Solid-State NMR Characterization of Mefloquine Resinate Complexes Designed for Taste-Masking Pediatric Formulations

Mefloquine (MQ) is an antimalarial medication prescribed to treat or malaria prevention.. When taken by children, vomiting usually occurs, and new doses of medication frequently need to be taken. So, developing pediatric medicines using taste-masked antimalarial drug complexes is mandatory for the success of mefloquine administration. The hypothesis that binding mefloquine to an ion-exchange resin (R) could circumvent the drug’s bitter taste problem was proposed, and solid-state 13C cross-polarization magic angle spinning (CPMAS) NMR was able to follow MQ–R mixtures through chemical shift and relaxation measurements. The nature of MQ–R complex formation could then be determined. Impedimetric electronic tongue equipment also verified the resinate taste-masking efficiency in vitro. Variations in chemical shifts and structure dynamics measured by proton relaxation properties (e.g., T1ρH) were used as probes to follow the extension of mixing and specific interactions that would be present in MQ–R. A significant decrease in T1ρH values was observed for MQ carbons in MQ–R complexes, compared to the ones in MQ (from 100–200 ms in MQ to 20–50 ms in an MQ–R complex). The results evidenced that the cationic resin interacts strongly with mefloquine molecules in the formulation of a 1:1 ratio complex. Thus, 13C CPMAS NMR allowed the confirmation of the presence of a binding between mefloquine and polacrilin in the MQ–R formulation studied.


