Thiolated α-cyclodextrin: The likely smallest drug carrier providing enhanced cellular uptake and endosomal escape

This study aimed to evaluate the effect of thiolated α-cyclodextrin (α-CD-SH) on the cellular uptake of its payload. For this purpose, α-CD was thiolated using phosphorous pentasulfide. Thiolated α-CD was characterized by FT-IR and 1 H NMR spectroscopy, differential scanning calorimetry (DSC), and powder X-ray diffractometry (PXRD). Cytotoxicity of α-CD-SH was evaluated on Caco-2, HEK 293, and MC3T3 cells. Dilauryl fluorescein (DLF) and coumarin-6 (Cou) serving as surrogates for a pharmaceutical payload were incorporated in α-CD-SH, and cellular uptake was analyzed by flow cytometry and confocal microscopy. Endosomal escape was investigated by confocal microscopy and hemolysis assay. Results showed no cytotoxic effect within 3 h, while dose-dependent cytotoxicity was observed within 24 h. The cellular uptake of DLF and Cou was up to 20- and 11-fold enhanced by α-CD-SH compared to native α-CD, respectively. Furthermore, α-CD-SH provided an endosomal escape. According to these results, α-CD-SH is a promising carrier to shuttle drugs into the cytoplasm of target cells.

Shortcomings of CDs, such as, in some cases, poor aqueous solubility or short gastrointestinal residence time, can be addressed by the design of derivatives. The covalent attachment of hydroxypropyl and sulfobutyl groups to CDs leads to highly soluble derivatives (Poulson et al., 2022). Among CD derivatives, thiolated CDs moved into the limelight of research more recently (Asim, Ijaz, Rösch, & Bernkop-Schnürch, 2020;Grassiri, Cesari, et al., 2022;Grassiri, Knoll, et al., 2022). Since thiolated CDs are able to form disulfide bonds with cysteine-rich subdomains of mucus glycoproteins, they exhibit high mucoadhesive properties, substantially prolonging the mucosal residence time of numerous drugs. Grassiri et al. showed a 4-fold increased concentration of dexamethasone in the aqueous humour of rabbits due to a significantly prolonged ocular residence time of this drug provided by thiolated hydroxypropylβ-CD versus the corresponding unmodified CD (Grassiri, Knoll, et al., 2022). Introducing thiol groups might even lead to other beneficial properties of CDs as well. In particular, cellular uptake-enhancing properties could be gained by thiolation, as thiol groups naturally present on the cell surface are known to enhance cellular association and internalization of various materials bearing thiol-reactive groups in their structure (Torres & Gait, 2012). The transfection efficacy of chitosan/pDNA nanoparticles was 2.5-fold enhanced when using thiolated chitosan J o u r n a l P r e -p r o o f Journal Pre-proof instead of the unmodified polysaccharide in these nanocarriers (Loretz, Thaler, & Bernkop-Schnürch, 2007). So far, however, the impact of thiol groups in the structure of CDs on their cellular uptake has not been investigated. We hypothesize that thiolated CDs enhance cellular uptake.
It was, therefore, the aim of this study to evaluate the cellular uptake of a thiolated CD. To that end, the smallest available CD (α-CD) was thiolated (α-CD-SH) according to a method recently developed by our research group (Kali, Haddadzadegan, Laffleur, & Bernkop-Schnürch, 2023) and characterized by FT-IR, 1 H NMR, XRD, and DSC. Cytotoxicity of α-CD-SH was evaluated on Caco-2, HEK 293, and MC3T3 cells. Dilauryl fluorescein (DLF) and coumarin-6 were chosen because of their fluorescence as surrogates for a pharmaceutical payload and incorporated in α-CD-SH to probe its drug-transporting potential. Cellular uptake was analyzed by flow cytometry and confocal microscopy. Based on the latter, we generated the new hypothesis that and endosomal escape could be promoted by thiolation. This was investigated via a hemolysis assay and confocal microscopy.

