The Role of Crosslinker Content of Positively Charged NIPAM Nanogels on the In Vivo Toxicity in Zebrafish

Polymeric nanogels as drug delivery systems offer great advantages, such as high encapsulation capacity and easily tailored formulations; however, data on biocompatibility are still limited. We synthesized N-isopropylacrylamide nanogels, with crosslinker content between 5 and 20 mol%, functionalized with different positively charged co-monomers, and investigated the in vivo toxicity in zebrafish. Our results show that the chemical structure of the basic unit impacts the toxicity profile depending on the degree of ionization and hydrogen bonding capability. When the degree of crosslinking of the polymer was altered, from 5 mol% to 20 mol%, the distribution of the positively charged monomer 2-tert-butylaminoethyl methacrylate was significantly altered, leading to higher surface charges for the more rigid nanogels (20 mol% crosslinker), which resulted in >80% survival rate (48 h, up to 0.5 mg/mL), while the more flexible polymers (5 mol% crosslinker) led to 0% survival rate (48 h, up to 0.5 mg/mL). These data show the importance of tailoring both chemical composition and rigidity of the formulation to minimize toxicity and demonstrate that using surface charge data to guide the design of nanogels for drug delivery may be insufficient.


Introduction
The development of nanomaterials as drug delivery systems is increasingly attractive thanks to the advantages that these materials bring, such as high loading capacity [1][2][3][4], sitespecific targeting [5,6], and controlled drug release via external stimuli, such as temperature and pH [7,8]. Covalently crosslinked acrylamide-based nanogels (NGs), with their highly swellable matrices, have high colloidal stability and higher biocompatibility compared to other nanomaterials [9][10][11][12][13][14][15][16]. In addition, the synthetic methods are simple, an important feature that allows the formulations to be easily tailored, facilitating not only the uploading of different drugs, but also the evaluation of how changes in formulation result in different properties [17][18][19][20].
The relationship between the formulation and properties of NGs has attracted considerable research, particularly given the importance of chemical structure and morphology in influencing the interaction with biological entities, in vivo, resulting in a protein/biomolecular corona [21]. We investigated the impact of changes in formulations and structure on the interfacial behavior of N-isopropylacrylamide (NIPAM) NGs, using a combination of neutron reflectivity and atomic force microscopy [22], and correlated formulations with changes in thermoresponsive properties [23], together with the role of monomer reactivities [24]. More recently, we reported our results on the role that charges play on biocorona formation with acrylamide-based NGs [25].
Despite advances in understanding the structure and properties of NGs, their use in clinical applications remains limited by the lack of comprehensive knowledge of their biocompatibility and long-term toxicity [26]. Important studies have highlighted how the for both NGs purification and characterization, was obtained using the Elga Option PURELAB Water Purification System. Fish water is a saltwater solution comprised of sodium bicarbonate, marine salts and calcium sulfate in a ratio of 7.5:1.8:0.84 g in 100 L of reverse osmosis (R/O) purified water, and this was used as taken from the zebrafish maintenance water system. Polytetrafluoroethylene (PTFE) syringe filters with pore size 0.2 µm were obtained from Fisher Scientific (Loughborough, UK).

Synthesis of NGs
NGs were prepared by high dilution radical polymerization (HDRP) according to a previously reported procedure [24]. Monomers were solubilized, together with the crosslinker MBA, in anhydrous DMSO in a 10 mL Wheaton™ bottle. The volume of dry DMSO was adjusted to obtain the desired total monomer concentration C M 1-3% (w/w) depending on the formulation. The initiator AIBN was then added to the polymerization mixture in concentrations of 1 or 2 mol% of total moles of double bonds present in the mixture.