Introduction
Mefloquine hydrochloride (MQ), when administered orally, either alone or in combination with artemisinin derivatives, is one of the primary agents used for the prevention, treatment, and emergency treatment of malaria infections caused by Plasmodium vivax and Plasmodium falciparum [1,2].Other diseases can be treated with MQ.In a recent review, Dobhal et al. [3] examined the last five years of work using natural biomolecules (BMs) to counteract the ZIKA virus through virtual screening and in vitro investigations.In vitro testing has shown that mefloquine can reduce ZIKA virus infections in cell lines.
Due to the quinine moiety in its chemical structure (Figure 1), MQ has a highly unpleasant, bitter taste.The bitter taste of antimalarial drugs is pointed to as the most likely cause of non-compliance to treatment, mainly by children [4,5].So, developing pediatric formulations that could mask the bitter taste of MQ would be a considerable advantage for the success of oral administration.However, adding flavors and sweeteners to the formulation may not be efficient enough to achieve acceptable drug palatability, requiring some technological process to improve the taste-masking efficiency [4,6].Among the different techniques described in the literature to overcome taste-masking issues, such as microencapsulation [7] or formation of inclusion complexes [8], complexation with ion-exchange resins (IER) has received considerable attention.It has often proved to be an efficient method for masking the bitter taste of different drugs [9][10][11][12][13][14][15].Guimaraes et al. [16] have investigated the taste masking of chloroquine-resin complexes using polacrilin potassium resin (Figure 2) and other ion-exchange resins.The results indicated increased taste properties of the chloroquine-polacrilin complex compared to chloroquine diphosphate.
Mefloquine hydrochloride (MQ) is a weak base with a pKa of approximately 8.6, and solubility has been reported to range from very slightly soluble to slightly soluble in water at room temperature.According to IUPAC nomenclature, it is termed (R, S)-erythroa-(2-piperidyl)-2,8-bis(trifluoromethyl)-4-chinolinmethanol-hydrochloride.The chemical structure is shown in Figure 1, with a molecular weight of 414.77 g/mol and presents a melting point of around 259-260 • C [17].Polacrilin potassium resin is a weakly acidic cation-exchange resin widely used as a tablet disintegrant in oral dosage formulations of drug products [18].Due to the quinine moiety in its chemical structure (Figure 1), MQ has a highly unpleasant, bi er taste.The bi er taste of antimalarial drugs is pointed to as the most likely cause of non-compliance to treatment, mainly by children [4,5].So, developing pediatric formulations that could mask the bi er taste of MQ would be a considerable advantage for the success of oral administration.However, adding flavors and sweeteners to the formulation may not be efficient enough to achieve acceptable drug palatability, requiring some technological process to improve the taste-masking efficiency [4,6].Among the different techniques described in the literature to overcome taste-masking issues, such as microencapsulation [7] or formation of inclusion complexes [8], complexation with ion-exchange resins (IER) has received considerable a ention.It has often proved to be an efficient method for masking the bi er taste of different drugs [9][10][11][12][13][14][15].Guimaraes et al. [16] have investigated the taste masking of chloroquine-resin complexes using polacrilin potassium resin (Figure 2) and other ion-exchange resins.The results indicated increased taste properties of the chloroquine-polacrilin complex compared to chloroquine diphosphate.
Mefloquine hydrochloride (MQ) is a weak base with a pKa of approximately 8.6, and solubility has been reported to range from very slightly soluble to slightly soluble in water at room temperature.According to IUPAC nomenclature, it is termed (R, S)-erythro-a-(2piperidyl)-2,8-bis(trifluoromethyl)-4-chinolinmethanol-hydrochloride.The chemical structure is shown in Figure 1, with a molecular weight of 414.77 g/mol and presents a melting point of around 259-260 °C [17].Polacrilin potassium resin is a weakly acidic cation-exchange resin widely used as a tablet disintegrant in oral dosage formulations of drug products [18].Due to the quinine moiety in its chemical structure (Figure 1), MQ has a highly unpleasant, bi er taste.The bi er taste of antimalarial drugs is pointed to as the most likely cause of non-compliance to treatment, mainly by children [4,5].So, developing pediatric formulations that could mask the bi er taste of MQ would be a considerable advantage for the success of oral administration.However, adding flavors and sweeteners to the formulation may not be efficient enough to achieve acceptable drug palatability, requiring some technological process to improve the taste-masking efficiency [4,6].Among the different techniques described in the literature to overcome taste-masking issues, such as microencapsulation [7] or formation of inclusion complexes [8], complexation with ion-exchange resins (IER) has received considerable a ention.It has often proved to be an efficient method for masking the bi er taste of different drugs [9][10][11][12][13][14][15].Guimaraes et al. [16] have investigated the taste masking of chloroquine-resin complexes using polacrilin potassium resin (Figure 2) and other ion-exchange resins.The results indicated increased taste properties of the chloroquine-polacrilin complex compared to chloroquine diphosphate.
Mefloquine hydrochloride (MQ) is a weak base with a pKa of approximately 8.6, and solubility has been reported to range from very slightly soluble to slightly soluble in water at room temperature.According to IUPAC nomenclature, it is termed (R, S)-erythro-a-(2piperidyl)-2,8-bis(trifluoromethyl)-4-chinolinmethanol-hydrochloride.The chemical structure is shown in Figure 1, with a molecular weight of 414.77 g/mol and presents a melting point of around 259-260 °C [17].Polacrilin potassium resin is a weakly acidic cation-exchange resin widely used as a tablet disintegrant in oral dosage formulations of drug products [18].Solid-state NMR spectroscopy is a powerful tool for organic and inorganic materials and crystalline or amorphous phases.Particularly for rare spins (as 13 C, 1.1% natural abundance), direct observation is sometimes prohibitive, and indirect excitation by abundant spins (as 1 H) allows for obtaining 13 C NMR spectra in a reasonable time through a cross-polarization magic-angle spinning pulse sequence (CPMAS).It is known [21] that in a given phase, by using the CPMAS sequence, the dependence of the 13 C magnetization, S(t c ) on contact time t c , is given by Equation (1): T CH and T 1ρ H measurements could provide information on the possible interactions between distinct components in a mixture.The procedure has considered that separate domains may have different proton relaxation properties.As a result, differences in dynamics could be used to establish the extension of mixing and specific interactions that would be present between distinct domains.That approach has been applied successfully to synthetic [22] and natural polymer mixtures [23].Molecular motion on the scale of several nanometers can be investigated by measuring the spin-lattice relaxation time of the 1 H nucleus (T 1ρ H ) in a rotating system [24].Recently, Durand et al. [25] addressed the interactions between mefloquine and cyclodextrins by 1 H and 19 F NMR in solution.They showed that the non-covalent interactions present enhanced the solubility of the inclusion complexes thus formed.
An electronic tongue is an analytical tool based on a multisensory system based on potentiometric, amperometric, voltammetric, or impedimetric transduction methods.Among them, the impedimetric method is attractive for non-faradaic processes and dismisses the use of reference and auxiliary electrodes.A "fingerprint" of the analyzed liquid is obtained from the electrical responses of an array of sensing units that provide a qualitative assessment of taste based on the global selectivity paradigm.Electronic tongues have been successfully applied for in vitro taste-masking assessment and will be applied to complement the NMR results [26,27].
In this work, 13 C solid-state NMR with a cross-polarization magic angle spinning pulse sequence was used to probe the interactions in mefloquine-polacrilin through chemical shift variations and 1 H relaxation times. 13C NMR 13 C CPMAS NMR spectra of MQ were acquired on a 9.4 T spectrometer under different acquisition conditions to obtain the NMR spectrum of solids under optimized conditions.The following parameters were varied: rotation speed (n r , Hz), recycle time (D1, s), and contact time (t c , µs).