Synthesis of α-CD-SH
α-CD was thiolated according to a method described previously for β-CD with some modifications . First, α-CD (1 g, 1.02 mmol) and phosphorus pentasulfide (4.8 g, 21.5 mmol) were dissolved in 30 mL of sulfolane. Triethylamine (2 mL, 14.3 mmol) was added to the solution. The mixture was heated to 150 °C and stirred for one hour. At the end of the incubation period, the temperature was decreased to 50°C, and distilled water was added to remove the remaining P2S5. The product was dialyzed for four days in the dark at 10°C and pH~6 using Spectra/Pore dialysis membrane (MWCO: 1000 Da) and then lyophilized.
The free thiol content of α-CD-SH was determined by the Ellman's test (Ijaz et al., 2015).
Briefly, modified α-CD was dissolved in Ellman's buffer, and Ellman's reagent (5,5'-dithiobis (2-nitrobenzoic)) was added to the sample. After 90 min incubation at room temperature, the absorbance value at 450 nm was recorded using Tecan Spark microplate reader (M-200 spectrometer, Tecan, Grödig, Austria). The amount of free thiol groups of α-CD-SH was calculated using L-cysteine as a standard for the calibration curve. Disulfide bridges were quantified using Ellman's test after reducing the disulfide bonds in the product with sodium borohydride (NaBH4) (Laquintana et al., 2019).
To determine the water solubility, 5 mg of α-CD-SH was added to 100 µL of distilled water, and the sample was incubated under agitation (1200 rpm) at 25 °C and 37 °C for 24 h. J o u r n a l P r e -p r o o f Journal Pre-proof Undissolved α-CD-SH was removed by centrifugation at 13 400 rpm for 10 minutes, the supernatant was lyophilized, and α-CD-SH was gravimetrically quantified . pKa of α-CD-SH was determined by measuring UV absorption at stepwise raised pH values (Asim, Nazir, Jalil, Laffleur, et al., 2020). Briefly, α-CD-SH (0.5% m/v) was dissolved in 100 mM NaCl containing 1 mM HCl, and the pH of the solution was adjusted to different values using 0.01% NaOH. The absorbance at 242 nm at different pH values was measured. The term -log ( − ) / as a function of pH was used to calculate pKa. (Amax: absorbance of completely deprotonated thiols at high pH Ai: Absorbance of thiols at different pH values)

Cytotoxicity Studies
Cytotoxicity of native and thiolated α-CD on Caco-2 (human colorectal adenocarcinoma cells), HEK 293 (Human embryonic kidney), and MC3T3-E1 subclone 4 (MC3T3, commercial murine calvaria pre-osteoblast cell line -ATCC CRL2593) cells were determined via resazurin assay (Veider, Akkuş-Dağdeviren, Knoll, & Bernkop-Schnürch, 2022). For this test, 96 well plates with a seeding concentration of 2 X 10 5 cells/well of Caco-2 and HEK 293 cells were grown in MEM, while for MC3T3, 2 X 10 4 cells/well were seeded and grown in α-MEM (Fisher Scientific) (Gong et al., 2020). Cells were incubated at 37 °C under 5 % CO2 and 95 % relative humidity until confluency was reached. Three concentrations (0.5, 1, and 5 mg/mL) of native and thiolated α-CD dissolved in HEPES buffer (pH:7.4) were applied to cells. The growing medium without phenol red was used as a positive control, and 0.1% v/v Triton TM X-100 was used as a negative control. After 3 h and 24 h incubation, test solutions were removed, and cells were washed twice with buffer. Resazurin solution (Sigma-Aldrich) was applied to each well and incubated for 3 hours at 37°C under 5 % CO2 and 95 % relative humidity environment. The fluorescence of the supernatant was measured (λEx 540 nm, λEm 590 nm) using Tecan Spark microplate reader. Cell viability was calculated using the following formula, and the data were normalized for the buffer effect.

