Design of Nanohydrogels for Targeted Intracellular Drug Transport to the Trans‐Golgi Network

2.1. Expression of hRS1‐Reg(S20E) and mRS1‐Reg(S19E) in E. coli

E. coli BL21 STAR cells were transformed by inserting the pET21a plasmid containing the genes for hRS1‐reg(S20E) or mRS1‐reg(S19E). After recombinant expression, cell harvest and disruption, the supernatant following a high‐speed centrifugation was used for protein affinity purification via His‐Tag. Figure  1A shows a SDS‐PAGE analysis of fractions, collected in the course of purification with Ni²⁺‐NTA‐Agarose. In the eluates, the purified peptides hRS1‐reg(S20E) and mRS1‐reg(S19E) showed a molecular weight of approximately 9 kDa, which corresponds well to the predicted MW of 9.3 kDa. For some follow‐up experiments RS1‐reg peptides were labeled via lysine residues with NHS‐ATTO488, NHS‐Alexa488 or NHS‐Alexa647 conjugates according to manufacturer´s protocol.

2.2. RS1‐Reg Peptides Are Taken Up via Polyglycidol‐Based Nanohydrogels into Caco‐2 Cells and Small Intestinal Mucosa of Mouse, Pig, and Human

Redox‐sensitive nanohydrogels based on thiol‐functionalized, hydrophilic polyglycidol prepolymers were formed by crosslinking via oxidative coupling in inverse miniemulsion or by inverse nanoprecipitation polymerization.^{[ 12} , ^{28 ]} In this context the hRS1‐reg(S20E) peptide or the mRS1‐reg(S19E) peptide and for some experiments also the cell‐penetrating TAT‐peptide were covalently bound by cysteine residues to the nanogel matrix in a hydrophilic, stabilizing environment (see scheme in Figure 1B; for further chemical details of the components see Figure S5, Supporting Information). Nanogel concentrations and sizes were measured using nanoparticle tracking analysis (NTA). Furthermore nanogel size distribution was measured by DLS and ζ potential by electrophoretic light scattering using a Zetasizer Nano ZSP (Malvern Instruments) instrument (see Figure S2, Supporting Information). Depending on the method (NTA or DLS) and nanogel charge a monodisperse size distribution (PdI < 0,2) with mean diameter in the range of 200 –280 nm was obtained for nanogels from inverse miniemulsions, and 250– 300 nm for nanohydrogels formed by nanoprecipitation. Interestingly, the ζ potential values showed no significant difference and were in the range of ‐20 to ‐25 mV. Cryo‐SEM analysis confirmed that nanogels from both manufacturing methods are present as individual colloidal particles in solution. To study the encapsulation efficiency regarding RS1reg or rather the TAT‐peptide, nanohydrogels were prepared using defined amounts of ATTO488‐RS1‐reg in presence or absence of Alexa647‐labeled TAT peptide. After washing and reductive degradation of the nanogels in presence of DTT the amount of released peptides was determined by fluorescence measurements, employing ATTO488‐RS1reg or rather Alexa647‐TAT calibration grades. The encapsulation efficiency (mass percentage of the drug loaded relative to the drug fed) of RS1reg without TAT or in presence of TAT was 83% or rather 78%. For TAT an encapsulation efficiency of 93 percent was calculated. The loading efficiency (mass percentage of the drug loaded relative to the dried nanogel) of TAT was 9%. To investigate whether the payload TAT was bound merely in the peripherie of nanogel particles or coupled to the matrix throughout the entire particle volume, the TAT peptide was fluorescently labeled with Alexa 647 and incorporated into nanoprecipitation nanogels. The sample, and especially particles in it with a much higher diameter than the mean to get a clear result, were analyzed using a ZEISS LSM 900 confocal microscope with Airyscan 2 in Airyscan mode with a laser at a wavelength of 640 nm. Z‐stacks of large particles revealed presence of the coupled TAT throughout the entire particle volume (Figure S6, Supporting Information).

Stability analysis of nanogels out of precipitation polymerization and inverse miniemulsion was performed by DLS and fluorescence measurements (Figure S3, Supporting Information ). DLS measurements of nanohydrogel stability were done after 1:100 dilution in 100 × 10⁻³ m HCl (pH 1), PBS (pH 7.4) and 20 × 10⁻³ m Tris/HCl (pH 8.5) for 0–60 min, mimicking pH conditions in stomach (pH 1), in small intestinal lumen and mucosa after fasting (neutral pH range, exact value varies from patient to patient, animal to animal, study to study and position in small intestine) and in presence of digestive juices within small intestine (pH 7.5–8.5) after a meal.^{[ 30} , ³¹ , ³² , ³³ , ³⁴ , ³⁵ , ³⁶ , ^{37 ]} While the nanogels were quite stable for 1 h under acidic conditions (pH 1) in the stomach and neutral pH conditions (pH 7.4) of small intestinal lumen and mucosa after fasting, nanogel size decreased significantly under slightly more alkaline conditions (pH 8.5) as within small intestine in presence of digestive juices. DLS analysis of nanohydrogel stability was conducted for 0, 10, 20, and 30 min after 1:100 dilution in PBS in absence or presence of 10 × 10⁻³ m DTT, mimicking the reducing conditions in the cytoplasm. Under reducing conditions the nanogels rapidly degraded. Furthermore, fluorescence measurements of the stability of precipitation polymerization based nanohydrogels loaded with ATTO488‐RS1reg were carried out for 10, 20, and 30 min after 1:100 dilution in PBS with or without DTT under liberation of the labeled peptide. In presence of DTT a pronounced peptide release was detectable.