The Wheaton™ bottle was sealed and the pre-polymerization mixture purged with N 2 . The solution was then placed in an oven at a temperature equal to 70 • C. After 24 h, the transparent NG solution was dialyzed against deionized water for 3 days with frequent changes (3 times per day) before being transferred to a 50 mL falcon tube and frozen in liquid N 2 . The frozen NG solution was lyophilized using an LTE Scientific Lyotrap freeze-drier (Greenfield, UK) to yield a fluffy white powder. All the NGs synthesized were stored in glass vials at room temperature.

Study of Monomer Conversions by 1 H NMR
The quantification of the monomer conversions of the NGs was carried out by 1 H NMR (Bruker AVIII 400 MHz BBO Probe), simultaneously to the NG synthesis, following a previously reported procedure [24]. Aliquots of 250 µL were drawn from the Wheaton™ bottle before (pre-polymerization, t = 0 h) and after (post-polymerization, t = 24 h) the 24 h polymerization and mixed with 250 µL of DMSO-d 6 in an NMR tube. The internal standard TMB, solubilized in DMSO-d 6 (concentration between 0.15 and 0.19 M), was added to both NMR tubes to obtain a molar concentration ratio equal to 1:4 with NIPAM. The 1 H NMR spectra were acquired using a method suppressing the non-deuterated solvent signal (δ DMSO = 2.50 ppm) and were then phased and integrated manually with the Bruker TopSpin 4.0.6 software. To calculate the concentration of unreacted monomers and MBA in the initial and final polymerization solutions, the intensities of monomer peaks (5.55 ppm for NIPAM, 5.67 ppm for t-BAEMA, 5.66 ppm for GUA, and 5.63 ppm for MBA) were compared to the intensity of TMB peak at 6.88 ppm.

Nanoparticle Size by Dynamic Light Scattering
The hydrodynamic diameter (Dh) of NGs was measured by DLS at 25 • C using a Zetasizer Nano Ultra operated with the ZS Xplorer software (version 1.5.0.163) (Malvern Instruments Ltd., Malvern, UK). NGs colloidal solutions (concentration 1 mg/mL) were prepared by dissolving the dry NG powder in either deionized water or fish water, followed by sonication for 10 min at room temperature, and filtration through 0.2 µm PTFE syringe filters. To avoid any contamination due to the filtration, the filters were preconditioned with 1 mL of the solvent and the first 3 drops of filtered NG solution discarded. All the DLS measurements were carried out in triplicate using backscatter angle mode (174.7 • ), allowing at least 4 min for the sample temperature to equilibrate prior to each measurement.
To study the colloidal stability of the nanogels, after obtaining nanogel solutions in fish water (pH = 7) at 0.5 mg/mL in the same way as described for hydrodynamic diameter characterization, samples were transferred into a DLS cuvette and incubated at 28 • C in a water bath for up to 96 h. Then, triplicate DLS measurements (number distribution) were collected at 0, 24, and 96 h without removing the sample from its cuvette; the colloidal stability of the nanogels was evaluated by comparing the size of the nanogels at different times compared to their initial size (0 h).

Nanoparticle Surface Charge by Electrophoretic Light Scattering
The electrophoretic mobility (µ) of positively charged NGs was investigated by ELS at 28.5 • C, using a Zetasizer Nano Ultra operated with the ZS Xplorer software (version 1.5.0.163) (Malvern Instruments Ltd., Malvern, UK). All the NGs were analyzed after dispersion in fish water (pH = 7) at a concentration equal to 1 mg/mL, sonication for 10 min, and filtration through 0.2 µm PTFE syringe filters. The samples were then transferred into disposable folded capillary cells (1080, Malvern Instruments Ltd., Malvern, UK) and the measurements performed after at least 4 min equilibration.