Determination of Hemolysis Activity
Since the cellular membrane of erythrocytes is susceptible and fragile, the hemolysis activity assay allows detailed conclusions about membrane damage and cell toxicity (Evans et al., 2013). Additionally, the cellular membrane of red blood cells (RBC) is frequently used to mimic the in vivo conditions after endosomal internalization. Hemolysis activity was determined spectroscopically, as previously described by Friedl et al. (2020) The percentage of hemolysis was calculated with the following formula:

Fluorescent Labeling of α-CDs
Fluorescent labeling of α-CDs was performed as previously reported by our group using dilauryl fluorescein (DLF) and coumarin-6 (Cou) dyes (Asim et al., 2018;Grassiri, Knoll, et al., 2022;Kali et al., 2023). Based on some previous studies, α-CDs are not forming real inclusion complex with Cou, but they J o u r n a l P r e -p r o o f Journal Pre-proof accumulate around the Cou micro aggregates (Ghosh, Das, Maity, Mondal, & Purkayastha, 2015). Firstly, 30 mg of thiolated and native α-CDs were suspended in 30 mL of demineralized water, and the pH was adjusted to 6.5. Then, 1 mL of two ethanolic dye solutions, 0.02% (m/v) DLF (molar ratio to CD 100:1) and 0.02% (m/v) Cou (molar ratio to CD 5:1), were separately added to α-CDs aqueous solutions and stirred at room temperature for 24 h. After that, the suspensions were filtered to remove free dye and lyophilized.

Thermomicroscopic characterization
DLF, Cou, thiolated and native α-CD were investigated with a BH2 polarizing microscope (Olympus, A) equipped with a Koffler hot-stage (Reichert, Austria) and an Olympus DP71 digital camera. The solid CDs were analyzed in cross-polarized light.

Differential Scanning Calorimetry
DSC thermograms were obtained with a DSC 7 (Perkin-Elmer, Norwalk, Ct, USA) using the Pyris 7. 0 software. Accordingly, 1-2 mg of samples (using a UM3 ultramicrobalance, Mettler, Greifensee, Switzerland) were weighed into one-pin hole aluminum capsules. The samples were heated using a rate of 10 °C min -1 with dry nitrogen. The instrument was calibrated for temperature with pure benzophenone (mp. 48.0 °C) and caffeine (236.0 °C), and the calibration of the energy was accomplished with indium (mp. 156.6 °C, the heat of fusion 28.45 J/g -1 )

Powder X-Ray Diffraction
The powder x-ray diffraction (PXRD) patterns for characterizing DLF, Cou, α-CD and α-CD-SH were obtained using an X'Pert PRO diffractometer (PANalytical, Almelo, NL) equipped with a θ/θ coupled goniometer in transmission geometry, programmable XYZ stage with well plate holder, Cu Ka1,2 radiation source with a focusing mirror, a 0.5° divergence slit and a 0.02° Soller slit collimator and a solid-state PIXcel detector. The patterns were registered at a tube OptiMEM) for 2 hours. However, since the dye was not toxic to the cells, the washing step after staining was not necessary. All fluorescence images were recorded under equivalent requirements. Image postprocessing was performed by ImageJ software: the yz-and xzprojections were performed from 5 XY images of an image stack at 0.2 µm z-step length.
Spectral unmixing was applied to eliminate fluorescence bleed between detection channels due to overlapping emission spectra. Furthermore, 2D image filtering was regulated by using a Gaussian filter.

Statistical Analysis
All experiments were performed at least three times. GraphPad Prism 5 (GraphPad Software, Inc., San Diego, CA, USA) was used to analyze the data. The one-way ANOVA and two-way Bonferroni multiple comparison test was applied, and statistical significance was defined as p<0.05 probability values.

