Preclinical Characterization of Relatlimab, a Human LAG-3-Blocking Antibody, Alone or in Combination with Nivolumab.

Novel therapeutic approaches combining immune-checkpoint inhibitors are needed to improve clinical outcomes for patients with cancer. Lymphocyte-activation gene 3 (LAG-3) is an immune-checkpoint molecule that inhibits T-cell activity and antitumor immune responses, acting through an independent mechanism from that of programmed death-1 (PD-1) and cytotoxic T-lymphocyte-associated antigen-4 (CTLA-4). Here, we describe the development and preclinical characterization of relatlimab, a human antibody that binds to human LAG-3 with high affinity and specificity to block the interaction of LAG-3 with the ligands MHC II and fibrinogen-like protein-1, and to reverse LAG-3-mediated inhibition of T-cell function in vitro. Consistent with previous reports, in mouse models, the combined blockade of LAG-3 and PD-1 with surrogate antibodies resulted in enhanced antitumor activity greater than the individual blockade of either receptor. In toxicity studies in cynomolgus monkeys, relatlimab was generally well tolerated when combined with nivolumab. These results are consistent with findings from the RELATIVITY-047 phase II/III trial showing that relatlimab combined with nivolumab is a well-tolerated regimen that demonstrates superior progression-free survival compared with nivolumab monotherapy in patients with unresectable or metastatic melanoma.

High-Throughput Platform to Identify Antibody Conjugation Sites from Antibody-Drug Conjugate Libraries.

Antibody-drug conjugates (ADCs) are a therapeutic modality that traditionally enable the targeted delivery of highly potent cytotoxic agents to specific cells such as tumor cells. More recently, antibodies have been used to deliver molecules such as antibiotics, antigens, and adjuvants to bacteria or specific immune cell subsets. Site-directed mutagenesis of proteins permits more precise control over the site and stoichiometry of their conjugation, giving rise to homogeneous chemically defined ADCs. Identification of favorable sites for conjugation in antibodies is essential as reaction efficiency and product stability are influenced by the tertiary structure of immunoglobulin G (IgG). Current methods to evaluate potential conjugation sites are time-consuming and labor intensive, involving multistep processes for individually produced reactions. Here, we describe a highly efficient method for identification of conjugatable genetic variants by analyzing pooled ADC libraries using mass spectrometry. This approach provides a versatile platform to rapidly uncover new conjugation sites for site-specific ADCs.

Design and characterization of mouse IgG1 and IgG2a bispecific antibodies for use in syngeneic models.

The development of antibody therapeutics relies on animal models that accurately recapitulate disease biology. Syngeneic mouse models are increasingly used with new molecules to capture the biology of complex cancers and disease states, and to provide insight into the role of the immune system. The establishment of syngeneic mouse models requires the ability to generate surrogate mouse counterparts to antibodies designed for humans. In the field of bispecific antibodies, there remains a dearth of technologies available to generate native IgG-like mouse bispecific antibodies. Thus, we engineered a simple co-expression system for one-step purification of intact mouse IgG1 and IgG2a bispecific antibodies from any antibody pair. We demonstrated proof of concept with CD3/CD20 bispecific antibodies, which highlighted both the quality and efficacy of materials generated by this technology.

A targeted proteomics-based pipeline for verification of biomarkers in plasma.

High-throughput technologies can now identify hundreds of candidate protein biomarkers for any disease with relative ease. However, because there are no assays for the majority of proteins and de novo immunoassay development is prohibitively expensive, few candidate biomarkers are tested in clinical studies. We tested whether the analytical performance of a biomarker identification pipeline based on targeted mass spectrometry would be sufficient for data-dependent prioritization of candidate biomarkers, de novo development of assays and multiplexed biomarker verification. We used a data-dependent triage process to prioritize a subset of putative plasma biomarkers from >1,000 candidates previously identified using a mouse model of breast cancer. Eighty-eight novel quantitative assays based on selected reaction monitoring mass spectrometry were developed, multiplexed and evaluated in 80 plasma samples. Thirty-six proteins were verified as being elevated in the plasma of tumor-bearing animals. The analytical performance of this pipeline suggests that it should support the use of an analogous approach with human samples.

A proteomics platform combining depletion, multi-lectin affinity chromatography (M-LAC), and isoelectric focusing to study the breast cancer proteome.

