Category Archives: Resource Library

The goal of this presentation is to show multiple ways to increase DMPK experimental efficacy. An enhancement of efficacy in this context is understood as how to get more knowledge out of the data that is produced. This process is translated on how to decrease the time from data to compound design and/or how to increase the knowledge learned from the data. To achieve this objective automation of data processing, maximizing the extraction of information out of the data and the translation the newly acquired information into practical design is necessary.

Methods

To increase the automation a new methodology will be shown where the input experimental files acquired on a Thermo Q-Exactive instrument for a clearance workflow are processed without human intervention to obtain the clearance and metabolite identification results. The procedure will be exemplified in a collection of 20 PROTACS structures where the Clearance measurement in Hepatocytes has been performed. Clearance end point data computation will be shown together with the automatic fragment assignment for each of the metabolites.

Preliminary data

For example, the fragmentation observed and assigned for a particular compound was used to develop a model to predict the compound fragmentation. In this context several published spectra interpretation from the MoNA database have been used to determine a Machine Learning model to predict the probability of a bond to be broken under certain Mass Spectrometry condition. That publicly available information is used to derive a model that was validated and monitored by using the own experimental data generated in the automatic procedure mentioned before. Considering the result from the comparison between the experimental and the prediction values the new data was incorporated into an automatically learning process to improve model quality. In this way a self-learning system has been developed for the improvement of the virtual fragmentation of a compound.

The above self learning procedure was also applied for other Mass Spectrometry derived property like for example the Cross Collision section (CCS) value to assess the reliability of a structural assignment to a certain peak. In this case a regression model with a regression coefficient between the observed and the calculated CCS values of 0.98 was obtained after developing a model based on published data and enriched with our own dataset.

Finally, an example of how the metabolite identification information extracted from the previously described automation system is used to design compound with a reduction in compound clearance will be shown.

Calibration lines in drug quantification workflows are traditionally determined using QQQ instruments that have high sensitivity and where the ion selectivity is gained by using multiple reaction monitoring. This workflow needs a previously identify transition to be monitored to obtain the needed selectivity for the compound of interest. In this work we demonstrate an automated methodology to perform quantification studies based on high-resolution mass spectra data. We obtain selectivity based on the high resolution, low ppm difference between the observed and the theoretical m/z computed from the compound formula. We gain the needed sensitivity by automatically considering multiple m/z ions, including adducts and fragments, in the computation of the peak areas.

Methods

The methodology has been developed using MSe data acquired on a Waters Xevo G2XS Qtof. A peak analysis is done in the high and low collision energy traces. The m/z of the each of the identified peaks in both MS functions are then compared to the m/z obtained for the parent molecular formula, the multiple charges, the adducts and/or the theoretically generated fragments. The list of experimental peaks that matches any of computed m/z is then submitted to the quality peak analysis (criteria: number of points/peak, ppm, signal/noise ratio, etc.). The area for each of the m/z is computed (isolated or in combination) using an auto-adjusting extraction window and later used in a regression analysis with multiple acceptance criteria.

Preliminary data

The data for the working example is obtained for a compound at 13 different concentration levels: 1, 2, 5, 10, 20, 50,1 00, 200, 500,1000, 2000, 5000, 10000 ng/mL measured in triplicate. The peak quality criteria were based on a minimum of eight scan points per peak, a difference in Retention Time between samples lower than 0.05 min, the ions with a difference between the observed and the computed m/z lower than 15 ppm.  The extraction window was auto-adjustable, the noise evaluation time range (peak units) was set to 6, a the Minimum signal/noise ratio to 3. The maximum variation of nominal concentration of each sample and the average, as well as the maximum CV was set to 25% for a point to be accepted for regression. For a line to be accepted a minimum number of concentration levels was set to 6 with a minimum replication of 2 and a maximum fold change between consecutive points was 10. Three lines can be derived to cover the wider dynamic range, the LLoQ and the ULOQ using different weighting factors. In the example that will be shown the best fragments for quantification are obtained even in absence of the M+H ion due to in-source fragmentation. The regression line obtained covered the dynamic range from 2 to 10000 with all the quality controls fulfilled with a combination of two m/z. It is also observed that due to the different ion intensity and noise evaluation for each ion it was necessary to perform an automatic optimization of parameter for the best peak quality.

A commonly used strategy in the peptide therapeutics field to introduce chemical modifications such as cyclisation, changing the stereochemistry of an amino acid, substitution of natural amino acids to chemically modified ones and others to improve their efficacy and ADME profile. The study of the metabolic degradation products for synthetically modified therapeutic peptides using LC-MS is a challenging issue. Different chemoinformatics approaches are used for automated metabolite identification. These tools propose metabolite structures based on the combination of metabolite prediction and analysis of MS data. This makes building the virtual set of all potential metabolites time and computational intensive. Fragmentation analysis requires even more computation time. Finally, a third challenge related to the depiction of the parent and the metabolites.