To get proof for differences regarding stability and cross‐linking density of miniemulsion nanogels versus nanoprecipitation nanogels, we performed time‐dependent turbidity measurements of samples of both nanogel types with equal particle density (1 × 10¹¹ particles mL⁻¹) following supplementation with 10 × 10⁻³ m DTT (Figure S7,  Supporting Information) for 1 h. The nanogels without DTT used as controls showed stable transmission over the 60 min, whereby the transmission value of the nanogels from precipitation polymerization was lower than that of the nanogels from inverse miniemulsion. The control approach with 10 × 10⁻³ m DTT alone showed a stable transmission of about 1 over the whole measurement period. The sample of nanogels from inverse emulsion with DTT showed an earlier, faster increase in transmission until the plateau value of slightly less than 1 was reached than the analog sample with precipitation nanogels. This is consistent with the assumption that the degree of disulfide cross‐linking is lower in the nanogels from inverse emulsion, the particles are therefore less stable and decompose more easily or more quickly in the presence of DTT than in case of nanoprecipitation nanogels.

To demonstrate that Caco‐2 cells can serve as a model for enterocytes of the intestinal mucosa in respect to peptide uptake via nanohydrogels in subsequent immunofluorescence experiments, Caco‐2 cells and segments of small intestinal mucosa of mouse, pig and human were preincubated with nanohydrogel, harboring ATTO488‐labeled RS1‐reg peptide. Confocal laser scanning microscopy analysis confirmed uptake of the green fluorescent peptide in all cases (Figure 1C).

Confocal LSM studies (Figure S4,  Supporting Information) to localize the uptake of ATTO488‐RS1‐reg via nanohydrogels out of inverse miniemulsion administered by gavage to a group of mice on sections of small intestinal segments one hour after administration showed strong accumulation in the initial, proximal region of the small intestine, 3 to 10 cm after gastric outlet. A weak signal was seen in the segment 15 cm after the gastric outlet, no more uptake in the later, distal segments (20 and 25 cm after the gastric outlet). No signal of the ATTO488‐RS1‐reg peptide was detected in stomach. These findings support the model concept that the uptake of the cargo occurs after gastric passage in the small intestine, where the nanogels develop mucoadhesive properties under the alkaline conditions, adhere to the mucosa of the small intestine and are absorbed there.

2.3. Glucose Uptake Is Inhibited in Caco‐2 Cells and Mouse Small Intestinal Segments to the Same Degree by Nanogels Prepared in Inverse Miniemulsion and by Inverse Nanoprecipitation

In previous studies inhibition of SGLT1 mediated glucose uptake in small intestine of RS1 knockout mice (RS1^−/−) was demonstrated by gavage of RS1^−/− mice with nanohydrogels prepared in inverse miniemulsions, containing the cell‐penetrating TAT‐peptide with or without (control) mRS1‐reg(S19E). Three hours later phlorizin‐inhibited AMG uptake was measured into segments of everted jejunum, that was downregulated by mRS1‐reg(S19E).^{[ 25 ]} In this work, we studied in more detail the influence on glucose uptake of nanohydrogels containing hRS1‐reg(S20E) and mRS1‐reg(S19E) with or without TAT‐peptide, prepared in inverse miniemulsion versus inverse nanoprecipitation, into segments of everted RS1^−/− mouse small intestine (mRS1‐reg(S19E)) and differentiated Caco‐2 cells (hRS1‐reg(S20E)), which represent an established model for human enterocytes for activity assays and express SGLT1.^{[ 38 ]} For that mouse jejunal segments or Caco‐2 cells were preincubated with the nanohydrogels or with the RS1‐reg‐peptides alone (as controls) and then phlorizin‐inhibited uptake of 10 × 10⁻⁶ m radiolabeled AMG was measured. AMG uptake was not altered after incubation with hRS1‐reg(S20E) or mRS1‐reg(S19E) peptides alone or by nanogels without those peptides, but it was downregulated to the same degree, respectively, by about 50% after treatment with mRS1‐reg(S19E) or hRS1‐reg(S20E), transported in nanohydrogels prepared in inverse miniemulsions and by nanoprecipitation, in presence or absence of TAT peptide (Figure  2A). Using a dilution series (500, 100, and 10 µg mL⁻¹) of nanogels containing hRS1‐reg(S20E) out of nanoprecipitation without TAT for preincubation with Caco‐2 cells, we demonstrated dose‐dependency of the inhibitory effect on AMG uptake (Figure 2B). These data suggest that the way of manufacturing the nanogels as well as the presence or absence of the cell‐penetrating TAT peptide did not influence the inhibitory effect of the RS1‐reg peptides on glucose uptake in mouse small intestine and Caco‐2 cells. Furthermore, the results also indicate, that Caco‐2 cells might be a good model for peptide uptake via nanohydrogels in small intestinal mucosa in the subsequent fluorescence experiments. Besides the inhibitory effect of the RS1‐reg loaded nanogel on AMG uptake in Caco2 cells and ex vivo in mouse small intestine presented here, this inhibition has also already been shown in vivo in mice.^{[ 25 ]}