Rotation Speed Optimization
The rotation speed was optimized using our 4 mm probe, the maximum speed of which is 14 kHz.The speed was varied from 8000 to 12,000 Hz. Figure S1 (Supplementary Material) shows the spectra of the mefloquine hydrochloride sample, obtained at two speeds: 9800 and 10,000 Hz.The speed chosen was 9800 Hz, as it presented rotational bands that did not overlap with the isotropic signals of mefloquine hydrochloride.

Recycle Time Optimization
13 C CPMAS NMR spectra with recycle time variation were acquired in a contact time of 1100 µs following the work of Guo et al. [28], varying the recycle time (D1) in the range of 3 to 20 s.The spectra obtained are shown in Figure S2 (Supplementary Material).The recycling time chosen was 20 s.

Contact Time Optimization
The last optimized parameter was the contact time in the cross-polarization pulse sequence.The 13 C CPMAS NMR spectra were obtained with a recycle time of 20 s, and the contact times were varied between 20 and 50,000 µs. Figure S3 (Supplementary Material) shows the results achieved.Under optimized conditions (Figure 3), the spectrum obtained in this work was comparable to that published by Guo et al. [28], which employed a 20 T spectrometer with an MAS frequency of 40,000 Hz.The aliphatic carbon C11 (C-OH) signal was observed at 67.8 ppm in MQ; in MQ-R, it was split into three distinct, low-intensity signals at 70.2 ppm, 68.6 ppm, and 66.9 ppm.Table 1 lists the chemical shifts observed for the samples MQ, R, and MQ-R.Chemical shifts previously published in the literature for MQ in solid phase were also included for comparison, as were the 13 C data of MQ in DMSO solution.The corresponding spectrum is depicted in Figure S6 (Supplementary Material).

Carbon Assignment
Mefloquine Hydrochloride (MQ) The last optimized parameter was the contact time in the cross-polarization pulse sequence.The 13 C CPMAS NMR spectra were obtained with a recycle time of 20 s, and the contact times were varied between 20 and 50,000 µs. Figure S3 (Supplementary Material) shows the results achieved.Under optimized conditions (Figure 3), the spectrum obtained in this work was comparable to that published by Guo et al. [28], which employed a 20 T spectrometer with an MAS frequency of 40,000 Hz.  1.

Characterization of the Ion-Exchange Resin Sample: Polacrilin Potassium (Sample R)
13 C CPMAS NMR spectra of sample R were obtained at contact times in the 200 to 20,000 µs range, with a recycle time of 20 s. Figure S4 (Supplementary Material) shows the spectra obtained.The signals observed correspond to the CO2 − K + groups in the region between 190 and 180 ppm, to the aromatic ring carbons, around 150 ppm (C1) and in the region from 120 to 135 ppm (C2), and to the C, CH, CH2, and CH3 groups of the aliphatic chain of the resin from the methacrylate portion of the copolymer.

Characterization of the Mefloquine Resinate Sample (MQ-R Sample)
The mefloquine-resin complex was prepared in line with previously published work [16]. 13C CPMAS NMR spectra of the mefloquine resinate sample were obtained under the same contact time conditions as the MQ sample (Figure S5, Supplementary Material).Significant changes in the chemical shifts of MQ in MQ-R were observed.Compared to what was observed in the pure MQ (Figure 3), some signals shifted for higher frequency (deshielding effect), and the unfolding of some signals was observed, as summarized in the bar graph of Figure 4.  1.

Characterization of the Ion-Exchange Resin Sample: Polacrilin Potassium (Sample R)
13 C CPMAS NMR spectra of sample R were obtained at contact times in the 200 to 20,000 µs range, with a recycle time of 20 s. Figure S4 (Supplementary Material) shows the spectra obtained.The signals observed correspond to the CO 2 − K + groups in the region between 190 and 180 ppm, to the aromatic ring carbons, around 150 ppm (C1) and in the region from 120 to 135 ppm (C2), and to the C, CH, CH 2 , and CH 3 groups of the aliphatic chain of the resin from the methacrylate portion of the copolymer.