Synthesis and Characterization of α-CD-SH
In this study, phosphorus pentasulfide in the presence of sulfolane and triethylamine was used for the thiolation of α-CD. The synthetic pathway is shown in Figure 1. The chemical structure of the α-CD-SH was confirmed by 1 H NMR and FT-IR analysis.
J o u r n a l P r e -p r o o f  Figure S1 shows the spectrum of the thiolated CD before the last purification, including solvent peaks, as afterward, due to the water absorption and the hydrogen bonding with the remaining hydroxyl groups, the peaks cannot be separately integrated. No significant difference in the 1 H NMR spectrum of fresh and old samples was found, referring to sufficient product stability. Due to the thiol hydroxyl exchange, the 13C NMR spectrum of thiolated α-CD show slightly shifted peaks for C-2, C-3, and C-6. Thiolation of α-CD was also confirmed with Fourier transform infrared (FT-IR) spectroscopy showing a weak peak that can be assigned to -S-H stretching (~2660 cm -1 ), together with a broad shoulder observed around 2333 cm -1 (Colthup, Daly, & Wiberley, 1990) ( Figure S2). The PXRD measurements confirmed that α-CD is present in hexahydrate form (Manor & Saenger, 1974) (Figure S3). The PXRD pattern of α-CD-SH showed that a different crystal structure is obtained upon thiolation, albeit the material exhibited a low crystallinity, as indicated by the low Bragg reflection intensities and the broad 'halo'. The latter is not surprising, as the synthesis comprised a final lyophilization step. The DSC thermogram ( Figure   S4) of the native α-CD exhibited a broad endotherm at temperatures below 150 °C, originating from water loss. Dehydration was accompanied by a structural collapse, i.e., the amorphous J o u r n a l P r e -p r o o f material was present at temperatures higher than 150 °C. From 275 °C onwards, applying a heating rate of 10 °C min -1 , decomposition of the α-CD occurred. The observations were additionally confirmed with hot-stage microscopy. In the case of α-CD-SH, the decomposition was seen at a slightly lower temperature than for α-CDs. Nevertheless, the α-CD-SH was thermally stable until above 200 °C (heating rate of 10 °C min -1 ). As described for the α-CDs, loss of crystallinity could be observed for α-CD-SH upon heating in the temperature range from 105 to 130 °C.
Ellman's and disulfide bond tests were used to determine the amount of free thiol and disulfide substructures on α-CD-SH. The free thiol concentration was 3270.83±213.46 µmol/g, while the disulfide bond concentration was 1376.09±43.13 µmol/g, that is 29.6% disulfide bond content for the product. Based on Ellman's and disulfide bond results, the degree of thiolation is about 32.69±1.62 %, which is in good agreement with the value calculated from the 1 H NMR spectrum.
The water solubility of α-CD-SH at 25 °C and 37 °C were 24.25 ± 0.29 mg/mL and 42.87 ± 0.46 mg/mL, respectively. This increase in solubility at higher temperature was previously shown for other CD derivatives (Jicsinszky, 2019).  Figure 2A illustrates the UV absorption of the α-CD-SH at indicated pH values. As the pH of the α-CD-SH solution increases, the concentration of thiolate anions increases, which can be quantified by a raised absorbance. As shown in Figure 2A, plotting pH vs. absorbance revealed a strong increase in thiolate anions above pH 6. A range of pKa of α-CD-SH is estimated from Figure 2B

J o u r n a l P r e -p r o o f
The results of cytotoxicity studies of α-CD and α-CD-SH on Caco-2, HEK 293, and MC3T3 cells are shown in Figure 3. Native α-CD did not show any significant cytotoxic effect during 3 h and 24 h of incubation (p˂0.05) at all concentrations applied to Caco-2 and MC3T3 cells ( Figure 3A, B, and 3E, F, respectively), whereas at a concentration of 5 mg/mL, a significant (p˂0.05) cytotoxic effect on HEK 293 was observed after 24 h of incubation ( Figure 3D).
In contrast, α-CD-SH showed a cytotoxic effect at a concentration of 5 mg/mL within 3 h in the case of all tested cells (Figure 3A, 3C, and 3E). Merely on MC3T3 cells, a low toxic effect was observed for concentrations of 0.5 and 1 mg/mL ( Figure 3E). Generally, α-CD-SH showed a time-and dose-dependent effect on all cells, leading to toxic effects from 1 mg/mL for Caco-2 ( Figure 3B) and HEK cells ( Figure 3D) and 0.5 mg/mL for MC3T3 ( Figure 3E J o u r n a l P r e -p r o o f

Determination of Hemolysis Activity
Native CD showed minor hemolysis activity at all applied concentrations after 3 h, while α-CD-SH exhibited significant hemolysis activity at 0.25 and 0.5 mg/mL concentrations. The results of this study are illustrated in Figure 4.
The hemolysis potential of materials was previously correlated to their interaction with endosomal membranes. As our results showed significantly higher membrane interactions of α-CD-SH compared to α-CD, the potential of α-CD for endosomal escape seems to be enhanced by thiolation. Hemolysis experiments showed that α-CD-SH has a high endosomal escape potential. However, this also indicates its limited hemocompatibility. Therefore, α-CD-SH may need to be used in mucosal drug applications or by injecting directly into target tissues. In contrast to the α-CD-SH tested in this study, however. some thiolated CDs with high hemocompatibility have been reported in previous studies (Asim et al., 2021;Fürst et al., 2023). and 24 h (B) of incubation at 37 °C. All results are shown as mean ± SD, n = 3. ns:not significant; **p˂0.005; *** p˂0.0005; **** p˂0.00005.

