Robust and High-Throughput Analytical Flow Proteomics Analysis of Cynomolgus Monkey and Human Matrices With Zeno SWATH Data-Independent Acquisition

Abstract

•Development of a robust and fast proteomics workflow with Zeno SWATH DIA.•Enhanced sensitivity, data quality, and pathway coverage in 15 biological matrices.•Reliable and consistent method over 1000+ uninterrupted injections.•Identification of ∼80% more potential disease biomarkers in clinical plasma samples. Modern mass spectrometers routinely allow deep proteome coverage in a single experiment. These methods are typically operated at nanoflow and microflow regimes, but they often lack throughput and chromatographic robustness, which is critical for large-scale studies. In this context, we have developed, optimized, and benchmarked LC-MS methods combining the robustness and throughput of analytical flow chromatography with the added sensitivity provided by the Zeno trap across a wide range of cynomolgus monkey and human matrices of interest for toxicological studies and clinical biomarker discovery. Sequential Window Acquisition of All Theoretical Fragment Ion Mass Spectra (SWATH) data-independent acquisition (DIA) experiments with Zeno trap activated (Zeno SWATH DIA) provided a clear advantage over conventional SWATH DIA in all sample types tested with improved sensitivity, quantitative robustness, and signal linearity as well as increased protein coverage by up to 9-fold. Using a 10-min gradient chromatography, up to 3300 proteins were identified in tissues at 2 μg peptide load. Importantly, the performance gains with Zeno SWATH translated into better biological pathway representation and improved the ability to identify dysregulated proteins and pathways associated with two metabolic diseases in human plasma. Finally, we demonstrate that this method is highly stable over time with the acquisition of reliable data over the injection of 1000+ samples (14.2 days of uninterrupted acquisition) without the need for human intervention or normalization. Altogether, Zeno SWATH DIA methodology allows fast, sensitive, and robust proteomic workflows using analytical flow and is amenable to large-scale studies. Modern mass spectrometers routinely allow deep proteome coverage in a single experiment. These methods are typically operated at nanoflow and microflow regimes, but they often lack throughput and chromatographic robustness, which is critical for large-scale studies. In this context, we have developed, optimized, and benchmarked LC-MS methods combining the robustness and throughput of analytical flow chromatography with the added sensitivity provided by the Zeno trap across a wide range of cynomolgus monkey and human matrices of interest for toxicological studies and clinical biomarker discovery. Sequential Window Acquisition of All Theoretical Fragment Ion Mass Spectra (SWATH) data-independent acquisition (DIA) experiments with Zeno trap activated (Zeno SWATH DIA) provided a clear advantage over conventional SWATH DIA in all sample types tested with improved sensitivity, quantitative robustness, and signal linearity as well as increased protein coverage by up to 9-fold. Using a 10-min gradient chromatography, up to 3300 proteins were identified in tissues at 2 μg peptide load. Importantly, the performance gains with Zeno SWATH translated into better biological pathway representation and improved the ability to identify dysregulated proteins and pathways associated with two metabolic diseases in human plasma. Finally, we demonstrate that this method is highly stable over time with the acquisition of reliable data over the injection of 1000+ samples (14.2 days of uninterrupted acquisition) without the need for human intervention or normalization. Altogether, Zeno SWATH DIA methodology allows fast, sensitive, and robust proteomic workflows using analytical flow and is amenable to large-scale studies. Recent advances in mass spectrometry (MS)–based proteomics, including instrumentation, chromatography, data acquisition strategies, and processing software, allow high-quality deep proteome coverage in a single-run experiment. These advancements are fueling a growing interest in utilizing those technologies in applications requiring high-throughput phenotypic readouts at scale. Notably, MS-based proteomics is emerging as a key player in drug discovery with the promise to accelerate (i) target identification and validation, (ii) lead discovery and optimization as well as (iii) preclinical and clinical assessments of drug safety and efficacy (1Meissner F. Geddes-McAlister J. Mann M. Bantscheff M. The emerging role of mass spectrometry-based proteomics in drug discovery.Nat. Rev. Drug Discov. 