2.4. Different Ways of Preparing Nanohydrogels out of the Same Chemical Compounds Result in Different Intracellular Localization of the Cargo RS1‐Reg

For investigating the influence of the nanogel preparation method and the presence or absence of the cell penetrating TAT peptide as transfection reagent on the intracellular targeting and localization of the example cargo RS1‐reg, Caco‐2 cells were preincubated with nanogels. The latter were formed either in inverse miniemulsions or by inverse nanoprecipitation and contained ATTO488 labeled hRS1‐reg(S20E) with or without TAT. The corresponding experiments have been performed by confocal laser scanning microscopy analysis (Leica TCS SP8). Nuclei were stained with Hoechst 33342 dye and the green fluorescence of the RS1‐reg peptide was detected (Figure  3 ). In case of the nanogel, which has been produced in inverse miniemulsion without TAT, a diffuse, homogenous distribution of the RS1‐reg signal in the cytosol of Caco‐2 cells could be observed. No signal was detected in the nuclei (Figure 3A). These data suggest endosomal escape of ATTO488‐hRS1‐reg(S20E) following uptake by endocytosis of the nanogel into Caco‐2 cells. The peptide distributes evenly in the cytosol and finally also reaches its site of action at the trans‐Golgi network and inhibits the release of SGLT1 containing vesicles, resulting in the observed inhibition of glucose uptake (Figure 2). In the approach of the nanogel out of inverse emulsion containing the TAT‐peptide, only a minor part of RS1‐reg signal was found in the cytosol. The majority of hRS1‐reg was localized in the nuclei of Caco‐2 cells, suggesting that TAT carried also hRS1‐reg with it into the cell nuclei (Figure 3B). This steering characteristic of TAT was also described for other compounds.^{[ 39 ]} Figure 3C,D show a similar, selective accumulation of ATTO488‐hRS1‐reg(S20E), taken up by nanogels formed by inverse nanoprecipitation with or without TAT‐peptide, in defined cytosolic vesicles. These findings could indicate an impaired endosomal escape of hRS1‐reg, which was not influenced by TAT. A direct comparison between CLSM measurements of the uptake of Alexa647‐labeled RS1‐reg via nanogels out of inverse emulsion versus nanoprecipitation without TAT is shown in Figure 3E. In case of the nanogel formed in inverse miniemulsions the Alexa647‐RS1‐reg signal is evenly distributed in the cytosol of Caco‐2 cells, whereas the RS1‐reg peptide transported by the nanohydrogel out of nanoprecipitation accumulates selectively in distinct vesicular structures within the cytosol. These results indicate that different ways of preparing nanohydrogels out of the same chemical compounds result in different intracellular localization of the cargo RS1‐reg.

2.5. hRS1‐Reg Follows the Endosomal Pathway after Uptake via Endocytosis Using Nanohydrogels from Precipitation Polymerization and Thus Reaches Its Site of Action at the TGN via Targeted Intracellular Transport

To identify the cytosolic vesicular structures, in which hRS1‐reg(S20E) selectively accumulates, we performed structured illumination microscopy (SIM) measurements, that offer a higher resolution than CLSM analysis. Caco‐2 cells were grown on coverslides or 8 Well Lab‐Tek chamber slides, pretreated with nanogels harboring Alexa488‐labeled hRS1‐reg(S20E) peptide without TAT, fixed with paraformaldehyde and permeabilized. After incubation with primary antibodies against marker proteins of endosomes (EEA1), lysosomes (Lamp‐1), endoplasmatic reticulum (Calnexin), the trans‐Golgi network (TGN46) and against the enzyme ODC1 the cells were treated with ATTO643‐labeled secondary antibodies. Cell nuclei were stained with Hoechst 34580 and SIM analysis (Zeiss Elyra S.1 SIM, Carl Zeiss Microscopy GmbH, Jena) was performed. In case of the markers EEA1, Lamp‐1, TGN46 and ODC1 a partial colocalization with the signal of Alexa488‐labeled RS1‐reg peptide were observed. No colocalization was detected with the marker for endoplasmatic reticulum, calnexin (Figure  4 ). To validate the data with higher accuracy, the SIM measurements were repeated with embedded samples for the approaches with the markers EEA1, Lamp‐1, TGN46 and ODC1, enabling a precise channel alignment. SIM‐analysis revealed the same result, a partial colocalization with Alexa488‐hRS1‐reg(S20E), respectively (Figure  5 ). Quantitative colocalization analysis was performed, employing all pictures of 4 stacks regarding the green signals of Alexa488‐RS1‐reg (Ch2) and red fluorescence of marker proteins of endosomes (EEA1), lysosomes (Lamp‐1), trans‐Golgi network (TGN46) and the enzyme ODC1 (Ch1, respectively). Pearson's coefficients were calculated using a custom‐written ImageJ macro (see Supporting Information). A high positive linear correlation between the respective characteristics under consideration is indicated by Pearson coefficients > 0.7. As a control, the colocalization between the signals of the marker proteins (Ch1) and nuclei (Ch3) was investigated. The resulting Pearson coefficients of approximately 0 indicate no dependence of the two characteristics on each other, as expected.