Thermoresponsive Properties of NGs by UV-Visible Spectroscopy
The volume phase transition temperatures (VPTTs) of the NGs were determined by turbidimetry assay, i.e., by measuring the light transmittance (at 500 nm) of the NG solutions as a function of temperature. The NG solutions, with concentration equal to 1 mg/mL, were prepared by dissolution of the dry NG in either deionized water or fish water followed by 10 min sonication and filtration with 0.2 µm PTFE syringe filters. The measurements were carried out in a temperature range between 25 and 65 • C, at a rate of 0.5 • C/minute, using a Cary 100 UV-Visible Spectrophotometer (Agilent Technologies, Santa Clara, CA, USA) coupled with a temperature controller. The raw data were analyzed using Origin2018 software, which allowed the plotting and fitting of the transmittancetemperature data with the sigmoidal Boltzmann fitting. For each measurement, the VPTT was considered equal to the temperature at which the curve presented its inflection point, after verifying the accuracy of this fitting in terms of adjusted R 2 (≥0.99).

Zebrafish Husbandry and Maintenance
In the animal facility, wild type Tübingen (TU) strain zebrafish were maintained at 28.5 • C in a recirculating system (Tecniplast, UK) on a 14 h:10 h light/dark cycle. Fish were fed twice daily, with ZM-400 fry food (Zebrafish Management Ltd., Winchester, UK) in the morning and brine shrimp in the afternoon, following standard husbandry protocols [52]. For breeding, fish were placed in groups of 3 (2 females and 1 male) in sloping breeding tanks (Tecniplast, UK) overnight with a clear divider separating males from females. The divider was removed on the following morning to allow fish to spawn, and the eggs were collected between 2 and 2.5 h post fertilization (hpf), when they were sorted fertile from infertile and placed individually in the wells of 48-well plates (1 mL of fish water or NG solution in each well). The embryos/larvae were then incubated at 28.5 • C for the entire duration of the experiment. All procedures were performed in accordance with UK Animals (Scientific Procedures) Act. No regulated procedures were undertaken (as no experiments were continued past 4 dpf).

Evaluation of the Biocompatibility of NGs in Zebrafish
The effects of NGs on zebrafish were assessed by monitoring the morphology and development of the embryos/larvae immersed in NG solutions. NG solutions in fish water, having concentration between 0.01 and 0.5 mg/mL, were obtained by dilution from 1 mg/mL stock solutions, prepared following the same protocol described for the characterization studies. After collection and sorting, the embryos at 64-256 cell stage (2 to 2.5 hpf) were placed individually in the wells of a 48-well plate, immersed in 1 mL of NG solution or, in case of control, of fresh fish water, and housed in an incubator at 28.5 • C. Their morphology and development were monitored under a Leica M165 FC stereo microscope (Leica Microsystems, Germany) at 4, 24, 48, 72, and 96 h after the immersion and compared with the development of unexposed fish used as control. Each condition was tested on a group of 8 fish and repeated 3 times, when no damage due to the nanoparticle exposure was observed.

Synthesis and Characterization of Neutral and Positively Charged NGs
A small library of covalently crosslinked NGs was synthesized via high dilution radical polymerization (HDRP), which allows nano-sized gel particles to be obtained with tunable size without addition of surfactants [53]. To introduce a thermoresponsive behavior in the NGs, N-isopropylacrylamide (NIPAM), see Scheme 1, was employed as backbone monomer. NIPAM has been commonly employed to obtain thermoresponsive functional nanomaterials due to its lower critical solubility temperature (LCST) of 32 • C, which is close to body temperature. N,N -methylenebisacrylamide (MBA) was used as the crosslinker in concentrations of 5-20 mol% to vary the flexibility of the particle and study its role on the physico-chemical and toxicological properties of the nanoparticles. All the formulations were prepared using 2,2 -azobisisobutyronitrile (AIBN) as the initiator, initially in concentration of 1 mol% (of the total moles of double bonds in the mixture) in DMSO. The particles were then purified via extensive dialysis, which ensures that the organic solvent is removed.
(Leica Microsystems, Germany) at 4, 24, 48, 72, and 96 h after the immersion and compared with the development of unexposed fish used as control. Each condition was tested on a group of 8 fish and repeated 3 times, when no damage due to the nanoparticle exposure was observed.