2022; 21: 637-654Crossref PubMed Scopus (41) Google Scholar). Recently, deep proteome profiling of ∼900 cancer cell lines by MS identified thousands of potential biomarkers of cancer vulnerabilities (2Goncalves E. Poulos R.C. Cai Z. Barthorpe S. Manda S.S. Lucas N. et al.Pan-cancer proteomic map of 949 human cell lines.Cancer Cell. 2022; 40: 835-849.e8Abstract Full Text Full Text PDF PubMed Scopus (17) Google Scholar). In addition, proteomics has been widely applied to the field of precision medicine enabling patient stratification (3Messner C.B. Demichev V. Wendisch D. Michalick L. White M. Freiwald A. et al.Ultra-high-throughput clinical proteomics reveals classifiers of COVID-19 infection.Cell Syst. 2020; 11: 11-24.e4Abstract Full Text Full Text PDF PubMed Scopus (285) Google Scholar) and improving patient management (4Demichev V. Tober-Lau P. Lemke O. Nazarenko T. Thibeault C. Whitwell H. et al.A time-resolved proteomic and prognostic map of COVID-19.Cell Syst. 2021; 12: 780-794.e7Abstract Full Text Full Text PDF PubMed Scopus (68) Google Scholar). Among discovery-driven acquisition strategies, data-dependent acquisition (DDA) and data-independent acquisition (DIA) methodologies are the most popular. While in DDA, a predefined number of the most abundant peptides from survey scans are selected for fragmentation, all peptides in a sliding m/z window in MS1 scans are selected for fragmentation in DIA. Therefore, DIA demonstrated better quantification robustness, data completeness, and sensitivity than DDA (5Barkovits K. Pacharra S. Pfeiffer K. Steinbach S. Eisenacher M. Marcus K. et al.Reproducibility, specificity and accuracy of relative quantification using spectral library-based data-independent acquisition.Mol. Cell Proteomics. 2020; 19: 181-197Abstract Full Text Full Text PDF PubMed Scopus (56) Google Scholar). Among DIA methodologies, Sequential Window Acquisition of All Theoretical Fragment Ion Mass Spectra (SWATH-MS) is a widely used methodology that allows deep proteome coverage with quantitative consistency and accuracy (6Ludwig C. Gillet L. Rosenberger G. Amon S. Collins B.C. Aebersold R. Data-independent acquisition-based SWATH-MS for quantitative proteomics: a tutorial.Mol. Syst. Biol. 2018; 14e8126Crossref PubMed Scopus (486) Google Scholar). A recent multilaboratory evaluation study demonstrated its reproducibility as well as quantitative and qualitative performances (7Collins B.C. Hunter C.L. Liu Y. Schilling B. Rosenberger G. Bader S.L. et al.Multi-laboratory assessment of reproducibility, qualitative and quantitative performance of SWATH-mass spectrometry.Nat. Commun. 2017; 8: 291Crossref PubMed Scopus (295) Google Scholar). To date, SWATH-MS methods have been primarily operated at nanoflow (7Collins B.C. Hunter C.L. Liu Y. Schilling B. Rosenberger G. Bader S.L. et al.Multi-laboratory assessment of reproducibility, qualitative and quantitative performance of SWATH-mass spectrometry.Nat. Commun. 2017; 8: 291Crossref PubMed Scopus (295) Google Scholar) and microflow regimes (2Goncalves E. Poulos R.C. Cai Z. Barthorpe S. Manda S.S. Lucas N. et al.Pan-cancer proteomic map of 949 human cell lines.Cancer Cell. 2022; 40: 835-849.e8Abstract Full Text Full Text PDF PubMed Scopus (17) Google Scholar) providing excellent proteome depth and data quality. However, these methods often lack throughput and chromatographic robustness necessary for the analysis of large studies. In this context, we developed, optimized, and benchmarked robust and high-throughput analytical flow LC-MS methods with Zeno SWATH DIA approach. Taking advantage of the Zeno trap, Zeno SWATH methods consistently provided higher sensitivity and protein coverage as well as superior data quality across a wide variety of biological matrices in comparison to SWATH. Furthermore, Zeno SWATH DIA improved pathway analysis and enhanced the ability to identify plasma biomarkers for two metabolic diseases in humans, demonstrating the clinical application of this novel method. This work was conducted according to the Declaration of Helsinki ethical principles. The following experiments were performed: (i) triplicate injections of 10-point