Characterization of the Mefloquine Resinate Sample (MQ-R Sample)
The mefloquine-resin complex was prepared in line with previously published work [16]. 13C CPMAS NMR spectra of the mefloquine resinate sample were obtained under the same contact time conditions as the MQ sample (Figure S5, Supplementary Material).Significant changes in the chemical shifts of MQ in MQ-R were observed.Compared to what was observed in the pure MQ (Figure 3), some signals shifted for higher frequency (deshielding effect), and the unfolding of some signals was observed, as summarized in the bar graph of Figure 4.
Figure 5 shows the corresponding 13 C CPMAS NMR spectra.A physical mixture of MQ and R (denoted MQ + R) was also prepared, and the 13 C CPMAS spectrum was included for comparison.Sample MQ-R is distinct, presenting shifted and new signals, compared with samples MQ and R, whereas sample MQ + R showed the same signals as the samples MQ and R, but some were superposed.Figure 5 shows the corresponding 13 C CPMAS NMR spectra.A physical mixture of MQ and R (denoted MQ + R) was also prepared, and the 13 C CPMAS spectrum was included for comparison.Sample MQ-R is distinct, presenting shifted and new signals, compared with samples MQ and R, whereas sample MQ + R showed the same signals as the samples MQ and R, but some were superposed.Some variations in chemical shifts, both in the carboxylate group of the resin and in the carbons present in the solid mefloquine structure, were also evident.The following can be highlighted: • A broadening of the signal corresponding to the CO 2 − groups: in R, it was observed at 186. 5  The aliphatic carbon C11 (C-OH) signal was observed at 67.8 ppm in MQ; in MQ-R, it was split into three distinct, low-intensity signals at 70.2 ppm, 68.6 ppm, and 66.9 ppm.Table 1 lists the chemical shifts observed for the samples MQ, R, and MQ-R.Chemical shifts previously published in the literature for MQ in solid phase were also included for comparison, as were the 13 C data of MQ in DMSO solution.The corresponding spectrum is depicted in Figure S6 (Supplementary Material).

Evaluation of Molecular Dynamics through T CH and T 1ρ H Measurements
To investigate the interaction between mefloquine and polacrilin resin, a measure of T CH (time constant of signal growth) and T 1ρ H (time constant of protons' relaxation in the rotatory axis) by cross-polarization pulse sequence was performed.The bar diagram in Figure S7 (Supplementary Material) shows the T CH variation obtained for MQ and R carbons.For the carbons of MQ, T 1ρ H relaxation times are in the range of 100-200 µs (Table 2).In contrast, for the carbons of polacrilin resin (R), T 1ρ H relaxation times showed a shorter value, between 9 and 11.7 ms (Table 3), thus evidencing that the carbons in polacrilin are rigid in the structure lattice.Conversely, the T CH values depend primarily on the carbon type: in MQ, it varies from 300 to 500 µs for CH 2 and CH sp 3 carbons of the piperazine ring and >700 µs for aromatic and pyridine CH and C. In sample R, the T CH values were between 99 and 640 µs.
The bar diagram depicted in Figure 6 illustrates the results of molecular dynamics probed by T 1ρ H relaxation values, listed in Tables 2 and 3 for each carbon of MQ, MQ-R, and R. Sample MQ-R presented T 1ρ H relaxation times for carbons in the range of 14.8 to 68.4 µs, highly contrasting with the values measured for MQ.On the other hand, those values are in the same order of magnitude as polacrilin resin R. Figure 7 depicts the curves of intensity variation with contact times obtained for carbon C2 in MQ USP and MQ-R samples.Figures S8-S12 (Supplementary Material) illustrate the dependence of C signal intensity on the contact time for carbons C2, C7, C5, C6, and C12 in MQ and MQ-R samples, respectively.The curves obtained for the carbons of sample R are shown in Figure S13 (Supplementary Material).
2).In contrast, for the carbons of polacrilin resin (R), T1ρ H relaxation times showed a shorter value, between 9 and 11.7 ms (Table 3), thus evidencing that the carbons in polacrilin are rigid in the structure la ice.Conversely, the TCH values depend primarily on the carbon type: in MQ, it varies from 300 to 500 µs for CH2 and CH sp 3 carbons of the piperazine ring and >700 µs for aromatic and pyridine CH and C. In sample R, the TCH values were between 99 and 640 µs.The bar diagram depicted in Figure 6 illustrates the results of molecular dynamics probed by T1ρ H relaxation values, listed in Tables 2 and 3 for each carbon of MQ, MQ-R, and R. Sample MQ-R presented T1ρ H relaxation times for carbons in the range of 14.8 to 68.4 µs, highly contrasting with the values measured for MQ.On the other hand, those values are in the same order of magnitude as polacrilin resin R. Figure 7 depicts the curves of intensity variation with contact times obtained for carbon C2 in MQ USP and MQ-R samples.Figures S8-S12  samples, respectively.The curves obtained for the carbons of sample R are shown in Figure S13 (Supplementary Material).