The discovery of breast cancer associated plasma/serum biomarkers is important for early diagnosis, disease mechanism elucidation, and determination of treatment strategy for the disease. In this study of serum samples, a multidimensional fractionation platform combined with mass spectrometric analysis were used to achieve the identification of medium to lower abundance proteins, as well as to simultaneously detect glycan and abundance changes. Immuno-affinity depletion and multi-lectin chromatography (M-LAC) were integrated into an automated HPLC platform to remove high abundance protein and fractionate glycoproteins. The collected glycoproteomes were then subjected to isoelectric focusing (IEF) separation by a digital ProteomeChip (dPC), followed by in-gel digestion and LC-MS analysis using an Orbitrap mass spectrometer. As a result, the total number of identified proteins increased significantly when the IEF fractionation step was included as part of the platform. Relevant proteins with biological and disease significance were observed and the dynamic range of the serum proteome measurement was extended. In addition, potential glycan changes were indicated by comparing proteins in control and cancer samples in terms of their affinity to the multi-lectin column (M-LAC) and the pI profiles in IEF separation. In conclusion, a proteomics platform including high abundance protein depletion, lectin affinity fractionation, IEF separation, and LC-MS analysis has been applied to discover breast cancer-associated proteins. The following candidates, thrombospondin-1 and 5, alpha-1B-glycoprotein, serum amyloid P-component, and tenascin-X, were selected as promising examples of the use of this platform. They show potential abundance and glycan changes and will be further investigated in future studies.

The development of an integrated platform to identify breast cancer glycoproteome changes in human serum.

Protein glycosylation represents one of the major post-translational modifications and can have significant effects on protein function. Moreover, changes in the carbohydrate structure are increasingly being recognized as an important modification associated with cancer etiology. In this report, we describe the development of a proteomics approach to identify breast cancer related changes in either concentration and/or the carbohydrate structures of glycoprotein(s) present in blood samples. Diseased and healthy serum samples were processed by an optimized sample preparation protocol using multiple lectin affinity chromatography (M-LAC) that partitions serum proteins based on glycan characteristics. Subsequently, three separate procedures, 1D SDS-PAGE, isoelectric focusing and an antibody microarray, were applied to identify potential candidate markers for future study. The combination of these three platforms is illustrated in this report with the analysis of control and cancer glycoproteomic fractions. Firstly, a molecular weight based separation of glycoproteins by 1D SDS-PAGE was performed, followed by protein, glycoprotein staining, lectin blotting and LC-MS analysis. To refine or confirm the list of interesting glycoproteins, isoelectric focusing (targeting sialic acid changes) and an antibody microarray (used to detect neutral glycan shifts) were selected as the orthogonal methods. As a result, several glycoproteins including alpha-1B-glycoprotein, complement C3, alpha-1-antitrypsin and transferrin were identified as potential candidates for further study.

Experiment-specific estimation of peptide identification probabilities using a randomized database.

Determining the error rate for peptide and protein identification accurately and reliably is necessary to enable evaluation and crosscomparisons of high throughput proteomics experiments. Currently, peptide identification is based either on preset scoring thresholds or on probabilistic models trained on datasets that are often dissimilar to experimental results. The false discovery rates (FDR) and peptide identification probabilities for these preset thresholds or models often vary greatly across different experimental treatments, organisms, or instruments used in specific experiments. To overcome these difficulties, randomized databases have been used to estimate the FDR. However, the cumulative FDR may include low probability identifications when there are a large number of peptide identifications and exclude high probability identifications when there are few. To overcome this logical inconsistency, this study expands the use of randomized databases to generate experiment-specific estimates of peptide identification probabilities. These experiment-specific probabilities are generated by logistic and Loess regression models of the peptide scores obtained from original and reshuffled database matches. These experiment-specific probabilities are shown to very well approximate "true" probabilities based on known standard protein mixtures across different experiments. Probabilities generated by the earlier Peptide_Prophet and more recent LIPS models are shown to differ significantly from this study's experiment-specific probabilities, especially for unknown samples. The experiment-specific probabilities reliably estimate the accuracy of peptide identifications and overcome potential logical inconsistencies of the cumulative FDR. This estimation method is demonstrated using a Sequest database search, LIPS model, and a reshuffled database. However, this approach is generally applicable to any search algorithm, peptide scoring, and statistical model when using a randomized database.

Charge state estimation for tandem mass spectrometry proteomics.

High-throughput protein analysis by tandem mass spectrometry produces anywhere from thousands to millions of spectra that are being used for peptide and protein identifications. Though each spectrum corresponds only to one charged peptide (ion) state, repetitive database searches of multiple charge states are typically conducted since the resolution of many common mass spectrometers is not sufficient to determine the charge state. The resulting database searches are both error-prone and time-consuming. We describe a straightforward, accurate approach on charge state estimation (CHASTE). CHASTE relies on fragment ion peak distributions, and by using reliable logistic regression models, combines different measurements to improve its accuracy. CHASTE's performance has been validated on data sets, comprised of known peptide dissociation spectra, obtained by replicate analyses of our earlier developed protein standard mixture using ion trap mass spectrometers at different laboratories. CHASTE was able to reduce number of needed database searches by at least 60% and the number of redundant searches by at least 90% virtually without any informational loss. This greatly alleviates one of the major bottlenecks in high throughput peptide and protein identifications. Thresholds and parameter estimates can be tailored to specific analysis situations, pipelines, and instrumentations. CHASTE was implemented in Java GUI-based and command-line-based interfaces.