dilution series from 10 cynomolgus monkey tissues, three human biofluids and two in vitro samples from air-liquid interface cultures, (ii) singlicate injection of clinical plasma samples from 20 healthy individuals and 20 individuals diagnosed with nonalcoholic steatohepatitis (NASH), and 10 individuals diagnosed with type 2 diabetes (T2D), and (iii) repetitive singlicate injection of commercial HeLa cell protein digest along the sequence (n = 17). All experiments were conducted in SWATH DIA mode with and without Zeno trap activated on the same instrument, and HeLa experiments were performed with Zeno trap activated. All data were acquired in a single uninterrupted sequence and are available via ProteomeXchange with identifier PXD039414. Dilution series were acquired to assess the performance of SWATH DIA experiments with Zeno trap activated (i.e., sensitivity, coverage, signal linearity) and technical replicates were used to evaluate quantitative robustness by calculating coefficients of variation (CVs). Dilution series were run from low to high loads to potential plasma samples were run in a injections of HeLa were used to evaluate over time and method processing was performed with on the samples data and processing LC-MS and were from HeLa protein digest and were from monkey samples including and were from healthy cynomolgus The the and of were and by and and the and of were conducted following the of the were conducted using a single from a single were from 10 healthy individuals using were with and at for 10 at The were and at A of all the samples was used in this was by The and were from all human samples from 20 individuals as healthy were from and samples from 20 individuals diagnosed with and 10 individuals diagnosed with were from and are available in samples were following with which in with the in to and to of human and were by of all biological samples using a for more A of all healthy plasma samples was used for the dilution series experiment. human samples from two healthy were from and for and samples were from in vitro interface cell and from experiments performed with from three to human human were from and were and at as A. of human of in of pathways in J. Cell PubMed Scopus Google P. E. of human cell for in vitro PubMed Scopus Google Scholar) with were in type and interface with were by of and of with with and of the the was using along with samples. All the samples were at to the of tissues were to 2 and of of in with were with at on of plasma was added to of and and samples were to a of and in of were on for at with a and for 10 at at The were and protein were using following the and of proteins were performed of the at for 10 by of and of by at for was performed using sample processing in μg following the proteins were by of by of to protein μg of were and on the following at for 2 at were three with and at for 2 was performed at with in of peptides was performed with of of in and of in by at for 2 were and a of and peptides were in of in were using and samples were at LC-MS data were using with a mass were on and in in and in were from the using a 10-min gradient at a flow of and the was at was at 10 The was with and operated in SWATH mode with or without Zeno trap activated. The were as The acquisition were as number of MS1 MS1 m/z to and or 15 of the m/z to window were for sample type from DDA using SWATH Window provided by All tissues were run with the same method on was performed three to using the and for dilution series were the to the at and from low to high in in SWATH DIA and Zeno SWATH DIA and plasma samples were in singlicate in a in SWATH mode by Zeno SWATH. All samples were more than days in the over time and method robustness were by repetitive injection of of HeLa protein digest in Zeno SWATH HeLa protein digest sample was in the and the were the of the All data were with with a m/z range of to of of mass accuracy at the and MS1 and protein of and quantification of dilution triplicate samples from dilution and acquisition mode were with and with plasma samples from study and acquisition mode were with and with All were performed using in from sequence protein for human biofluids and in vitro samples and sequence protein for cynomolgus monkey were using the following peptide to to Mass for and was in on the run in the experiment. on of and data to spectral in the V. C.B. M. and deep proteome coverage in high 2020; PubMed Scopus