Synthesis and Characterization of Neutral and Positively Charged NGs
A small library of covalently crosslinked NGs was synthesized via high dilution radical polymerization (HDRP), which allows nano-sized gel particles to be obtained with tunable size without addition of surfactants [53]. To introduce a thermoresponsive behavior in the NGs, N-isopropylacrylamide (NIPAM), see Scheme 1, was employed as backbone monomer. NIPAM has been commonly employed to obtain thermoresponsive functional nanomaterials due to its lower critical solubility temperature (LCST) of 32 °C, which is close to body temperature. N,N'-methylenebisacrylamide (MBA) was used as the crosslinker in concentrations of 5-20 mol% to vary the flexibility of the particle and study its role on the physico-chemical and toxicological properties of the nanoparticles. All the formulations were prepared using 2,2′-azobisisobutyronitrile (AIBN) as the initiator, initially in concentration of 1 mol% (of the total moles of double bonds in the mixture) in DMSO. The particles were then purified via extensive dialysis, which ensures that the organic solvent is removed. Positively charged NGs were obtained by introducing 10 mol% of 2-(tert-butylamino)-ethyl-methacrylate (t-BAEMA), which presents a weak basic character (pKa of 8 once polymerized [54]) that may reduce its impact on the thermoresponsive behavior of the NGs and limit toxicity. Good monomer conversions (71-88%) and chemical yields >70% (Table 1) were obtained for neutral NGs when a total monomer concentration (CM) of 1% (w/w) was employed, as found in previous works [24,55]. Positively charged NGs were obtained by introducing 10 mol% of 2-(tert-butylamino)ethyl-methacrylate (t-BAEMA), which presents a weak basic character (pKa of 8 once polymerized [54]) that may reduce its impact on the thermoresponsive behavior of the NGs and limit toxicity. Good monomer conversions (71-88%) and chemical yields >70% (Table 1) were obtained for neutral NGs when a total monomer concentration (C M ) of 1% (w/w) was employed, as found in previous works [24,55]. A lower yield compared to the overall monomer conversion is usually observed because of the loss of low-molecular-weight polymer particles during the purification step by dialysis (molecular cut-off 3.5 kDa). The addition of t-BAEMA in the formulation was associated with lower monomer conversions and yields at C M 1 and 2% (Table S1), possibly due to a quenching effect of the methacrylate monomer. Increasing the C M to the optimized value of 3% was found to bring both the overall monomer conversions and yields to satisfactory values, encouraging further characterization of the nanoparticles.
The hydrodynamic diameter (D h ) of NGs is commonly determined by dynamic light scattering (DLS), which provides sizes based on the intensity of the scattered light; however, when analyzing particles with very small sizes, the presence of even very small amounts of larger particles, including dust, results in an over-estimation of their relative abundance in the sample. For this reason, in this case, size distribution by number provides a more realistic estimate of the particles' diameters. The neutral NGs were found to be <10 nm in deionized water ( Figure S1), consistent with previously reported NGs [25,55,56]; for MBA concentrations of 5, 10, and 20 mol%, a D h of 3.6, 4.7, and 8.0 nm, respectively, was observed, showing the impact of the crosslinker on the size of the nanoparticles. A similar trend was seen for the NGs incorporating 10 mol% t-BAEMA, although the size of the particles was found to be overall slightly smaller (Table 1, Figure S2); this might be due to the hydrophobicity of the t-butyl substituent of the t-BAEMA unit, which is only partially ionized in water due to its weakly basic character. Given that the size of nanoparticles is affected by the nature of the media [56], the characterization was carried out in fish water, which is the medium used for the in vivo studies with zebrafish [52,57]. Data show that the size of the neutral NGs was mostly unaffected by the change in medium ( Figure S3); however, positively charged NGs were found to have slightly larger D h (Figure S4), possibly due to the lower pH of the fish water (around 7) compared to the unbuffered deionized water, leading to a larger degree of ionization of the t-BAEMA unit and, consequently, a more swollen NG matrix.