In summary, the results suggest (Figure  6 ) that in case of nanogels, formed in inverse miniemulsions, after cellular uptake by endocytosis the endosomal escape of RS1‐reg succeeds, resulting in an even distribution in the cytosol and also in reaching its site of action at the TGN. On the other hand, the endosomal escape of RS1‐reg, taken up via nanogels from nanoprecipitation, is impaired. Due to this process, we detected the colocalization with the endosomal marker EEA1. RS1‐reg passes from the early endosomes to the late endosomes, where it can be transported via vesicle exchange to the trans‐Golgi network. The inhibition of glucose uptake by RS1‐reg via the enzyme ODC1 by blocking the release of SGLT1‐containing vesicles from the TGN could be detected in activity assays (see Figure 2). Furthermore, we observed both partial colocalization with the TGN marker TGN46 and with ODC1. The residual hRS1‐reg in late endosomes is degraded after fusion with lysosomes in the resulting endolysosomes. In this context, a partial colocalization with the lysosomal maker Lamp‐1 was detected. After cellular uptake via nanohydrogels, formed by nanoprecipitation, RS1‐reg follows the endosomal pathway and reaches its site of action at the TGN. It is an example of a targeted intracellular drug transport, where endosomal escape is not required. Thus we developed a bespoken nanogel for specific organelle targeting of a drug to the TGN.

2.6. Nanohydrogels out of Nanoprecipitation Exhibit a Higher Stability in Terms of Form and Structure

To further corroborate the assumption that the impaired endosomal escape of hRS1‐reg after endocytosis of nanohydrogels formed by nanoprecipitation may be due to a higher stability of these nanogels as a result of a higher degree of cross‐linking compared to nanogels prepared from inverse miniemulsion, we studied adsorbed nanogel colloids by PeakForce Tapping mode AFM in air. The spin‐coated miniemulsion nanogels formed film‐like structures on mica. No distinct nanogel particles could be observed within this structure. By contrast, for the precipitation method, distinct nanogel particles could be clearly resolved within in the larger aggregates. Cryo‐SEM analysis as well as DLS measurements demonstrated that nanogels from both manufacturing methods were initially present as individual colloidal particles in the suspensions (Figure S2,  Supporting Information). The AFM measurements revealed the different stability of the two types of nanogels. Nanogels prepared by the inverse precipitation method are more stable due to the higher degree of cross‐linking.

Animals

C57BL/6 wild‐type mice were kept in temperature controlled environment with 12 h light/dark cycles. The mice were provided with standard chow and water ad libitum (Altromin 1324; Altromin Spezialfutter GmbH, Lage, Germany). Animals were handled in compliance with the guidelines of the University of Würzburg and German laws and ethical approval was given by the government of Upper Bavaria (file number: 55.2.2‐2532‐2‐704).

Materials

α‐Methyl‐D‐[14C]glucopyranoside ([14C]AMG) (11.1 GBq per mmol) was obtained from American Labeled Chemical Inc. (St. Louis, MO). Alloxan‐monohydrate, Saponin (from Quillaja bark), Span 80 and Tween 80 were obtained from Sigma‐Aldrich GmbH (Taufkirchen, Germany). DTT was from Applichem (Darmstadt, Germany). Soluene‐350 was obtained from PerkinElmer Inc. (Waltham, MA, USA). TAT‐peptide (CGRLLRRQRRR) was delivered by Peptides International (Louisville, Kentucky, USA). Nickel(II)‐charged nitrilotriacetic acid‐agarose (Ni²⁺‐NTA‐agarose) was purchased from QIAGEN (Hilden, Germany). Escherichia coli strain BL21 Star(D3) (cat. no. C6010‐03) was purchased from Invitrogen/Thermo Fisher Scientific (Carlsbad, CA, USA) and plasmid pET21a (cat. no. 69740‐3) from Merck Life Science GmbH (Eppelheim, Germany). Other chemicals were purchased as described.^{[ 16} , ²⁵ , ^{26 ]}

Antibodies

Rabbit polyclonal antibody against TGN46 was obtained from Novus Biologicals (Centennial, Colorado, USA) and rabbit monoclonal antibody against EEA1 was from Cell Signaling Technology (Danvers, MA, USA). Rabbit polyclonal antibody against ODC1 and mouse monoclonal antibody against Calnexin were purchased from ThermoFisher (Waltham, MA, USA). Rabbit polyclonal anti‐human‐LAMP‐1 antibody was delivered by Abcam (Cambridge, UK). Goat‐anti‐mouse‐ATTO643 antibody was from Sigma‐Aldrich GmbH (Taufkirchen, Germany). Polyclonal donkey‐anti‐Rabbit antibody was obtained from Jackson ImmunoResearch Europe Ltd (Ely, Cambridgeshire, UK) and fluorescently labeled with NHS‐ATTO643 conjugate (ATTO‐TEC GmbH, Siegen, Germany) according to manufacturer´s protocol.

