Investigation of Praziquantel/Cyclodextrin Inclusion Complexation by NMR and LC-HRMS/MS: Mechanism, Solubility, Chemical Stability, and Degradation Products

Praziquantel (PZQ) is a biopharmaceutical classification system (BCS) class II anthelmintic drug characterized by poor solubility and a bitter taste, both of which can be addressed by inclusion complexation with cyclodextrins (CD). In this work, a comprehensive investigation of praziquantel/cyclodextrin (PZQ/CD) complexes was conducted by means of UV–vis spectroscopy, spectrofluorimetry, NMR spectroscopy, liquid chromatography-high-resolution mass spectrometry (LC-HRMS/MS), and molecular modeling. Phase solubility studies revealed that among four CDs tested, the randomly methylated β-CD (RMβCD) and the sulfobutylether sodium salt β-CD (SBEβCD) resulted in the highest increase in PZQ solubility (approximately 16-fold). The formation of 1:1 inclusion complexes was confirmed by HRMS, NMR, and molecular modeling. Both cyclohexane and the central pyrazino ring, as well as an aromatic part of PZQ are included in the CD central cavity through several different binding modes, which exist simultaneously. Furthermore, the influence of CDs on PZQ stability was investigated in solution (HCl, NaOH, H2O2) and in the solid state (accelerated degradation, photostability) by ultra-high-performance liquid chromatography–diode array detection–tandem mass spectrometry (UPLC-DAD/MS). CD complexation promoted new degradation pathways of the drug. In addition to three already known PZQ degradants, seven new degradation products were identified (m/z 148, 215, 217, 301, 327, 343, and 378) and their structures were proposed based on HRMS/MS data. Solid complexes were prepared by mechanochemical activation, a solvent-free and ecologically acceptable method.

1. Phase solubility 1     3. Full assignment of praziquantel in DMSO-d 6 NMR analysis of 1D and 2D NMR spectra recorded in DMSO-d 6 at 25 °C revealed two sets of resonance lines which both correspond to the proposed structure ( Figure S5), their ratio being 55% : 45%. Variable temperature experiments showed coalescence of most signals at 80 °C ( Figure S6), showing that two sets of resonance lines belong to two conformational isomers in chemical exchange. This conclusion was further corroborated by exchange peaks between the two conformers observed in NOESY spectrum ( Figure S7). When comparing the two isomers, it immediately comes to view that the biggest difference in chemical shifts is around N13-C15=O amide bond, pointing towards these conformational isomers being the consequence of slow rotation around the amide bond.

.1. Validation of methods
Chemical shift for C12 in major rotamer (ca 55%) is 45.6 ppm, which is lower than the chemical shift of the same atom (48.4 ppm) in minor isomer (ca 45%). This suggests that amide oxygen is in vicinity of C12 in case of major rotamer. [1] Similarly, chemical shift of C14 (44.4 ppm) in minor rotamer is lower than its counterpart in major rotamer (48.0 ppm). This corresponds oxygen atom being oriented towards C-14 in minor isomer.

Chemical shift comparisons
Comparison of proton and carbon chemical shifts was performed using the DMSO-d 6 solutions of PZQ, β-CD and praziquantel/β-CD mixture obtained as solid from the mill, showing no difference in resonance lines in either proton, or carbon spectra. The comparison is shown in Figures S8, S9, S10 and S11, the result suggests that the complex was not formed.

Diffusion Spectroscopy (DOSY)
Diffusion spectroscopy was used to measure the diffusion coefficients (D) of all species in the mixture, as well as the percentage of complexation. After the baseline correction, signals were integrated and automatically processed. The fitted curves were examined and only the signals with good fit were taken for further calculations. The omitted signals were overlapped and/or signals with low signal : noise ratio resulting in large integration errors. The final diffusion coefficient was calculated as average of all selected signals. The results are summarised in Table  S2 and graphical depiction is shown in Figure S14.   Table S2, the percentage of complexed PZQ with β-CD in DMSO-d 6 was calculated to be 20%. This low percentage can explain the lack of intermolecular interactions in NOESY/ROESY spectra, as well as no noticeable chemical shifts differences in proton and carbon spectra.    Figure S17) yielded no additional information on the conformation of the free compound. All observed nOe contacts were expected and belong to neighbouring protons. Comparison to the ROESY spectrum of the complex revealed couple of additional nOe interactions, which are probably the result of slight change in chemical shifts resulting in less overlap or less accidental water suppression after the binding.
Analysis of PZQ/β-CD complex ROESY spectrum ( Figure S18) unfortunately did not reveal any intermolecular interactions which would confirm the formation of the complex or the orientation of the PZQ within the β-CD cavity.
NOESY spectrum was also recorded, but showed no usable data, most probably due to the complex undergoing zero-crossing of the signals under experimental conditions.   6. Molecular modeling   7. Linearity of UPLC-DAD method used for identification of the degradation products