The electrophoretic mobility (µ) of the NGs was then measured to confirm the presence of the charge on the surface of the particles. Data in Figure 1a show that the NGs displayed a positive charge in fish water, which confirmed the presence of the t-BAEMA residues on the surface of the particle. Importantly, NGs with 5 mol% MBA displayed significantly lower µ values compared to those with 10 and 20 mol%, although no statistical difference between the 10 and 20% crosslinked NGs could be observed. These data suggest that the presence of a more loosely crosslinked matrix leads to a more uniform distribution of the t-BAEMA unit in the NG compared to the more surface-localized in the more crosslinked particles, thus leading to a lower surface charge.
by dialysis (molecular cut-off 3.5 kDa). The addition of t-BAEMA in the formulation was associated with lower monomer conversions and yields at CM 1 and 2% (Table S1), possibly due to a quenching effect of the methacrylate monomer. Increasing the CM to the optimized value of 3% was found to bring both the overall monomer conversions and yields to satisfactory values, encouraging further characterization of the nanoparticles.
The hydrodynamic diameter (Dh) of NGs is commonly determined by dynamic light scattering (DLS), which provides sizes based on the intensity of the scattered light; however, when analyzing particles with very small sizes, the presence of even very small amounts of larger particles, including dust, results in an over-estimation of their relative abundance in the sample. For this reason, in this case, size distribution by number provides a more realistic estimate of the particles' diameters. The neutral NGs were found to be <10 nm in deionized water ( Figure S1), consistent with previously reported NGs [25,55,56]; for MBA concentrations of 5, 10, and 20 mol%, a Dh of 3.6, 4.7, and 8.0 nm, respectively, was observed, showing the impact of the crosslinker on the size of the nanoparticles. A similar trend was seen for the NGs incorporating 10 mol% t-BAEMA, although the size of the particles was found to be overall slightly smaller (Table 1, Figure S2); this might be due to the hydrophobicity of the t-butyl substituent of the t-BAEMA unit, which is only partially ionized in water due to its weakly basic character. Given that the size of nanoparticles is affected by the nature of the media [56], the characterization was carried out in fish water, which is the medium used for the in vivo studies with zebrafish [52,57]. Data show that the size of the neutral NGs was mostly unaffected by the change in medium ( Figure S3); however, positively charged NGs were found to have slightly larger Dh ( Figure S4), possibly due to the lower pH of the fish water (around 7) compared to the unbuffered deionized water, leading to a larger degree of ionization of the t-BAEMA unit and, consequently, a more swollen NG matrix.
The electrophoretic mobility (µ) of the NGs was then measured to confirm the presence of the charge on the surface of the particles. Data in Figure 1a show that the NGs displayed a positive charge in fish water, which confirmed the presence of the t-BAEMA residues on the surface of the particle. Importantly, NGs with 5 mol% MBA displayed significantly lower µ values compared to those with 10 and 20 mol%, although no statistical difference between the 10 and 20% crosslinked NGs could be observed. These data suggest that the presence of a more loosely crosslinked matrix leads to a more uniform distribution of the t-BAEMA unit in the NG compared to the more surface-localized in the more crosslinked particles, thus leading to a lower surface charge. To further investigate this finding, we characterized the thermoresponsive properties of the NGs by measuring their volume phase transition temperature (VPTT). The transmittance of NGs solutions (1 mg/mL) was monitored as a function of temperature in the To further investigate this finding, we characterized the thermoresponsive properties of the NGs by measuring their volume phase transition temperature (VPTT). The transmittance of NGs solutions (1 mg/mL) was monitored as a function of temperature in the range 25-65 • C using UV-Vis spectroscopy. The VPTT coincides with the temperature at which the transmittance of the sample drops to 50%. Neutral NGs with a different degree of crosslinking displayed similar VPTT values in fish water between 39 and 41 • C, while the VPTT of the t-BAEMA-NGs crosslinked with 5, 10, or 20 molar% of MBA were found to be 37.4, 40.3, and 44.7 • C, respectively, in fish water; these values are higher than the neutral NGs due to the presence of the positive charge, which increases the solvation of the NG's polymer chains. Interestingly, a clear effect of the crosslinker content was observed in the case of the positively charged NGs, with higher VPTT values being observed for more crosslinked matrices. These data provide clear evidence that the more crosslinked NGs present better colloidal stability compared to the more loosely crosslinked NGs, possibly Having obtained the characterization data, the next step focused on evaluating whether the change in composition of the NGs and the difference in surface charge could influence the biocompatibility profiles of the particles. This was evaluated using zebrafish as an in vivo model.