Methods

New algorithms that address the challenges in highly modified therapeutic peptide structure elucidation have been developed. The peak detection algorithm was improved to use the Most Abundant Isotope for parent and potential metabolites. A new algorithm is producing all virtual metabolites applying the library of chemical reactions to each monomer while maintaining the connection to the atoms. A third improvement is the fragmentation algorithm that has two layers of analysis, one at the monomer level and the other one at the bond level. Finally, we will also show the results of the implemented algorithm for the analysis results visualization.

Preliminary data

Using MassMetaSite we analyzed a collection of experimental data for a set of peptides where the data was collected on a Q-Exactive Thermo instrument. The metabolite identification study was performed using a peptide set that included eight compounds: somatostatin and its seven synthetic analogues. All test compounds were incubated in serum. These peptides are all cyclic peptides and seven of them had unnatural amino acids. The structural assignments were performed for 17 degradation products with high mass accuracy (ppm<3). Most of the metabolites resulting from one or two metabolic modifications were produced by amide hydrolysis and were detected by the new algorithm. Each of these compounds showed a suitable fragmentation pattern that could be compared to the parent by the fragmentation algorithm to assign the shifted and non-shifted fragments. All the metabolites were checked manually including a review of the assigned fragments. Metabolites were considered as reliable because the fragmentation was adequate, isotope pattern was as expected, the small differences between the m/z of observed and theoretical, and the mass score was high. As previously reported in the literature, the first two most abundant metabolites with assigned structure correspond to cleavage of the linear part of the somatostatin. Finally, we analyzed experimental data for the insulin and semaglutide data where the data was collected on a Waters QToF. Insulin is a cyclic peptide that contains 3 disulfide bridges and semaglutide is a linear peptide that contains linkages and fatty acid. The visualization algorithm developed allows to show results for these complex structures in a such a manner that it is easy to interpret due to the constraint structure alignment between the substrate and the metabolite (keeping same orientation), a possibility to combine monomer and atom/bond notation. Therefore, the metabolic changes in the structure can be easily seen by the User.

Collisional-induced dissociation (CID) has been the main workhorse for small molecule structure elucidation in drug metabolite identification studies.  Software tools, such as Massmetasite, have been developed to assist in the automatic interpretation of CID MS/MS spectra.  The challenge with CID is that for many drug metabolites, non-informative fragmentation occurs, resulting in a lack of useful structural assignments for these metabolites.  Commercialization of electron activated dissociation (EAD) on a quadrupole time of flight mass spectrometer provides a potentially powerful new tool for small molecule structure elucidation in drug discovery.  Automated interpretation of EAD MS/MS spectra using existing algorithms needs to be tested, modified, and implemented. A comparison of CID and EAD fragmentation using automated interpretation will be presented in this work.

Methods

Small molecule drugs were incubated in rat hepatocytes at 1 µM. Time points:  0, 30 and 120 min were pulled from the incubation and quenched with 1 volume of CH₃CN. Samples were vortexed, centrifuged, and the supernatant transferred to an HPLC vial for analysis.

LC separation was performed on a Phenomenex Luna Omega Polar C18, 150 mm column using 0.5µL or 5 µL injection volumes. Gradient separation 0.1% formic acid in water and acetonitrile was performed over 4.75 minutes from 5%B to 95%B with a total of 6.5 minutes.

The samples were analyzed using ZenoTOF 7600 CID-IDA(DDA) and EIEIO IDA(DDA). TOFMS was scanned between m/z 100-1000, CID/EAD MS/MS from 60-1000.  Data was processed in MassMetaSite with CID and EAD fragmentation.

Preliminary data

EAD is a free electron fragmentation mode available recently introduced to accurate mass LC-MS/MS. It utilizes high energy electrons which allows for the dissociation of singly charged precursors, in this work an electron kinetic energy of 10 eV was utilized.  The MS/MS spectra show significant increases in the number of fragment ions observed when going from Zeno CID to Zeno EAD spectra.  The larger number of fragment ions makes it even more important for automated assignment using software tools such as Massmetasite.  Many of the new fragments are the result of radical bond cleavage driven by electron-impact excitation of ions from organics (EIEIO) mechanism.  Typical software algorithms for MS/MS focus on even electron species, typical of CID fragmentation.  With EAD and the production of odd electron fragments, modification of the algorithm is required.  The results to date show much richer fragmentation spectra using EAD versus CID and that the modified algorithm can assign the new odd electron fragments.