In Vitro Taste-Masking Evaluation
Electrical impedance measurements were performed at 1 kHz for 5 min in the simulated salivary solution.Considering the difference between the response of the resin in the presence (Z) of MQ and the absence of the mefloquine drug (Z0), the drug release kinetic was monitored (Figure 8).A theoretical exponential fi ing of the experimental data suggests that most of the drug was released from the MQ-R sample within 3.4 min.Such

MQ carbons
MQ-R carbons R carbons

In Vitro Taste-Masking Evaluation
Electrical impedance measurements were performed at 1 kHz for 5 min in the simulated salivary solution.Considering the difference between the response of the resin in the presence (Z) of MQ and the absence of the mefloquine drug (Z 0 ), the drug release kinetic was monitored (Figure 8).A theoretical exponential fitting of the experimental data suggests that most of the drug was released from the MQ-R sample within 3.4 min.Such liberation decreased the solution conductivity from ~753 to 827 Ω, with a dispersion of ~5% for two replicates.liberation decreased the solution conductivity from ~753 to 827 Ω, with a dispersion of ~5% for two replicates.
Figure 8.The dissolution test for mefloquine resinate in simulated salivary conditions was performed using the impedance method, measured through the conductivity difference between MQ-R (|Z|) and R (|Z0|).
An impedimetric electronic tongue was also used to assess the taste-masking effectiveness and confirm the dissolving results.The magnitude impedance spectra of mefloquine (MQ), polacrilin potassium resin (R), and the mefloquine resinate (MQ-R) were projected using interactive document mapping (IDMAP) and hierarchical cluster analysis (HCA) pa ern recognition methods [16].The assay was carried out using an array An impedimetric electronic tongue was also used to assess the taste-masking effectiveness and confirm the dissolving results.The magnitude impedance spectra of mefloquine (MQ), polacrilin potassium resin (R), and the mefloquine resinate (MQ-R) were projected using interactive document mapping (IDMAP) and hierarchical cluster analysis (HCA) pattern recognition methods [16].The assay was carried out using an array of six sensing units.The IDMAP plot, Figure 9A, shows a clear distinction of the three solutions, which is the reason for achieving a high silhouette coefficient value of 0.96 (from 0.71 to 1, a high data correlation is obtained).The drug and the resin clusters are displaced on the opposite edges of the map.Meanwhile, the MQ-R complex is in the center.However, the distance between MQ and MQ-R is smaller than between MQ-R and R, which suggests that the bitter taste of mefloquine was not completely masked.The Euclidean distances between the clusters are projected in Figure 9B through the dendrogram, confirming the last result.
Figure 8.The dissolution test for mefloquine resinate in simulated salivary conditions was performed using the impedance method, measured through the conductivity difference between MQ-R (|Z|) and R (|Z0|).
An impedimetric electronic tongue was also used to assess the taste-masking effectiveness and confirm the dissolving results.The magnitude impedance spectra of mefloquine (MQ), polacrilin potassium resin (R), and the mefloquine resinate (MQ-R) were projected using interactive document mapping (IDMAP) and hierarchical cluster analysis (HCA) pa ern recognition methods [16].The assay was carried out using an array of six sensing units.The IDMAP plot, Figure 9A, shows a clear distinction of the three solutions, which is the reason for achieving a high silhoue e coefficient value of 0.96 (from 0.71 to 1, a high data correlation is obtained).The drug and the resin clusters are displaced on the opposite edges of the map.Meanwhile, the MQ-R complex is in the center.However, the distance between MQ and MQ-R is smaller than between MQ-R and R, which suggests that the bi er taste of mefloquine was not completely masked.The Euclidean distances between the clusters are projected in Figure 9B through the dendrogram, confirming the last result.