Impact of Surface Chemistry and Morphology of NGs on their In Vivo Toxicity on Zebrafish Model
The in vivo toxicity of NGs was tested by direct environmental exposure on zebrafish up to 4 dpf, which involved adding the NGs to the medium where the zebrafish embryos were immersed (Figure 2a). The colloidal stability of the NGs in fish water was confirmed at the highest concentration studied (0.5 mg/mL) up to 96 hrs by DLS ( Figure S5a-f). To evaluate the toxicity, the embryos/larvae were monitored under a stereomicroscope to check for any morphological deformities and/or any delay in their development. Moreover, after the hatching, the capability of swimming and response to touch stimuli were used as further parameters for healthy status. When the neutral NGs were tested at concentrations of 0.5 mg/mL, a 100% survival rate was observed for exposure times up to 72 h (Figure 2b) with no effects on the development of the embryos (Figure 2c for NG 20 -1, and Figure S6 for NG 10 -1 and NG 5 -1), suggesting that these formulations are well tolerated by the zebrafish model (regardless of their crosslinker content). When the zebrafish embryos were exposed to positively charged NG20-tBAEMA-3 (20 mol% crosslinker) in concentrations ranging between 0.1 and 0.5 mg/mL, survival rates around 80% were recorded after only 48 h of incubation (Figure 2d), with higher concentration of 0.5 mg/mL leading to a drop to 50% survival rate at 72 h. This indicates that, as expected, the incorporation of a positively charged co-monomer in the NGs induces toxicity in the animal, likely a consequence of the interaction between the particle and the negatively charged cellular structures, resulting in a loss of cellular function and structure. When the zebrafish embryos were exposed to positively charged NG 20 -tBAEMA-3 (20 mol% crosslinker) in concentrations ranging between 0.1 and 0.5 mg/mL, survival rates around 80% were recorded after only 48 h of incubation (Figure 2d), with higher concentration of 0.5 mg/mL leading to a drop to 50% survival rate at 72 h. This indicates that, as expected, the incorporation of a positively charged co-monomer in the NGs induces toxicity in the animal, likely a consequence of the interaction between the particle and the negatively charged cellular structures, resulting in a loss of cellular function and structure.
It is important to highlight that as the embryos are transparent, evidence of gross morphological changes can be visible generally without the need for staining. An example of images showing embryos exposed to toxic nanogels is shown in Figure S7.
Interestingly, positively charged NGs that presented a lower degree of crosslinker displayed a higher toxicity on the embryos (Figure 2e,f), with NG 5 -tBAEMA-3 leading to a 20% survival rate in 24 h even at the lowest concentration of 0.1 mg/mL. This result is somewhat counterintuitive, as characterization of the NGs indicated that NG 5 -tBAEMA-3 presented the lowest surface charge. However, NGs with less crosslinked matrices have been shown to possess higher flexibility and adaptability to interfaces [22,55]. The crosslinker, once incorporated in the matrix, does not impact toxicity on its own, as previously reported [16]. This, together with the data presented here, suggests that NGs with higher flexibility may be able to interact better with cellular structures, thus increasing the formulation's toxicity. This finding also suggests that the characterization of nanoparticles may lead to inappropriate assumptions of their behavior in vivo, and that a complete characterization of NGs should be carried out before hypothesizing about their potential interactions with biological entities.
