Tailoring thermoresponsiveness of biocompatible polyethers: copolymers of linear glycerol and ethyl glycidyl ether

Linear polyglycerol is known as a highly hydrophilic and biocompatible polymer that is currently considered for numerous medical applications. Derived from this well-known structure, the synthesis of highly biocompatible, thermoresponsive polyether copolymers via statistical anionic ring-opening copolymerization of ethyl glycidyl ether (EGE) and ethoxy ethyl glycidyl ether (EEGE) is described. Subsequent deprotection of the acetal groups of EEGE yields copolymers of linear glycerol (linG) and EGE, P(linG-co-EGE). These copolymers showed monomodal and narrow molecular weight distributions with dispersities Đ ≤ 1.07. The microstructure was investigated via in situ1H NMR kinetics experiments, revealing reactivity ratios of rEEGE = 1.787 ± 0.007 and rEGE = 0.560 ± 0.002, showing a slightly favored incorporation of EEGE over EGE. Due to the deliberate incorporation of rather hydrophobic EGE units into the water soluble linPG, tunable thermoresponsive behavior is achieved with cloud point temperatures Tcp between 9.0–71.4 °C. Besides the commonly utilized method turbidimetry, temperature-dependent 1H NMR measurements were used for more accurate and reproducible results. The change of the hydrodynamic radii rH of the copolymers and their aggregates upon reaching Tcp was investigated via DOSY NMR spectroscopy. To explore possible biomedical applications, as an example, the cell viability and immunology of an exemplary P(linG-co-EGE) copolymer sample was investigated. Since both, cell viability and immunology are comparable to the gold standard PEG, the herein presented copolymers show high potential as biocompatible and thermoresponsive alternatives to PEG for biomedical applications.


Jaacks method for calculation of the reactivity ratios
The Jaacks method is a simplification of the well-known Mayo-Lewis equation 1 (Eq. S1) for the copolymerization of the monomers M 1 and M 2 . (S1) with r 1 = and r 2 = . The simplification of the Jaacks method is based on the use of an excess of M 1 . Therefore, the active chain ends consist almost completely of M 1 units, and the active chain ends with M 2 units can approximately be neglected. 2 The rates of monomer consumption can be simplified as followed: where is the active chain end with a M 1 unit. The division of Eq. S2 and Eq. S3 results in which can also be obtained from Eq. S1 for » 1 and « 1. 2 Integration of Eq. S4 yields Eq. S5.
The integrated equation (Eq. S5) is valid also for high conversions, if the excess of M 1 over M 2 is ensured. 2 In a plot of vs. , r 1 can be obtained as the slope of the graph. For non- terminal models, r 2 can be calculated from r 1 , since r 1 · r 2 = 1. 2,3

Synthesis of the monomer ethoxy ethyl glycidyl ether (EEGE)
The monomer EEGE was synthesized according to a literature synthesis by Fitton et al. 4 Fig. S1 shows the 1 H NMR spectrum of the synthesized EEGE.

NMR characterization of P(EEGE-co-EGE)
The 1 H NMR spectra of all synthesized homo-and copolymers are shown in Fig. S2, the 13 C NMR spectra are summarized in Fig. S3.

Calculation of copolymer composition and molecular weight
Calculation of the copolymer composition is exemplary shown for sample P(EEGE 0.43 -co-EGE 0.57 ) (Fig. S4). The spectrum was referenced to the solvent signal (DMSO-d 6 , δ = 2.50 ppm). The integral of the methylene group of the initiator end group (integral 2) was defined as 2. The degree of polymerization of EEGE (DP EEGE ) is calculated by division of integral 5 by 3 protons of the methyl group (H5). For the calculation of the degree of polymerization of EGE (DP EGE ), the number of protons of integral 5 is subtracted from integral 4. DP EGE is then calculated by division by 3 protons of the methyl group (H4) (Eq. S6).
The molecular weight is calculated by multiplying the degree of polymerization of the monomers with the corresponding molar mass and addition of the molar mass of the end groups.