J o u r n a l P r e -p r o o f
Considering that α-CD-SH contains 1376 µmol/g disulfide bonds, the enhanced endosomal escape of this thiolated CD is in good agreement with previous studies. Kanjilal et al. have recently designed polymeric nanogels bearing disulfides at various positions that enabled endosomal escape, while without disulfide bounds, the polymeric nanogels did not show this result. Moreover, the exposure of disulfides on the outer surface of the nanogel further enhanced both cellular uptake and endosomal escape. The authors explained their observations by an endosomal escape mechanism based on thiol/disulfide exchange reactions between disulfidebearing nanogels and membrane thiols, forcing the packaging of endosomal membranes and disrupting membrane integrity (Kanjilal, Dutta, & Thayumanavan, 2022). Endosomal escape is a major bottleneck in the design of delivery systems to transport therapeutic biomolecules into cells (Pei & Buyanova, 2019). Thiolated delivery systems might therefore be a promising approach to provide endosomal escape. In our system, disulfide bonds are already present, and above the pKa of the thiolated CDs, it is expected that disulfide bonds develop with time, which can be advantageous for cellular uptake and endosomal escape.

Cellular Uptake Studies
To compare the cellular uptake of native and thiolated α-CD complexes with the fluorescent molecules DLF and coumarin-6 were prepared. The complex formation was confirmed by 1 H NMR and PXRD. In the 1 H NMR spectra, some broad peaks of the fluorescent dyes could be detected in the region of 1.0-3.0 ppm for the methyl and methylene protons and small peaks at the aromatic region (7.5 -8.0 ppm) (see Figure S5). In the PXRD diffractograms no peak positions of the fluorescent dyes were detectable ( Figure S3).
The PXRD measurements of DLF and coumarin-6 revealed that the two compounds were suggests that the two α-CD complexes may be structurally related to the α-CD succinic acid clathrate hydrate (Saouane & Fabbiani, 2016), an inclusion complex. The PXRD pattern of the α-CD-SH modified with the fluorescent molecules resulted in amorphous materials, as additionally confirmed with hot-stage microscopy and DSC. The absence of eutectic melting in the DSC curves of the labeled α-CD(-SH)s confirmed that neither DLF nor Cou were present in crystalline form. DLF and Cou exhibited sharp melting endotherms at 63.4°C and 208.2°C, respectively ( Figure S4). Decomposition of the labeled α-CD(-SH)s occurred above the dyes' melting points and at higher temperatures than for unlabeled α-CD(-SH)s. Therefore, it may be inferred that labeled α-CD(-SH)s exhibit, due to complexation, a higher thermal stability than the single component molecules.  Figure 5A, C, E). This was reflected in an increase in RMFI for α-CD-SH samples compared to native α-CDs for all tested cell lines (p˂0.01). α-CD-SH labeled with DLF showed approximately 20 times higher cellular uptake in Caco-2 cells than native α-CD (Figure 5 B). Using coumarin-6 in the same experimental setup led to a 5fold higher cellular uptake of the thiolated complex. α-CD-SH labeled with DLF (α-CD-SH/DLF) showed 3-fold higher cellular uptake than α-CD, while this difference was approximately 6.5-fold in samples labeled with coumarin-6 for HEK 293 cells ( Figure 5D).