To further understand the role of the NG's composition on their toxicity, we then evaluated how the addition of positively charged co-monomers having different structures affected the survival rate of the zebrafish embryo.

Impact of Positively Charged Group Structure on the Toxicity of NGs
Two additional positively charged monomers, 3-acrylamidopropyl-trimethylammonium chloride (ATC) and N-(3-methacrylamidopropyl)-guanidinium chloride (GUA), were separately added at concentrations of 10 mol% to the NGs formulations (Figure 3a). Given the lower toxicity observed for the more rigid t-BAEMA-based NGs, optimized formulations NG 20 -ATC-2 and NG 20 -GUA-2 were obtained using a fixed amount of 20 mol% of MBA (Table S2). Monomer conversions were found to be consistent with formulations containing t-BAEMA, although slightly lower chemical yields were observed for both NG 20 -ATC-2 (60%) and NG 20 -GUA-2 (54%).
The nanoparticles showed D h around 7.5 nm (number distribution) in fish water (colloidally stable up to 96 h, Figure S5), which is comparable to NG 20 -tBAEMA-3. However, when electrophoretic mobility of the nanoparticles was analyzed (in fish water) values of 2 µm cm/V s were observed for both the ATC and GUA NGs, almost twice as high compared to the related t-BAEMA containing NG.
This result is likely due to the fully ionized nature of the ATC and GUA monomers in fish water, which leads to a higher amount of charged groups present on the surface of the NGs (despite all incorporating 10 mol% of charged co-monomer). Solutions of both NG 20 -ATC-2 and NG 20 -GUA-2 in fish water did not show any change in transmittance as a function of temperature up to 65 • C ( Figure S8), a finding that further supports the observation that these NGs present a more charged and hydrated shell. These data indicate that the NGs incorporating three different types of positively charged groups are comparable in terms of size and display different surface properties that can be linked to their formulation. On this basis, we decided to evaluate these polymers in vivo.
We initially conducted toxicological studies on the 20 mol% crosslinked NG 20 -tBAEMA-3 at concentrations < 0.1 mg/mL; this was done to identify a suitable concentration at which the acute toxicity of the positively charged formulation could be considered negligible within the experimental time window. As observed in Figure 3b, NG 20 -tBAEMA-3 did not impact the survival rate of the zebrafish embryo (over 96 h) when used at a concentration of 0.01 mg/mL ( Figure S9), indicating that the formulation could be used to compare the toxicity of the other two positively charged formulations. nium chloride (ATC) and N-(3-methacrylamidopropyl)-guanidinium chloride (GUA), were separately added at concentrations of 10 mol% to the NGs formulations (Figure 3a). Given the lower toxicity observed for the more rigid t-BAEMA-based NGs, optimized formulations NG20-ATC-2 and NG20-GUA-2 were obtained using a fixed amount of 20 mol% of MBA (Table S2). Monomer conversions were found to be consistent with formulations containing t-BAEMA, although slightly lower chemical yields were observed for both NG20-ATC-2 (60%) and NG20-GUA-2 (54%). The nanoparticles showed Dh around 7.5 nm (number distribution) in fish water (colloidally stable up to 96 h, Figure S5), which is comparable to NG20-tBAEMA-3. However, when electrophoretic mobility of the nanoparticles was analyzed (in fish water) values of 2 µm cm/V s were observed for both the ATC and GUA NGs, almost twice as high compared to the related t-BAEMA containing NG.