Cloning

Cloning of hRS1‐reg(S20E) and mRS1‐reg(S19E) into the plasmid pET21a was performed as described earlier.^{[ 25 ]}

Expression and Purification of hRS1‐Reg(S20E) and mRS1‐Reg(S19E)

E. coli bacteria (strain BL21 Star) were transformed with pET21a plasmids containing genes for His‐tagged hRS1‐reg(S20E) or mRS1‐reg(S19E) and grown at 30 °C to mid‐log phase. Protein expression was induced by adding 1 × 10⁻³ m isopropyl‐1‐thio‐b‐D‐galactopyranoside, and bacteria were subsequently grown for 3 h at 30 °C. After 15 min centrifugation at 6000 × g, bacteria were washed and suspended in 20 × 10⁻³ m Tris‐HCl (pH 8.0) containing 500 × 10⁻³ m NaCl, 50 × 10⁻³ m imidazole and an EDTA‐free protease inhibitor cocktail (Roche, Basel, Switzerland). The cells were lysed by sonication at 4 °C, and cellular debris was removed by 1 h centrifugation at 100 000 × g. For protein purification, the supernatants were mixed with Ni²⁺‐NTA‐Agarose (Qiagen, Hilden, Germany), and the beads were incubated for 1 h under rotation and poured into an empty gravity flow column. After washing with 20 × 10⁻³ m Tris (pH 8.0) containing 500 × 10⁻³ m NaCl and 50 × 10⁻³ m imidazole, protein was eluted with 20 × 10⁻³ m Tris (pH 8.0) containing 500 × 10⁻³ m NaCl and 500 × 10⁻³ m imidazole. Fractions containing purified protein were pooled and dialyzed against PBS. For some experiments RS1‐reg peptides were labeled with NHS‐ATTO488 (ATTO‐TEC GmbH, Siegen, Germany), NHS‐Alexa488 (ThermoFisher Scientific, Waltham, MA, USA) or NHS‐Alexa647 (ThermoFisher Scientific, Waltham, MA, USA) conjugates according to manufacturer´s protocol.

SDS‐Polyacrylamide Gel Electrophoresis (SDS‐PAGE)

SDS polyacrylamide gel electrophoresis (5% stacking gel and 12% separating gel) was performed as described by Laemmli in a Mini‐PROTEAN II slab gel apparatus (Bio‐Rad, Hercules, California, USA) for 60 min at 200 volt.^{[ 79 ]} For standard sample preparation the protein samples were pretreated for 30 min at 37 °C in 60 × 10⁻³ m Tris‐HCl, pH6.8, 100 × 10⁻³ m dithiothreitol, 2% (w/v) SDS, and 7% (v/v) glycerol and then separated by SDS‐PAGE. The gels were stained with a colloidal coomassie brilliant blue staining solution according to Candiano et al.^{[ 80 ]}

Preparation of Nanohydrogels in Inverse Miniemulsion

The preparation of nanohydrogels in inverse miniemulsion was carried out on the basis of a previously published protocol.^{[ 25 ]} Inverse miniemulsions were formed from an organic phase consisting of 625 µL hexane, which contained 14.1 mg Span 80 and 4.7 mg Tween 80, and an aqueous phase of 62.5 µL PBS with 25 mg thiol‐functionalized linear polyglycidol prepolymer, 44 µg RS1‐reg polypeptide with N‐terminal cysteine and in some cases 125 µg of the cell‐penetrating TAT peptide (CGRLLRRQRRR) or without cell‐penetrating peptide.^{[ 29 ]} For the preparation of empty control nanohydrogels, the RS1‐reg polypeptide was omitted. To form the inverse miniemulsion, the organic and aqueous phases were combined, stirred and treated by ultrasonication for 60 s with a Branson sonifier 250 (duty cycle of 40% and output control of 10%). During the sonication, the reaction vessel was cooled with an ice bath. By adding 30 µL of 50 × 10⁻³ m alloxan, the formation of the nanohydrogel and the covalent attachment of the cell‐penetrating peptide and the RS1‐reg peptide, respectively, were achieved by oxidative coupling under formation of a disulfide bridge network. For this purpose, the mixture was immediately treated with ultrasonification for 60 s after alloxan addition and then incubated for 25 min at room temperature with stirring. The oxidation was then stopped by acidification and the formed nanohydrogels were separated by centrifugation. The organic phase was removed, the aqueous phase containing the nanohydrogels was washed twice with n‐hexane and twice with tetrahydrofuran to remove detergents and unreacted prepolymer. Remaining organic solvents and acid were finally removed by multiple dialysis against water. Suspensions of the purified, loaded nanohydrogels were stored in millipore water at 4 °C for further use for up to 4 weeks.