Discussion
Mefloquine (MQ) is an essential drug that requires a slightly acidic medium to ionize its salt.This ionization allows the drug to complex with the ion-exchange resin polacrilin potassium and form mefloquine resinate (MQ-R).At pH 5.0, the resin displayed a Drug Loading Efficiency of 99.64 ± 0.14%, expressed as mefloquine hydrochloride.The pKa range of ion-exchange resins with carboxylic acid groups is between 4.0 and 6.0 [29].This range is also where MQ is ionized (pKa 8.6), facilitating the formation of resinate complexes. 13C CPMAS NMR spectra of the MQ were acquired under different acquisition conditions to obtain the NMR spectrum of solids under optimized conditions.The following parameters were varied: rotation speed (nr, Hz), recycle time (D1, s), and contact time (tc, µs).Solution 13 C NMR spectrum (DMSO-d 6 , a polar aprotic solvent) was also acquired for comparison.Prado et al. [1] employed a rotation speed of 15 kHz to obtain the 13 C CPMAS NMR spectrum of mefloquine, and Guo et al. [28] used probes that allowed the sample to be rotated between 20 and 60 kHz.Previous experiments indicate that the optimized spin rate for MQ was 9500 to 10,500 Hz.The speed chosen was 9800 Hz, as it presented rotational bands that did not overlap with the isotropic signals of mefloquine hydrochloride.It is worth mentioning that the spectrum obtained was comparable to that published by Guo et al. [28]. 13C CPMAS NMR spectra with recycle time variation were acquired in a contact time of 1100 µs following the work of Guo et al. [28], varying the recycle time (D1) from 3 to 20 s.Although a recycle time of 5 s was used in the literature, more resolved signals were obtained with 20 s, so this value was chosen in this work.
The last optimized parameter was the contact time in the cross-polarization pulse sequence.The 13 C CPMAS NMR spectra were obtained with a recycle time of 20 s, and the contact times were varied between 20 and 50,000 µs.Prado et al. [1] obtained the spectrum of mefloquine hydrochloride with a contact time of 1500 µs; the authors did not visualize the signals corresponding to the CF 3 carbons of the mefloquine structure in the expected position in the highest frequency region of the spectrum.Guo et al. [28] published data on mefloquine's solid-state 13 C NMR spectra, with a rotation speed of 20 kHz and contact times between 0.5 and 1.4 µs.Under optimized conditions (Figure 3), the spectrum obtained in this work was comparable to that published by Guo et al. [28].It was found that the signals from carbons linked to fluorine, in the region between 115 and 125 ppm, could be identified in the spectrum obtained with a longer contact time.Also, in contrast with the published spectrum, the carbon signals of the piperidine ring (C15, C16, and C17) could be seen with ν r 9800 Hz, D1 20 s, and t c 10,000 µs.
Polacrilin potassium is the salt of a crosslinked polymer derived from methacrylic acid, resulting in a weakly acidic form of a cationic ion-exchange resin.The fact that the spectrum with the highest intensity signals was observed in a shorter contact time range (Figure S4) confirms the rigid structure of the polymeric resin analyzed.The spectra of MQ-R showed significant changes in the chemical shifts in comparison with MQ sample spectra.As can be seen in Table 1, the presence of more than one signal for carbons C4, C5, C6, C12, and C14 in the 13 C CPMAS NMR spectrum of MQ suggests that in the sample analyzed, there is a mixture of diastereoisomers since, in the case of this compound, only one crystalline structure is reported [28].In the case of C12 and C14 carbons, the presence of more than one signal would also result from dipolar interaction with the quadrupolar 14 N neighbor.Conversely, in the spectrum of MQ-R, the carbons C4, C2, C8a, and C11 were unfolded in two or three signals.That result would indicate that the diastereoisomers interact with the resin in distinct forms that can be detected in the solid-state 13 C spectrum.Comparing the chemical shifts of MQ in solution with the ones obtained for solid MQ and MQ-R complexes, it could be observed that the packing effect in MQ solid vanishes in the sample MQ-R: the chemical shifts of carbons C8, C4a, C7, and C6 in MQ-R are somewhat more shielded to the ones in MQ in DMSO solution (Figure 4, green bars).From that result, we can suggest that in a 1:1 MQ-R complex, the MQ is highly dispersed in the R moiety through an electron donation mechanism from resin to MQ structure that shields the trifluoromethyl benzenoid ring.The presence of R also affected MQ's molecular dynamics.In the pure MQ, an increase in contact time caused an increase in signal intensities until 10,000 contact time.Still, in contrast, in the case of MQ-R, it was possible to observe a decrease in the signal intensities of the MQ carbons with the increased contact time, with molecular dynamics like those presented by pure resin.Distinct molecular dynamics were established between MQ and resin by observing the curves of variation of intensity of signals in the function of contact time t c (Figures S8-S12, Supplementary Material).For MQ, the maximum is reached at around 10,000 µs, whereas for polacrilin resin, this value is about 1000 µs, except for COO -carbon (Figure S13, Supplementary Material).Strongly dipolar-coupled hydrogen nuclei in rigid solids tend to have similar T 1ρ H values because spin diffusion is fast on the NMR timescale [21].However, T CH is determined by the local strength of the dipolar coupling between C and H and is specific to the C environment.We observed that for the non-protonated C2 and C8a of MQ, the T CH values were higher than the protonated sp 2 carbons.On the other hand, following the theory, the sp 3 carbons presented the lowest T CH in the 300 to 500 s range.T CH values decreased for all the carbons measured for the MQ-R complex except for carbon C5 (Table 2).The most prominent decrease was observed for CH sp 3 carbons C11 and C12.That result suggests that MQ in MQ-R strongly interacts with the resin structure through a proton spin diffusion mechanism.Durand et al. [25] showed that the non-covalent interactions between mefloquine and cyclodextrins enhanced the solubility of the inclusion complexes thus formed in solution.Further studies will be conducted to establish the structure of the MQ-R complex.The electronic tongue system evaluated the taste-masking efficiency of the MQ, R (resin), and resinate (MQ-R).The impedance spectra of all three samples were analyzed using IDMAP and HCA pattern recognition techniques.The IDMAP method further converted the dissimilarities between each spectrum to Euclidean distance and discriminated the clusters formed by resinate (MQ-R) from those of MQ.The distance between the cluster of resinate and that of the drug increased with the effectiveness of the taste masking.The results pointed out that the bitter taste of MQ was not wholly masked, but it was sufficiently compelling to make MQ tolerable for pediatric patients.This method has been considered effective in evaluating drug taste masking [16,30] and can be applied in the early stages of product development.