J o u r n a l P r e -p r o o f
Similarly, a 14.5-fold increase was obtained for α-CD-SH/DLF and an 11-fold increase for α-CD-SH/Cou in MC3T3 (Figure 5F).
In order to better understand the mechanism involved in native and thiolated α-CD cellular uptake, we performed uptake experiments at 4°C to analyze the energy-dependent endocytosis pathway. Most endocytic pathways are energy-dependent processes that can be inhibited at low temperatures (Hsu et al., 2017) since cells consume less ATP at this temperature and active transport is blocked. Here, a dramatic decrease in cellular uptake at 4°C compared with 37°C was observed. Uptake of α-CD labeled with DLF by Caco-2 cells at 4°C was 4-fold lower compared to 37°C. This ratio was approximately 5-fold higher in samples labeled with coumarin-6 (Figure 5 B). For HEK 293 cells, uptake was 2.5 times higher in α-CD-SH/DLF, while 7-fold in α-CD-SH/Cou at 4°C (Figure 5 C and D). In the MC3T3 cell line, uptake of α-CD labeled with DLF and coumarin-6 was at 4°C 3-fold and 6-fold lower than at 37°C, respectively (Figure 5 F). It has been demonstrated in studies using fluorescently labeled CDs that the main internalization pathways of CDs are macropinocytosis and clathrin-dependent endocytosis (Fenyvesi et al., 2014;Plazzo et al., 2012;Réti-Nagy et al., 2015;Rusznyák et al., 2022). However, the cellular uptake of thiolated CDs has not been thoroughly investigated, and thiol-mediated cellular uptake is poorly understood (Laurent et al., 2021). For thiol-mediated cellular uptake, J o u r n a l P r e -p r o o f mechanisms such as endocytosis, fusion, and direct translocation across the plasma membrane to the cytosol have been proposed. Thiol modification increases the cellular delivery efficiency of the carrier system by enabling interactions with cell surface thiols (Cheng et al., 2021).
Recently, a thiol-reactive chloride channel CLIC1, an epithelial growth factor receptor EGFR, a transferrin receptor protein-disulfide isomerase, and Ca-activated scramblase TMEM16F proteins have been identified in detail as possible cell surface targets of thiol groups (Laurent et al., 2021). It has been revealed that thiol-modified carrier systems show higher cell permeability-enhancing effects than their non-thiolated counterparts due to thiol-disulfide exchange reactions (Li T & Takeoka S, 2014;Zhang, Qin, Kong, Chen, & Pan, 2019). These findings support the data obtained in our study. Considering the efficacy of thiolation on α-CD on alternative thiol-mediated cellular uptake and endosomal escape, the current strategy can be applied to chemically diverse biomolecules, as well as used to target multiple cell types.
To support flow cytometry results, we examined cellular uptake by confocal microscopy. As shown in Figure 6, cells treated with α-CD-SH/DLF exhibit much higher fluorescence compared to native α-CD. This supports the cellular uptake data obtained by flow cytometry.
Moreover, in the case of DLF, fluorescence was also found in the nucleus of cells. This coincides with hemolysis experiments, confirmed that thiolation promotes endosomal escape.
An additional indication for endosomal escape is the diffuse fluorescence observed in these cells rather than a dotted fluorescence that would indicate endocytosis without endosomal escape. Confocal microscopy results showed furthermore that native and thiolated α-CDs exhibited a significant fluorescence in Caco-2 and HEK cells (Figure 6A and B, respectively).
When comparing confocal images of samples labeled with coumarin-6 with those labeled with DLF, results obtained for thiolated α-CDs were correlated with those of flow cytometry for all cells. I.e., uptake followed the pattern α-CD-SH/Cou = α-CD-SH/DLF for Caco-2 (

CONCLUSION
In this study, we demonstrated the impact of thiol groups on CD-based drug carriers on cellular uptake and endosomal escape for the first time. The formation of thiolated α-CD was confirmed by FT-IR and 1 H NMR spectroscopy, and the crystalline state of the materials was identified by PXRD and thermal analysis. Cellular uptake experiments on Caco-2, HEK 293, and MC3T3 cells showed that thiolated α-CD could enter cells at a greater extent than native CD, with uptake driven by energy-dependent endocytosis. The results confirmed the hypothesis that thiolated α-CD increases cellular uptake. Additionally, the hypothesis that endosomal escape is promoted by thiolation was generated and confirmed by confocal microscopy and hemolysis assays. The striking effects of thiol modifications on cellular uptake and endosomal escape in CDs can be used in the pharmaceutical field to deliver hydrophobic drugs into target cells. Therefore, this study will provide strong motivation for the design of α-CD-SH based drug delivery systems in the future.