Electron transfer versus proton transfer in gas-phase ion/ion reactions of polyprotonated peptides.

The ion/ion reactions of several dozen reagent anions with triply protonated cations of the model peptide KGAILKGAILR have been examined to evaluate predictions of a Landau-Zener-based model for the likelihood for electron transfer. Evidence for electron transfer was provided by the appearance of fragment ions unique to electron transfer or electron capture dissociation. Proton transfer and electron transfer are competitive processes for any combination of anionic and cationic reactants. For reagent anions in reactions with protonated peptides, proton transfer is usually significantly more exothermic than electron transfer. If charge transfer occurs at relatively long distances, electron transfer should, therefore, be favored on kinetic grounds because the reactant and product channels cross at greater distances, provided conditions are favorable for electron transfer at the crossing point. The results are consistent with a model based on Landau-Zener theory that indicates both thermodynamic and geometric criteria apply for electron transfer involving polyatomic anions. Both the model and the data suggest that electron affinities associated with the anionic reagents greater than about 60-70 kcal/mol minimize the likelihood that electron transfer will be observed. Provided the electron affinity is not too high, the Franck-Condon factors associated with the anion and its corresponding neutral must not be too low. When one or the other of these criteria is not met, proton transfer tends to occur essentially exclusively. Experiments involving ion/ion attachment products also suggest that a significant barrier exists to the isomerization between chemical complexes that, if formed, lead to either proton transfer or electron transfer.

SO2-* electron transfer ion/ion reactions with disulfide linked polypeptide ions.

Multiply-charged peptide cations comprised of two polypeptide chains (designated A and B) bound via a disulfide linkage have been reacted with SO2-* in an electrodynamic ion trap mass spectrometer. These reactions proceed through both proton transfer (without dissociation) and electron transfer (with and without dissociation). Electron transfer reactions are shown to give rise to cleavage along the peptide backbone, loss of neutral molecules, and cleavage of the cystine bond. Disulfide bond cleavage is the preferred dissociation channel and both Chain A (or B)-S* and Chain A (or B)-SH fragment ions are observed, similar to those observed with electron capture dissociation (ECD) of disulfide-bound peptides. Electron transfer without dissociation produces [M + 2H]+* ions, which appear to be less kinetically stable than the proton transfer [M + H]+ product. When subjected to collision-induced dissociation (CID), the [M + 2H]+* ions fragment to give products that were also observed as dissociation products during the electron transfer reaction. However, not all dissociation channels noted in the electron transfer reaction were observed in the CID of the [M + 2H]+* ions. The charge state of the peptide has a significant effect on both the extent of electron transfer dissociation observed and the variety of dissociation products, with higher charge states giving more of each.

Electron transfer ion/ion reactions in a three-dimensional quadrupole ion trap: reactions of doubly and triply protonated peptides with SO2*-.

Ion-ion reactions between a variety of peptide cations (doubly and triply charged) and SO2 anions have been studied in a 3-D quadrupole ion trap, resulting in proton and electron transfer. Electron transfer dissociation (ETD) gives many c- and z-type fragments, resulting in extensive sequence coverage in the case of triply protonated peptides with SO2*-. For triply charged neurotensin, in which a direct comparison can be made between 3-D and linear ion trap results, abundances of ETD fragments relative to one another appear to be similar. Reactions of doubly protonated peptides with SO2*- give much less structural information from ETD than triply protonated peptides. Collision-induced dissociation (CID) of singly charged ions formed in reactions with SO2*- shows a combination of proton and electron transfer products. CID of the singly charged species gives more structural information than ETD of the doubly protonated peptide, but not as much information as ETD of the triply protonated peptide.

Targeted biomarker detection via whole protein ion trap tandem mass spectrometry: thymosin beta4 in a human lung cancer cell line.