Mefloquine Resinate Preparation
Equal amounts of MQ and resin (drug: resin ratio 1:1 w:w) were weighed and suspended in an HCl aqueous solution (pH 5.0) and kept under magnetic stirring for 24 h at room temperature.Resinates were separated by vacuum filtration and dried in a vacuum oven Memmert VO 200 (Memmert GmbH + Co.KG, Schwabach, Germany) at 60 • C and 1100 mbar for 48 h.The supernatant was filtered through a 0.45-µm membrane filter and measured by UV-visible spectrophotometry at λ = 222 nm [31].The drug loading efficiency (%) was calculated by comparing the amount of the drug before and after complex formation, according to Equation (2).The samples, as received, were transferred to zirconium oxide rotors (4 mm diameter).After weighing, the rotors were closed using 4 mm Kel-F caps. 13C Solid NMR Spectra 13 C Solid-state NMR spectra were acquired using Bruker equipment, model Avance III WB400, 9.4 T, equipped with a 4 mm triple channel probe.The spectra acquisition conditions were as follows: pulse sequence-cross-polarization magic angle spinning; spinning rate-between 8 and 12 kHz; recycle time-between 3 and 20 s; contact time for cross-polarization-between 0.02 and 50 ms.During the signal acquisition step, a two-pulse phase-modulated (TPPM) sequence was used for proton decoupling [32].Glycine was used as a reference (C=O 176.03 ppm).

Measurement of T CH and T 1ρ H Relaxation Times
The relaxation time values were measured indirectly through the cross-polarization sequence.To this end, 13 C CPMAS NMR spectra were acquired, with variation in contact times ranging from 20 µs to 50,000 µs, rotation speed of 9800 Hz, and recycle time of 20 s.Bruker's Dynamics Center data processing platform version 2.8.3 (12 January 2023) was used to obtain the signal growth time constants and the decay time constant of the signals on the rotational axis for the different signals.
Relaxation time values for all carbons were possible for the mefloquine hydrochloride sample.The same happened for the resin.In the case of the mefloquine resinate sample, the overlap of some MEF signals with resin signals restricted the range of measurements.

Conclusions
The 13 C CPMAS NMR spectra under different acquisition conditions allow us to characterize the samples of mefloquine hydrochloride, polacrilin resin, and mefloquine resinate.The comparison of the T 1 ρ H values for the carbons of mefloquine, ion-exchange resin, and a mefloquine-resin complex allows us to confirm the absence of a physical mixture and the presence of the complex in a formulated sample of mefloquine-resin.Furthermore, the electronic tongue is a viable economical alternative for drug development screening.The results show that the API mefloquine could be distinguished from complex resinate and ion-exchange resin, proving that the taste masking was effective.Although it did not completely cover the API, it was sufficiently compelling to make MQ tolerable for pediatric patients.The mefloquine resinate complex presents itself as a feasible option for potential pediatric formulations in the future.The incorporation of excipients such as sweeteners and flavorings into the ultimate formulation will notably contribute to masking the inherently bitter taste of mefloquine.