Google Scholar). HeLa were using from HeLa and cell provided by The mass spectrometry proteomics data have been to the ProteomeXchange via the Y. A. J. M. S. et and and in improving for quantification PubMed Scopus Google Scholar) with the identifier PXD039414. were performed in and all to dilution of and were used for and protein and of clinical plasma samples was performed using of for and protein and in samples was by calculating of across using protein in at of the samples The was using in and the used was analysis in tissues was performed using in and for was used for the were performed using the proteins in and all the proteins in the were used as with were The number of protein pathway was in pathways with and without Zeno trap activated were across triplicate and the of and protein with a were relative to proteins in of the and analysis was performed using a coefficients of of peptide loads were in all matrices and in at three in SWATH and Zeno SWATH were in the A was used for The ability to across loads was on with a across triplicate injections and analysis was performed using a of variation in HeLa protein were using protein and in at of the samples of plasma samples from healthy and individuals was performed as (i) protein in than of the plasma samples (ii) and (iii) analysis using a The were using the and proteins with were analysis was performed using and pathways with were The Zeno trap on is a trap the cell that of by up to A novel trap that high and wide m/z range on injection mass Mass PubMed Scopus Google T. P. E. K. A. B. et of by in a high sensitivity mass Mass 2021; PubMed Scopus Google Z. M. A. K. T. et proteomics of samples with Zeno SWATH 2022; Scopus Google Scholar). This increased sensitivity the of analytical flow methods in comparison to more conventional microflow or nanoflow To assess the in with Zeno trap we protein from a variety of biological matrices including 10 cynomolgus monkey tissues and three human biofluids and and two in vitro samples from interface and was performed at analytical flow with a 10-min gradient chromatography to throughput and robustness, and LC-MS were to identify the most protein and in biological The flow MS1 and time as well as SWATH window number and window and and Notably, the selected were for sample type protein and and protein coverage were on 10-point dilution series across all biological matrices to using LC-MS methods with and without Zeno trap activated and data were using in in V. C.B. M. and deep proteome coverage in high 2020; PubMed Scopus Google Scholar). Zeno SWATH DIA methods more and protein at all peptide loads across all sample types in comparison to conventional SWATH DIA methods run on the same and enabling Zeno trap increased the number of by 20 to and protein by 20 to Among the sample plasma the from the in sensitivity, and the Zeno SWATH methods to to and to in tissues, and in vitro to to to and to protein While Zeno trap improved protein coverage at high the in sensitivity was the most at low peptide loads and with in protein and coverage by to across the matrices and up to for and and and in and the most from the in sensitivity to in protein a number of proteins were with SWATH DIA and the in sensitivity provided by the Zeno trap identification of Finally, SWATH methods were to at low peptide loads and Zeno SWATH methods typically of and protein at these demonstrating the of this sample is and we protein of cynomolgus monkey tissues using to assess data quality. samples from the and and tissues and in protein These in human tissues M. L. C. P. A. et map of the human PubMed Scopus Google Scholar). we the in protein coverage with Zeno trap activated pathway analysis was performed using M. D. H. et for the of The PubMed Scopus Google C. the a 2021; PubMed Scopus Google Scholar) and a in pathway coverage across and peptide loads with gains at loads and in and pathways by and at peptide in pathway coverage increased by and on across in biofluids and in vitro at peptide load. While the in pathway coverage was tissues and in vitro from were more In pathway coverage in plasma was in comparison to biofluids on across at peptide the in the number of proteins In addition, the number of proteins pathway was in pathways in SWATH and Zeno SWATH a in the number of proteins identified pathway with and proteins on across in tissues, and in vitro at peptide and Altogether, Zeno SWATH DIA provided more by