N-Terminally acetylated thymosin beta4, a species implicated for use as a cancer biomarker, was identified in a human lung cancer cell line using ion trap tandem mass spectrometry at the whole protein level. Ion-ion proton transfer reactions were used for parent ion concentration/manipulation and to simplify interpretation of product ion spectra. Dissociation data for the +6 to +3 charge states are reported. As is usually the case, structural information available from the ion trap collisional activation of the protein is sensitive to parent ion charge state. Each parent ion charge state selected, however, provided sufficient information to make a confident identification. Furthermore, each charge state provided relatively rich fragmentation. Therefore, any of the charge states can be used to detect with high specificity thymosin beta(4) in a complex protein mixture. There are advantages associated with the rapid detection of protein biomarkers at the whole protein level, as opposed to the peptide level following protein digestion, particularly for relatively small protein and polypeptide biomarkers. Having identified and characterized the protein, product ion spectra obtained directly, without recourse to ion-ion proton transfer reactions, can be used for library matching. However, ion-ion proton transfer reactions for parent ion concentration and charge state purification are advantageous in addressing relatively complex mixtures.

Randomized sequence databases for tandem mass spectrometry peptide and protein identification.

Tandem mass spectrometry (MS/MS) combined with database searching is currently the most widely used method for high-throughput peptide and protein identification. Many different algorithms, scoring criteria, and statistical models have been used to identify peptides and proteins in complex biological samples, and many studies, including our own, describe the accuracy of these identifications, using at best generic terms such as "high confidence." False positive identification rates for these criteria can vary substantially with changing organisms under study, growth conditions, sequence databases, experimental protocols, and instrumentation; therefore, study-specific methods are needed to estimate the accuracy (false positive rates) of these peptide and protein identifications. We present and evaluate methods for estimating false positive identification rates based on searches of randomized databases (reversed and reshuffled). We examine the use of separate searches of a forward then a randomized database and combined searches of a randomized database appended to a forward sequence database. Estimated error rates from randomized database searches are first compared against actual error rates from MS/MS runs of known protein standards. These methods are then applied to biological samples of the model microorganism Shewanella oneidensis strain MR-1. Based on the results obtained in this study, we recommend the use of use of combined searches of a reshuffled database appended to a forward sequence database as a means providing quantitative estimates of false positive identification rates of peptides and proteins. This will allow researchers to set criteria and thresholds to achieve a desired error rate and provide the scientific community with direct and quantifiable measures of peptide and protein identification accuracy as opposed to vague assessments such as "high confidence."

Phosphorylation site identification via ion trap tandem mass spectrometry of whole protein and peptide ions: bovine alpha-crystallin A chain.

Tandem mass spectrometry was applied both to ions of a tryptic fragment and intact protein of bovine alpha-crystallin A chain to localize the single site of phosphorylation. The [M + 19H](19+) to [M + 11H](11+) charge states of both phosphorylated and unphosphorylated bovine alpha-crystallin A chain whole protein ions were subjected to collisional activation in a quadrupole ion trap. Ion parking was used to increase the number of parent ions over that yielded by electrospray. Ion-ion proton-transfer reactions were used to reduce the product ion charge states largely to +1 to simplify spectral interpretation. In agreement with previous studies on whole protein ion fragmentation, both protein forms showed backbone cleavages C-terminal to aspartic acid residues at lower charge states. The phosphorylated protein showed competitive fragmentation between backbone cleavage and the neutral loss of phosphoric acid. Analysis of which backbone cleavage products did or did not contain the phosphate was used to localize the site of phosphorylation to one of two possible serine residues. A tryptic digest of the bovine alpha-crystallin A chain yielded a phosphopeptide containing one missed cleavage site. The peptide provided information complementary to that obtained from the intact protein and localized the modified serine to residue 122. Fragmentation of the triply charged phosphopeptide yielded five possible serine phosphorylation sites. Fragmentation of the doubly charged phosphopeptide, formed by ion/ion proton-transfer reactions, positively identified the phosphorylation site as serine-122.

Gas-phase concentration, purification, and identification of whole proteins from complex mixtures.

Five proteins present in a relatively complex mixture derived from a whole cell lysate fraction of E. coli have been concentrated, purified, and dissociated in the gas phase, using a quadrupole ion trap mass spectrometer. Concentration of intact protein ions was effected using gas-phase ion/ion proton-transfer reactions in conjunction with mass-to-charge dependent ion "parking" to accumulate protein ions initially dispersed over a range of charge states into a single lower charge state. Sequential ion isolation events interspersed with additional ion parking ion/ion reaction periods were used to "charge-state purify" the protein ion of interest. Five of the most abundant protein components present in the mixture were subjected to this concentration/purification procedure and then dissociated by collisional activation of their intact multiply charged precursor ions. Four of the five proteins were subsequently identified by matching the uninterpreted product ion spectra against a partially annotated protein sequence database, coupled with a novel scoring scheme weighted for the relative abundances of the experimentally observed product ions and the frequency of fragmentations occurring at preferential cleavage sites. The identification of these proteins illustrates the potential of this "top-down" protein identification approach to reduce the reliance on condensed-phase chemistries and extensive separations for complex protein mixture analysis.