SEC data for P(EEGE-co-EGE)
SEC measurements were performed in DMF and toluene as internal standard and calibrated with PEG standards. The SEC curves of additional P(EEGE-co-EGE) copolymers are shown in Fig. S5. The molecular weight distributions are narrow and monomodal for all copolymers.

MALDI-ToF MS characterization of P(EEGE-co-EGE)
An enlargement of the MALDI-ToF mass spectrum of P(EEGE 0.57 -co-EGE 0.43 ) (see Fig. 1) is shown in Fig. S6 where a glycidol repeating unit is highlighted. Parts of the acetal protecting groups of the EEGE units were deprotected by the acidic salt additive trifluoroacetic acid potassium salt during the measurement.

NMR characterization of P(linG-co-EGE)
The 1 H NMR spectra of all deprotected homo-and copolymers are shown in Fig. S7, the 13 C NMR spectra are summarized in Fig. S8. The spectra display no acetal proton signals, the cleavage of the EEGE acetal protecting groups has been successful.

SEC data for P(linG-co-EGE)
The SEC curves of additional P(linG-co-EGE) copolymers are shown in Fig. S9. The molecular weight distributions are narrow and monomodal for all copolymers but P(linG 0.57 -co-EGE 0.43 ). The elugram of this copolymer is broadened towards lower elution volume which is possibly caused by aggregation during the SEC measurement since no allylic initiation is observed in the 1 H NMR spectrum and the MALDI-ToF mass spectrum (see Fig. S11, ESI).   S10 shows the shift in elution volume of the deprotected P(linG 0.77 -co-EGE 0.23 ) copolymer compared to the protected P(EEGE 0.77 -co-EGE 0.23 ) copolymer. Due to the abstraction of the acetal protecting group, the molecular weight of the deprotected copolymer P(linG 0.77 -co-EGE 0.23 ) decreases during the deprotection and therefore the elution volume increases.

MALDI-ToF MS characterization of P(linG-co-EGE)
All P(linG-co-EGE) copolymers and the homopolymers PEGE and linPG were characterized via MALDI-ToF MS. For all samples but linPG, DCTB was chosen as matrix and KTFA as salt additive. The matrix of linPG was α-cyano-4-hydroxycinnamic acid (HCCA) and lithium chloride was added as salt additive. Fig. S11 shows an exemplary MALDI-ToF mass spectrum of the copolymer P(linG 0.57 -co-EGE 0.43 ).

Determination of the reactivity ratios r EEGE and r EGE
The monomer consumption was detected by the decrease of the proton signals in the 1 H NMR spectra.
Since the chemical shifts of the monomer proton signals of EGE and EEGE are similar, only the nonoverlapping part of one proton signal of the epoxide methylene group of each monomer is considered (red frame in Fig. S12).

Investigation of the critical solution behavior of P(linG-co-EGE) via turbidimetry
The data points of the turbidimetry vs. temperature curves were fitted by the following equation: The cloud point temperature T cp is defined as the temperature with 50% transmittance.

Investigation of immune cell viability and immunophenotype of P(linG-co-EGE)
The cell viability of P(linG 0.57 -co-EGE 0.43 ) and the reference monomethyl poly(ethylene glycol) (mPEG) was investigated via fluorescence-activated cell scanning (FACS) for dendritic cells (DC), polymorphonuclear cells (PMN) and T cells, the immunology was investigated via expression of the surface protein CD80 for dendritic cells (DC) and polymorphonuclear cells (PMN), as shown in Fig. S18. Further immunology tests were carried out via expression of the surface proteins CD86 (Fig. S19, top) and MHCII (Fig. S19, bottom) for the following cell lines: B cells, dendritic cells (DC), natural killer cells (NK), macrophages and polymorphonuclear cells (PMN).