Preparation of Nanohydrogels via Inverse Nanoprecipitation Polymerization

Nanogels were produced by nanoprecipitation polymerization of the polymer solution in acetone.^{[ 13 ]} In the standard procedure, 6 mg thiol‐functionalized, linear polyglycidol was dissolved in 90 µL PBS with 44 µg RS1‐reg polypeptide with N‐terminal cysteine and in some cases with 125 µg TAT‐peptide or without cell‐penetrating peptide and precipitated in 45 mL acetone by rapid addition. For the preparation of empty control nanohydrogels, RS1‐reg polypeptide was omitted. After 30 min, 30 µL of alloxane solution (32 mg mL⁻¹ in water) was added and oxidation was carried out for another 30 min, followed by the addition of 4 mL water. The acetone still contained in the sample was evaporated overnight in a fume hood and the formed nanogels were then washed three times with water by centrifugation (16.100 x g, 15 min). Suspensions of the loaded nanohydrogels were stored at 4 °C for up to 4 weeks.

Determination of Particle Size Distribution, Particle Concentration, and ζ Potential

Measurements were performed using the nanoparticle tracking analysis (NTA) device (NS500, Nanosight). Samples were diluted in a ratio of 1:1000 and 1:10 000 in a total volume of 1 mL of Millipore water. Each sample was measured three to five times (camera level 16, threshold 4−5 with automated settings) at 23 °C. The particle concentration and size distribution of the sample were calculated as a mean value overall measurements. ζ potential measurements were performed by electrophoretic light scattering and size determination by dynamic light scattering (DLS) using a Zetasizer Nano ZSP (Malvern Instruments) instrument. Measurements were performed at 25 °C in water.

Stability Analysis of Nanogels out of Precipitation Polymerization and Inverse Miniemulsion by DLS and Fluorescence Measurements

DLS measurements of nanohydrogel stability after 1:100 dilution in 100 × 10⁻³ m HCl (pH 1), PBS (pH 7.4) and 20 × 10⁻³ m Tris/HCl (pH 8.5) were performed for 0–60 min, mimicking pH conditions in stomach (pH 1), in small intestinal lumen and mucosa after fasting (neutral pH range, exact value varies from patient to patient, animal to animal, study to study and position in small intestine) and in presence of digestive juices within small intestine (pH 7.5–8.5) after a meal.^{[ 30} , ³¹ , ³² , ³³ , ³⁴ , ³⁵ , ³⁶ , ^{37 ]} DLS analysis of nanohydrogel stability was carried out for 0, 10, 20, and 30 min after 1:100 dilution in PBS in absence or presence of 10 × 10⁻³ m DTT, mimicking the reducing conditions in the cytoplasm. Fluorescence measurements of the stability of precipitation polymerization based nanohydrogels loaded with ATTO488‐RS1reg for 10, 20, and 30 min after 1:100 dilution in PBS with or without DTT under liberation of the labeled peptide were done with a SPARK Multimode Microplate Reader (TECAN Trading AG, Switzerland), Excitation wavelength: 500 nm, Emission wavelength: 520 nm.

AFM Imaging

AFM imaging of miniemulsion and nanoprecipitation nanogels was performed using a Dimension Icon equipped with a Nanoscope V controller (Bruker, Santa Barbara, CA) by means of PeakForce Tapping mode in air. Prior to each AFM measurement the aqueous nanogel suspension was vortexed and treated in an ultrasonic bath (120 W) each for 180 s and subsequently spin‐coated (120 s; 2100 rpm, EC101DT, Headway Research Inc., Garland, TX) onto a freshly cleaved mica sheet (#71856‐01; Electron Microscopy Sciences, Hatfield, PA). The surface of the dry mica sheet was imaged using silicon nitride cantilevers (ScanAsyst‐Air, Bruker) with a nominal spring constant of 0.4 N m⁻¹. A PeakForce Tapping amplitude of 150 nm and a frequency of 2 kHz were used. Processing of the imaging data was performed with NanoScope Analysis 1.50 (Bruker).

Turbidity Measurements Regarding Nanogel Stability

To investigate differences concerning stability and cross‐linking density of miniemulsion nanogels versus nanoprecipitation nanogels, time‐dependent turbidity measurements of samples of both nanogel types with equal particle density (1 × 10¹¹ particles mL⁻¹) following supplementation with 10 × 10⁻³ m DTT (Figure S7, Supporting Information) in a Tecan Spark 20 M multimode microplate reader (Tecan Group Ltd., Männedorf, Switzerland) at a wavelength of 500 nm at 25 °C for 1 h. Nanogels without DTT and a 10 × 10⁻³ m DTT solution served as controls.

Cell Culture

Caco‐2 cells were grown and routinely maintained at 37 °C in EMEM containing 10% (v/v) FBS and penicillin (100 U mL⁻¹)–streptomycin (100 µg mL⁻¹) in an atmosphere of 5% CO₂ and 90% relative humidity. The cells were harvested with Accutase and seeded out on 8 Well Lab‐Tek chamber slides or coverslips for immunostaining or on 12 well plates for uptake measurements. The culture medium was replaced every 2–3 d. Caco‐2 cells were used between the passage 43 and 70.