Figure 4 .
Figure 4. 13 C chemical shifts variation (D chemical shift = dMQ-R − dMQ) observed for MQ carbon skeleton in DMSO-d6 solution (green), and in the solid state (orange).The change for higher frequency (deshielding) is denoted by the positive bar; the shift for lower frequency (shielding) is denoted by the negative bar.DMQ chemical shifts (±1 ppm interval) for each carbon signal are indicated by the horizontal lines in blue.

Figure 5
Figure 5  shows the corresponding13 C CPMAS NMR spectra.A physical mixture of MQ and R (denoted MQ + R) was also prepared, and the 13 C CPMAS spectrum was included for comparison.Sample MQ-R is distinct, presenting shifted and new signals, compared with samples MQ and R, whereas sample MQ + R showed the same signals as the samples MQ and R, but some were superposed.

Figure 4 .
Figure 4. 13 C chemical shifts variation (D chemical shift = d MQ-R − d MQ ) observed for MQ carbon skeleton in DMSO-d 6 solution (green), and in the solid state (orange).The change for higher frequency (deshielding) is denoted by the positive bar; the shift for lower frequency (shielding) is denoted by the negative bar.DMQ chemical shifts (±1 ppm interval) for each carbon signal are indicated by the horizontal lines in blue.

Figure 4 .
Figure 4. 13 C chemical shifts variation (D chemical shift = dMQ-R − dMQ) observed for MQ carbon skeleton in DMSO-d6 solution (green), and in the solid state (orange).The change for higher frequency (deshielding) is denoted by the positive bar; the shift for lower frequency (shielding) is denoted by the negative bar.DMQ chemical shifts (±1 ppm interval) for each carbon signal are indicated by the horizontal lines in blue.

Figure 7 .
Figure 7. Variation of the intensity of carbon C2 signal with the cross-polarization contact time (tc, s) for samples of mefloquine hydrochloride (MQ, pink) and mefloquine resinate (MQ-R, blue).The curve obtained for polacrilin carbons at 47 ppm (R, black) was included for comparison.The vertical scale is arbitrary.

Figure 7 .
Figure 7. Variation of the intensity of carbon C2 signal with the cross-polarization contact time (tc, s) for samples of mefloquine hydrochloride (MQ, pink) and mefloquine resinate (MQ-R, blue).The curve obtained for polacrilin carbons at 47 ppm (R, black) was included for comparison.The vertical scale is arbitrary.

Figure 8 .
Figure 8.The dissolution test for mefloquine resinate in simulated salivary conditions was performed using the impedance method, measured through the conductivity difference between MQ-R (|Z|) and R (|Z 0 |).

Figure 9 .
Figure 9.The taste-masking efficiency was determined with an electronic tongue analysis.(A) IDMAP plot points to the clusters formed by the mefloquine (MQ), the resin (R), and the mefloquine

Figure 9 .
Figure 9.The taste-masking efficiency was determined with an electronic tongue analysis.(A) IDMAP plot points to the clusters formed by the mefloquine (MQ), the resin (R), and the mefloquine resinate (MQ-R).(B) Dendrogram with the Euclidean distances and similarities between the clusters formed.
7nd 185.6 ppm, and in MQ-R, between 187.7and 181.6 ppm; • In MQ, the signals of aromatic carbons, non-protonated C4 (unfolded in two, at 148.5 and 148.1 ppm), C2 linked to N and the CF 3 group (at 145.2 ppm), and C8a of ring junction and linked to the N of the nitrogen ring (at 143.8 ppm), moved to 155, 150.6 ppm, 148.1, 147.2, and 143.3 ppm, respectively; • In MQ, the signals from carbons C7, C5, and C6 of the benzene ring at 130.1, 127.1, and 125.2 ppm, respectively, shifted to a lower frequency, at 128 and 122.9 ppm in the MQ-R sample; • The narrow signal from the protonated aromatic carbon C3, observed at 114.4 ppm in MQ, was observed as a broad signal at 115.2 ppm in the MQ-R sample.

Table 3 .
Chemical shifts (δ, ppm), T CH , and T 1ρ H relaxation times obtained for carbons of polacrilin potassium resin in sample R and a mefloquine resinate (MQ-R) sample.

Table 3 .
Chemical shifts (δ, ppm), TCH, and T1ρ H relaxation times obtained for carbons of polacrilin potassium resin in sample R and a mefloquine resinate (MQ-R) sample.
(Supplementary Material) illustrate the dependence of C signal intensity on the contact time for carbons C2, C7, C5, C6, and C12 in MQ and MQ-R n.d.-not determined.ls 2024, 17, x FOR PEER REVIEW 9 of 16