
Firstly, the cytotoxicity of the PTHOXMA was studied on the mouse fibroblast cell line L929 using PrestoBlue assay, at pH 7.4 ( Figure 6A). This assay works as a cell health indicator which uses the reducing ability of living cells in order to measure the cellular viability [28]. A cellular viability below 0.7 (or 70%) indicates cytotoxic behaviour [29]. These assays showed that PTHOXMA presented no cytotoxicity on L929 cells for concentrations below 400 μg/mL, while the cellular viability was above 0.9 (90%). These promising results further proved the interest to investigate PTHOXMA for blood compatibility. Then, further investigations were conducted to study the red blood cell aggregation activity of PTHOXMA in sheep blood (i.e., pH 7.4) and in a slightly acidic environment representing the pH of the early endosome (i.e., pH 6). The results are summarised in Figure 6B, where the aggregation activity was expressed by the aggregation rate and compared to a polycationic commercial polymer (i.e., polyethyleneimine PEI) which is known for its high aggregation rate [30]. For both pH values tested and all PTHOXMA concentrations, an aggregation rate equal to 1 was observed. These results confirm that PTHOXMA did not cause the undesired cell aggregation.
Since PTHOXMA did not provoke the aggregation of red blood cells, further attempts were developed in order to investigate if the thiomorpholine oxide-containing homopolymer could damage the red blood cell membrane. To this regard, the release of haemoglobin from the erythrocytes was measured [29]. As evidenced in Figure 6C, the haemoglobin release was studied in two selected media: at pH 7.4 (to mimic the physiological conditions) and at pH 6 (specific for endosomal escape process which is a reference step in cellular transfection applications). At physiological pH (i.e., pH 7.4), haemoglobin was released in low amounts (below 1%), without significant influence of the polymer concentration in the blood medium. Haemoglobin release value below 2% is correlated with a non-haemolytic activity, [31] this result thus suggested that the PTHOXMA is blood compatible at pH 7.4 for concentrations between 10-100 μg/mL and that it did not damage the Since PTHOXMA did not provoke the aggregation of red blood cells, further attempts were developed in order to investigate if the thiomorpholine oxide-containing homopolymer could damage the red blood cell membrane. To this regard, the release of haemoglobin from the erythrocytes was measured [29]. As evidenced in Figure 6C, the haemoglobin release was studied in two selected media: at pH 7.4 (to mimic the physiological conditions) and at pH 6 (specific for endosomal escape process which is a reference step in cellular transfection applications). At physiological pH (i.e., pH 7.4), haemoglobin was released in low amounts (below 1%), without significant influence of the polymer concentration in the blood medium. Haemoglobin release value below 2% is correlated with a non-haemolytic activity, [31] this result thus suggested that the PTHOXMA is blood compatible at pH 7.4 for concentrations between 10-100 µg/mL and that it did not damage the plasma membrane of the erythrocytes. In slightly acidic conditions (i.e., pH 6), a concentration-dependent haemoglobin release profile was observed. Particularly, an increase in PTHOXMA concentration from 10 µg/mL to 100 µg/mL led to an increase in the haemoglobin percentage from 0.8% to 6%, which indicates a variable haemolytic activity. For example, at the middle concentration of 50 µg/mL, a haemoglobin release slightly above 2% was detected. At this concentration, PTHOXMA was slightly haemolytic, so the blood cells were not significantly damaged. On the contrary, at 100 µg/mL, the PTHOXMA is highly haemolytic and it interacts with the cell membrane. Compared to the results obtained at physiological pH, the different hemocompatibility observed at pH 6 is likely a result of partial protonation of the thiomorpholine oxide heterocycles, since this pH is close to the pK a of PTHOXMA (around 5.6). Despite this slight haemolytic activity at high concentrations, the results at pH 6 are very promising for a prospective use of PTHOXMA in cellular transfection application, which requires a polymer that promotes the release of active substances from the endosome. Statistical and block copolymers of DMAEMA and MEMA were also shown to be biocompatible and showed interesting results as transfection agents [17,32]. Given its in vitro non-cytotoxicity and hemocompatibility, PTHOXMA is also potentially suitable for drug delivery and transfection strategies. Its higher pK a compared to PMEMA could also be advantageous for drug release approaches. These applications are under investigation in our laboratories and will be reported in due course.

Conclusions
New thiomorpholine oxide-containing polymers with tailored LCST at different pHs were developed in order to design stimuli-responsive materials for biological applications. Hydrophilic poly(thiomorpholine oxide ethyl methacrylate) polymers possessed a pK a around 5.6 which denotes a weak acid character which could be exploited in biological applications such as cellular transfection. Interesting results were obtained concerning the behaviour of hydrosoluble PTHOXMA in environments with different pHs. At physiological and alkaline pHs, LCSTs around 65 • C and 58 • C were reported, while in acid conditions the polymer remained hydrophilic. In acid medium, the protonation of the PTHOXMA amino groups increased the solubility of the polymer and no aggregation occurred at any temperature. The high LCST values (65 • C in physiological conditions and 58 • C in alkaline conditions) open the gates to explore a large area of applications, such as thermal therapy in biomedicine (or thermal tumour ablation which requires temperatures above 50 • C) [33]. Lastly, PTHOXMA showed no-cytotoxicity and no haemolytic behaviour, without any cellular aggregation, which proved promising biocompatibility. To conclude, this study highlights the development of non-cytotoxic, blood compatible, pHand temperature-responsive polymers based on thiomorpholine oxide ethyl methacrylate with tailored LCST, which may find applications in biosciences.

Conflicts of Interest:
The authors declare no conflict of interest.