Uptake Measurements with Caco‐2 Cells

The non‐metabolizable glucose analog α‐methylglucose (AMG), a specific substrate for Na⁺/glucose cotransporters was used to measure cotransport activity in Caco‐2‐cell monolayers.^{[ 81 ]} Cells were seeded in 12 well plates, grown to confluence and further cultivated for 18 d, at which time they were fully differentiated. Transport assays were performed at 37 °C in PBS as transport medium, supplemented with 10 × 10⁻⁶ m α‐methyl‐D‐glucopyranoside (AMG) containing tracer amounts of [14C]AMG (11.1 GBq per mmol, American Labeled Chemical, St. Louis, MO, USA). To evaluate nonspecific uptake of AMG SGLT1 was selectively inhibited by 1 × 10⁻³ m phlorizin in controls. Cells were preincubated without or with 0.5 mg mL⁻¹ nanogel for 30 min (nanogels with TAT‐peptide) or 120 min (nanogels without TAT). Thereafter AMG‐uptake was performed for 10 min and then stopped by adding ice‐cold transport medium containing 1 × 10⁻³ m phlorizin and incubation for 5 min. The monolayers were washed twice with the same buffer and solubilized in 2% (w/v) sodium dodecyl sulfate (SDS). Triplicate cultures were assayed for radioactivity by liquid scintillation counting. Counts were normalized in relation to 1 for the approach with Caco‐2 cells without any addition.

Uptake Measurements with Everted Small Intestinal Segments of Mice

For preparation of small intestinal segments, mice were killed by cervical dislocation, and the small intestine was removed. After perfusion with PBS, the small intestine was everted using a steel rod. Eight 1‐cm long segments of the proximal jejunum were isolated and washed with PBS. The small intestinal segments and mucosal pieces were preincubated at 37 °C with PBS in the absence (control) and presence of 0.5 mg mL⁻¹ nanogel for 30 min (nanogels with TAT‐peptide) or 120 min (nanogels without TAT). Thereafter the intestinal samples were washed thoroughly with PBS. For measurements of α‐methyl‐D‐glucopyranoside (AMG) uptake, the samples were incubated for 2 min at 37 °C with PBS containing 10 × 10⁻⁶ m AMG traced with [14C]AMG. The incubations were performed in the absence of phlorizin or in the presence of 1 × 10⁻³ m phlorizin. Uptake was stopped with ice‐cold PBS containing 1 × 10⁻³ m phlorizin. The segments were washed twice with the same buffer, solubilized with Tissue Solubilizer, Soluene‐350 (PerkinElmer, Waltham, MA), and their radioactivity was determined by liquid scintillation counting. Counts were normalized in relation to 1 for the approach with a small intestinal segment without any addition.

Staining of Small Intestinal Segments of Mice, Pig and Human with ATTO488‐Labeled RS1‐reg for Confocal Laser Scanning Microscopy (CLSM) Analysis

After killing of mice by cervical dislocation, removal and perfusion of small intestine 1 cm long segments of the proximal jejunum were isolated. Furthermore, 1 cm diameter slices were punched out of pig and human small intestinal segments and washed with PBS. Human small intestine segments were obtained as waste material after bariatric surgery at the Surgical Clinic of the University Hospital of Würzburg. Written informed consent was obtained from all patients. The small intestinal samples were preincubated with 0.5 mg mL⁻¹ nanogel in PBS containing ATTO488‐RS1‐reg and TAT‐peptide for 30 min. After washing with PBS the samples were fixed for 24 h with 4% (v/v) p‐formaldehyde and embedded in Tissue Tek (Sakura), snap frozen in liquid nitrogen and histological slices produced with a Leica CM3050 S cryotome that were analyzed by CLSM (Leica TCS SP8).

Localization of ATTO488‐RS1‐reg Uptake via Nanohydrogel in Mouse Gastrointestinal Tract

One hour after administration of 30 µL 0.5 mg mL⁻¹ nanogel in PBS containing ATTO488‐RS1‐reg and TAT peptide via gavage, mice were killed by cervical dislocation. After removal and washing of small intestine 1 cm long segments at different positions (3, 5, 10, 15, 20, 25 cm) after stomach outlet were isolated. Furthermore, stomach segments were excised and washed with PBS. Samples were fixed for 24 h with 4% (v/v) p‐formaldehyde, embedded in Tissue Tek (Sakura), snap frozen in liquid nitrogen, and histological slices produced with a Leica CM3050 S cryotome that were analyzed by CLSM (Leica TCS SP8). A corresponding animal welfare application was approved (file number: 55.2.2‐2532‐2‐704, Government of Upper Bavaria).