more pathways and improving pathway data quality. In to proteome coverage and sensitivity, the quantitative quality of the data in Zeno SWATH mode was dilution series and technical at by the of with a across triplicate improved in tissues at all peptide loads and of were with a in Zeno SWATH and SWATH at peptide load. protein and of proteins were with a in Zeno SWATH and SWATH at peptide were in biofluids and in vitro samples linearity was by calculating coefficients of of peptide While high coefficients of were in in Zeno SWATH improved linearity in all matrices and the ability of Zeno SWATH and SWATH methods to across loads by on the that robust quantification of across triplicate a clear in the number of reliable was at loads with Zeno SWATH DIA the in sensitivity and demonstrated that the of analytical flow chromatography with a 10-min gradient and Zeno SWATH DIA enhanced sensitivity, protein coverage, and higher quality However, for this method to well to large-scale method over time to To evaluate method HeLa (n = were along the study which in uninterrupted injections of 1000+ samples to μg peptides in that a in the number of and protein demonstrating method robustness In addition, signal stable without the need for with coefficients of variation across all HeLa digest injections of and for and protein Altogether, we have demonstrated robust and consistent performance of high-throughput LC-MS method over the injection of 1000+ samples without the need for human intervention or and data normalization. Finally, we the ability of Zeno SWATH to identify biomarkers and of disease In this context, we protein in plasma from diagnosed with and with and without Zeno trap activated. in analysis of samples from healthy individuals identified and dysregulated proteins in SWATH and Zeno SWATH and most proteins in SWATH mode were with Zeno SWATH and proteins that have been associated with S. J. Liu J. H. Y. et is that and 2021; Scopus Google N. E. K. M. H. et of are associated with with nonalcoholic from to human 19: PubMed Scopus Google and G. L. H. S. M. D. et and are associated with of disease at to a 2021; 11: PubMed Scopus Google as well as protein Liu T. S. T. et of and in Full Text Full Text PDF PubMed Scopus Google Zeno SWATH identified dysregulated proteins which were and were in SWATH mode of the proteins that were but in SWATH that Zeno SWATH data were patient with which translated into of these proteins were in the of and G. V. V. D. A. G. et in nonalcoholic PubMed Scopus Google Scholar) and and D. A. M. S. L. and associated with and 2022; PubMed Scopus Google Scholar). analysis identified pathways associated with and as and Y. Y. a of and PubMed Scopus Google Scholar) and H. of in nonalcoholic the PubMed Scopus Google Scholar) as well as D. A. M. S. L. and associated with and 2022; PubMed Scopus Google Scholar) (i.e., While these pathways were in SWATH and Zeno SWATH was enhanced with Zeno trap activated In the of improved = in SWATH mode and = in Zeno SWATH that proteins in Zeno SWATH mode to this and proteins were identified in plasma from with healthy individuals in SWATH and Zeno SWATH and Among protein has been associated with F. N. Y. Z. G. et as a of via Commun. 2021; PubMed Scopus Google is a that in Y. as in to its and 2020; PubMed Scopus Google and G. P. F. S. M. C. et and but disease are associated with in from the study of PubMed Google Scholar). Zeno SWATH identified proteins in the of as potential biomarker of S. of in of in diabetes J. 2017; PubMed Scopus Google Scholar) and of metabolic and diabetes Y. T. T. of on the of diabetes in a 2018; PubMed Scopus Google Scholar). Among proteins with Zeno were and were in SWATH mode and the pathway were with and without Zeno trap activated pathway improved for a of including = in SWATH mode and = in Zeno SWATH and in and J. PubMed Scopus Google Scholar). Altogether, Zeno trap in the identification of ∼80% more proteins in plasma from diagnosed with metabolic diseases and these proteins and associated pathways were to disease In this we and analytical flow LC-MS methods with Zeno SWATH DIA across a range of biological matrices and the performance relative to SWATH DIA the Zeno trap improved all the performance coverage in proteins and data quality, and ability to identify plasma biomarkers in clinical samples. MS-based proteomics is to and (6Ludwig C. Gillet L. Rosenberger G. Amon S. Collins B.C. Aebersold R. Data-independent acquisition-based SWATH-MS for quantitative proteomics: a tutorial.Mol. Syst. Biol. 2018; 14e8126Crossref PubMed Scopus (486) Google V. C.B. M. and deep proteome coverage in high 2020; PubMed Scopus Google F. S. J. Mann M. acquisition method proteomics at a depth of proteins in 2018; PubMed Scopus Google F. M. A. E. et fragmentation with data-independent 2020; PubMed Scopus Google F. Mann M. spectrometry and fragmentation in Cell Proteomics. 