This result is likely due to the fully ionized nature of the ATC and GUA monomers in fish water, which leads to a higher amount of charged groups present on the surface of the NGs (despite all incorporating 10 mol% of charged co-monomer). Solutions of both NG20-ATC-2 and NG20-GUA-2 in fish water did not show any change in transmittance as a function of temperature up to 65 °C ( Figure S8), a finding that further supports the observation that these NGs present a more charged and hydrated shell. These data indicate that the NGs incorporating three different types of positively charged groups are comparable in terms of size and display different surface properties that can be linked to their formulation. On this basis, we decided to evaluate these polymers in vivo. The data in Figure 3 show that both NG 20 -ATC-2 and NG 20 -GUA-2 showed significant toxicity at 0.01 mg/mL; NG 20 -ATC-2 initially showed no toxicity for 48 h, but led to 20% survival rate after 72 h, demonstrating its more toxic profile compared to the t-BAEMA NG. NG 20 -GUA-2, on the other hand, immediately led to a loss of 40% of the initial embryos within 24 h, to then result in 0% survival rate after 72 h of exposure.
These results show that the NG toxicity depends on the type of positively charged monomer incorporated into the NG structure, with increasing toxicity being observed in the order t-BAEMA < ATC < GUA. NG characterization suggests that NG 20 -tBAEMA-3 has a lower surface charge compared to the other two NGs, likely due to its weakly basic character and consequent partial ionization, while, both NG 20 -ATC-2 and NG 20 -GUA-2 are fully ionized in the fish water medium, with similar surface charge. This indicates that their different toxicological profile might be determined by the chemical structure of the positively charged moieties. Indeed, GUA has been reported to strongly bind to the carboxylate groups of proteins [58] thanks to its ability to donate 2 hydrogen bonds; this is not possible in the case of the quaternary ammonium group in ATC, which may lead these NGs to form weaker bonds with proteins present on the surface of the cells. These data further expand the understanding of how the surface chemistry of nanoparticles requires careful design.

Conclusions
We synthesized and characterized covalently crosslinked NIPAM-based NGs, neutral or incorporating three different positively charged co-monomers, t-BAEMA, ATC, and GUA, with varying content of MBA (5-20 mol%). The NGs were obtained with a similar hydrodynamic diameter and polydispersity, while the surface charge and thermoresponsive behavior, resulting from the use of NIPAM as backbone monomer, was found to be consistent with their formulations. The data demonstrated that the crosslinker content directly influenced the location of the positively charged monomer t-BAEMA in the matrix, with higher values of surface charge observed when the crosslinker was increased from 5 mol% to 10-20 mol%.
The toxicity of the nanoparticles was evaluated by environmental exposure of the zebrafish larvae (up to 4 days post fertilization) to the different nanoparticles, and by monitoring the survival rate up to 96 h. Neutral NGs did not impact the survival of the zebrafish larvae, regardless of their rigidity, while the presence of the positively charged t-BAEMA in the matrix highlighted the key role played by the crosslinker content in influencing the toxicity. The nanogels with 20 mol% crosslinker, previously shown to have higher surface charge, resulted in >80% survival rate over a period of 48 h with concentrations up to 0.5 mg/mL, while the more flexible polymers with 5 mol% crosslinker were found to be highly toxic at the same concentrations (0% survival rate at 48 h).
The choice of the positively charged moiety also played an important role, where formulations having similar surface charge values but incorporating a quaternary ammonium group (ATC) led to lower toxicity compared to NGs incorporating the guanidinium group (GUA) as a consequence of different hydrogen bonding capabilities.
These results provide important insights into the roles that crosslinker content and functional groups' structure play in influencing the toxicity of nanoparticles, as a result of influencing the position of charged units in the matrix, as well as determining its ability to interact with biomolecules, ultimately influencing their behavior in vivo. The development of new drug delivery systems based on nanoparticles requires a comprehensive characterization of nanoparticles that may lead to incorrect hypotheses regarding their behavior in vivo. The data revealed the importance of the degree of crosslinking, both in terms of its impact on the positioning of positively charged co-monomers in the matrix, and on the toxicity of the particle. The present study thus provides a tool to develop positively charged NGs with higher biocompatibility.