Immunostaining of Caco‐2 Cells for Confocal Laser Scanning Microscopy (CLSM) and Structured Illumination Microscopy (SIM) Analysis

For immunostaining, Caco‐2 cells were grown on 8 Well Lab‐Tek chamber slides (0.7 cm² cultivated area per well) or coverslips (12 mm diameter) to 90% confluence. The cells were preincubated with 10 µg mL⁻¹ nanogel out of inverse miniemulsion or 100 µg mL⁻¹ nanoprecipitation nanohydrogel containing fluorescently labeled hRS1‐reg(S20E) for 30 min (nanogels with TAT) or 120 min (nanogels without TAT), washed twice with PBS, fixed for 15 min with 4% (w/v) paraformaldehyde diluted in PBS, and washed twice again. Free aldehyde groups were quenched by 10 min incubation with 100 × 10⁻³ m glycine (pH 9). For immunoreactions, washed cells were permeabilized and unspecific binding sites were blocked by a 60 min incubation with PBS containing 2.5% (w/v) BSA and 0.1% (w/v) saponin and incubated for 1 h with primary antibodies diluted in PBS. The dilutions of primary antibodies were as follows: rabbit‐anti‐EEA1‐Ab, 1:100; rabbit‐anti‐TGN46‐Ab, 1:50; mouse‐anti‐Calnexin, 1:50; rabbit anti‐ODC1, 1:50; and rabbit‐anti‐LAMP‐1, 1:200. After incubation with primary antibodies, cells were washed three times with PBS and incubated for 1 h at room temperature with fluorochrome‐linked secondary antibodies donkey‐anti‐rabbit‐ATTO643 or goat‐anti‐mouse‐ATTO643, diluted 1:200 each. The cells were washed five times with PBS, rinsed shortly with double‐distilled water, and for some experiments embedded in Pro Long Glass Anti Fade Mounting Medium from Invitrogen/Thermo Fisher Scientific (Carlsbad, CA, USA). Nuclei were stained with Hoechst 33342 or Hoechst 34580 (Sigma‐Aldrich GmbH, Taufkirchen, Germany).

Quantitative Colocalization Analysis

To estimate colocalization between two channels of each slice from SIM z‐stacks Pearson's coefficient was calculated using a custom‐written ImageJ macro (see Supplementary). The macro is based on the JACoP plugin (https://imagej.nih.gov/ij/plugins/track/jacop2.html) to calculate Pearson's coefficients for each individual slice of all slices in a z‐stack between two channels.^{[ 82 ]} The results for each slice are written into a single results table for each z‐stack, respectively. Individual Pearson's coefficients of each stack, with at least 39 slices/data points, are depicted as superviolin plot using the python package superviolin 1.0.6 (https://pypi.org/project/superviolin/).^{[ 83 ]}

Airyscan Analysis of TAT Localization

To analyze whether TAT was bound merely in the peripherie of nanogel particles or coupled to the matrix throughout the entire particle volume, the TAT peptide was fluorescently labeled with Alexa 647 and incorporated into nanogels out of nanoprecipitation. Especially particles in the sample with a much higher diameter than the mean were analyzed to get a clear result using a ZEISS LSM 900 confocal microscope with Airyscan 2 (Leica Microsystems, Wetzlar, Germany) in Airyscan mode with a laser at a wavelength of 640 nm.

Cryo Scanning Electron Microscopy (Cryo‐SEM)

The hydrogel nanoparticles were imaged using a Crossbeam 340 field emission scanning electron microscope (SEM) (Carl Zeiss Microscopy, Oberkochen, Germany). To visualize the nanoparticles in a swollen state, nanogels in a concentrated form were placed between two alumina holders (d = 3 mm), both containing a notch with a diameter of 2 mm, inclosing the sample and rapidly frozen in nitrogen slush (SN) at ‐210°C. The samples were subsequently transferred with an EM VCT100 cryo‐shuttle into an EM ACE600 sputter coater (both Leica Microsystems, Wetzlar, Germany) at ‐140°C. Here, under high vacuum (<1 × 10⁻³ mbar) the upper half of the sample was knocked off, creating a freshly fractured surface, and freeze‐etched at ‐85 °C for 15 min. After etching, the samples were sputtered with 2.5 nm of platinum and again transferred with the cryo‐shuttle into the SEM. Here the morphology of the fractured surface with the uncovered nanoparticles was imaged at ‐140 °C, by setting an acceleration voltage of 8 kV and detection of secondary electrons with an Everhart Thornley detector. Supplemental measurements were taken by energy dispersive X‐ray spectroscopy (EDX), to verify the positions of the nanogel on the surface. The EDX measurements were taken with an X‐MaxN 50 Silicon Drift Detector (Oxford Instruments, Abington, England), also setting the acceleration voltage to 8 kV.

Statistics

Uptake measurements indicated in Figure 2 are presented as mean ± S.E. Uptake rates of phlorizin‐inhibited AMG uptake into Caco‐2 cells were calculated from three independent experiments. In each experiment, three uptake measurements were performed in the absence of phlorizin and corrected for uptake measured in the presence of phlorizin. Uptake rates measured in segments of intestinal mucosa were calculated from measurements in three animals. For each animal, the uptake rate was calculated from three measurements using two experimental conditions (AMG uptake without and with phlorizin). When three or more groups of data were compared, the significance of differences was determined by analysis of variance (ANOVA) using post hoc Tukey comparison. The significance of the differences between two groups was determined by Student's t test. P < 0.05 was considered significant.