2021; Full Text Full Text PDF PubMed Scopus Google K. E. M. D. L. N. et of analysis for data-independent acquisition proteomics using a large-scale Commun. 2022; PubMed Scopus Google Scholar). and acquisition methods are more and processing more the proteomics is and higher flow to application of the methodology to studies requiring large sample a gradient method at to plasma proteins C.B. Demichev V. N. White M. M. et proteomics with 2021; PubMed Scopus Google Scholar). While a method is for plasma analysis to its in protein protein coverage in sample types to The in sensitivity provided by the Zeno trap improved protein and coverage by up to at 2 μg peptide with a 10-min gradient analytical flow These methods up to 3300 proteins from tissues which is to performances at microflow using R. Hunter C. C. N. S. et protein biomarker discovery from samples using gradient microflow SWATH 2020; 19: PubMed Scopus Google Scholar). The in protein coverage was the most at low peptide loads with up to the of the sample is technologies more they are used in studies high quality for disease and as well as disease on and proteomics are the two most and for a wide range of including and diseases T. J. G. M. D. D. et 2022; PubMed Scopus Google F. E. D. J. S. J. of metabolic biomarkers for type 2 diabetes in the 2022; PubMed Scopus Google Z. C. V. et Biol. PubMed Scopus Google R. J. K. et al.A proteomic for that is to of in 2022; Scopus Google Scholar). proteomics has been in of human disease from has improved sensitivity and specificity of in studies of diseases (i.e., type 2 K. T. S. et al.A data for precision PubMed Scopus Google Scholar) and (i.e., and K. S. D. S. et of 2020; Full Text Full Text PDF PubMed Scopus Google K. D. et of the and 2021; PubMed Scopus Google Scholar). In addition, proteomics has the potential to studies with the discovery of safety and efficacy biomarkers as well as the of of drug and drug large-scale plasma proteomics by which protein that was to and C. A. R. et assessment of drug safety proteomics: and of the of 2018; PubMed Scopus Google Scholar). The were to identify a that in from of In to the of proteomics technologies in clinical is a clear need to high throughput and robust methods that the proteome the need for flow chromatographic robustness and signal over time as demonstrated by the in sensitivity the uninterrupted injection of 1000+ samples and high reproducibility of quality samples the methods have been to reproducibility over they are typically and in variation R.C. R. Lucas N. D. Manda S.S. et to large-scale proteomics for Commun. 2020; 11: PubMed Scopus Google J. T. Collins B.C. M. et and of in large-scale proteomic a tutorial.Mol. Syst. Biol. 2021; PubMed Scopus Google Scholar). In study reveals that SWATH DIA with Zeno trap sensitivity, quantitative robustness, and signal linearity and protein coverage in all sample types The in protein coverage translated into better biological pathway representation and identification of potential disease biomarkers and dysregulated pathways in a In addition, the methods using analytical flow chromatography were robust the need for human intervention in large-scale studies. Zeno SWATH DIA is well for applications that a phenotypic at and drug and clinical biomarker discovery studies. The mass spectrometry proteomics data have been to the ProteomeXchange via the Y. A. J. M. S. et and and in improving for quantification PubMed Scopus Google Scholar) with the identifier PXD039414. All data are available from the on This Y. J. S. A. and K. C. are of and have in the Y. L. was a for to and from for providing technical and on the and Demichev of for with are to Liu and from for and commercial plasma and samples. K. Y. Y. and A. R. Y. and K. C. Y. L. and S. Y. and K. C. S. and K. C. data K. and Y. L. K. and Y. L. K. C. Y. J. and S. T. K. C. and A. R. K. and A. R. K. and A. R. K. A. Y. Y. J. and S. T. with with with with with with with data

Topics & Concepts