Prostate‐specific membrane antigen (PSMA) glycoforms in prostate cancer patients seminal plasma - Mackay - The Prostate - Wiley Online Library
Abstract
Introduction
Prostate-specific membrane antigen (PSMA) is a US Food and Drug Administration-approved theranostic target for prostate cancer (PCa). Although PSMA is known to be glycosylated, the composition and functional roles of its N-linked glycoforms have not been fully characterized.
Methods
PSMA was isolated from pooled seminal plasma from low-risk grade Groups 1 and 2 PCa patients. Intact glycopeptides were analyzed by mass spectrometry to identify site-specific glycoforms.
Results
We observed a rich distribution of PSMA glycoforms in seminal plasma from low and low-intermediate-risk PCa patients. Some interesting generalities can be drawn based on the predicted topology of PSMA on the plasma membrane. The glycoforms at ASN-459, ASN-476, and ASN-638 residues that are located at the basal domain facing the plasma membrane in cells, are predominantly high mannose glycans. ASN-76 which is located in the interdomain region adjacent to the apical domain of the protein shows a mixture of high mannose glycans and complex glycans, whereas ASN-121, ASN-195 and ASN-336 that are located and are exposed at the apical domain of the protein predominantly possess complex sialylated and fucosylated N-linked glycans. These highly accessible glycosites display the greatest diversity in isoforms across the patient samples.
Conclusions
Our study provides novel qualitative insights into PSMA glycoforms that are present in the seminal fluid of PCa patients. The presence of a rich diversity of glycoforms in seminal plasma provides untapped potential for glycoprotein biomarker discovery and as a clinical sample for noninvasive diagnostics of male urological disorders and diseases including PCa. Specifically, our glycomics approach will be critical in uncovering PSMA glycoforms with utility in staging and risk stratification of PCa.
What are PSMA glycoforms and Why Important
Prostate-specific membrane antigen (PSMA) is a cell surface protein that is overexpressed in prostate cancer cells. Glycoforms refer to different forms of a protein that vary in their glycosylation patterns. Glycosylation is the process of covalently attaching carbohydrate molecules to proteins, and it plays a crucial role in the structure and function of many proteins.
In the context of PSMA, glycoforms may refer to different versions of the protein that have distinct patterns of glycosylation. These variations in glycosylation can affect the protein's stability, localization, and interactions with other molecules.
Understanding the specific glycoforms of PSMA is important for several reasons:
-
Diagnostic and Prognostic Applications: The glycosylation pattern of PSMA may have diagnostic and prognostic implications in prostate cancer. Changes in glycosylation patterns are often associated with disease states, and analyzing PSMA glycoforms may provide additional information about the nature and aggressiveness of prostate cancer.
-
Therapeutic Targeting: Glycoforms of PSMA may serve as specific targets for therapeutic interventions. By understanding the variations in glycosylation, researchers can design targeted therapies that specifically interact with certain forms of PSMA, potentially improving the precision and efficacy of treatment.
-
Biomarker Discovery: Different glycoforms of PSMA may serve as potential biomarkers for prostate cancer. Biomarkers are measurable indicators of a biological state or condition, and specific glycoforms of PSMA could be used for diagnostic or monitoring purposes.
-
Drug Development: Knowledge of PSMA glycoforms may influence the development of drugs targeting PSMA. Understanding the specific glycosylation patterns can aid in designing drugs that interact with the protein in a selective and effective manner.
Research in the field of cancer biology, particularly in prostate cancer, is ongoing, and scientists continue to investigate the role of PSMA and its glycoforms in health and disease. Advanced analytical techniques, such as mass spectrometry and glycoproteomics, are often employed to study protein glycosylation patterns and identify specific glycoforms associated with disease states.
1 INTRODUCTION
Prostate cancer (PCa) is the leading nonskin malignancy in men and with significant health-related consequences in the United States and worldwide.1, 2 In 2023, approximately 288,300 new cases of PCa will be diagnosed in the United States and with 34,700 deaths.1 PCa is a heterogeneous disease ranging from indolent asymptomatic cases to very aggressive life-threatening forms. Distant metastasis remains the leading cause of mortality in PCa. Approximately 40% of newly diagnosed PCas are indolent and of no serious health consequences during a man's life. However, the remainder range from intermediate to high risk, and despite definitive local therapy, up to 20% of the cases progress to aggressive metastatic disease.3, 4 No effective cures exist for metastatic PCa and 5-year survival rate for metastatic disease is 30%.1, 5 The most commonly used PCa biomarkers include two glycoproteins, prostate-specific antigen and prostate-specific membrane antigen (PSMA). A critical limitation in the management of PCa is the availability of biomarkers with utility in the discrimination of aggressive versus indolent disease accurately. This has resulted in systemic over- and under-treatment of PCa patients. Therefore, it is imperative that novel molecular biomarkers of PCa are still required to facilitate the identification and management of the disease through screening, early detection/diagnosis, and more accurate prognosis and risk stratification.6, 7
PSMA, also known as glutamate carboxypeptidase 2, N-acetylated-alpha-linked acidic dipeptidase I (NAALADase I) or folate hydrolase 1 (FOLH1) is a glycoprotein diagnostic biomarker and therapeutic target in PCa. In addition to prostatic tissue, PSMA is also expressed at lower abundances in other tissues including neo-vasculature of certain solid tumors, healthy tissues including salivary glands, proximal renal tubes, duodenum, and neural ganglia.8 In PCa, PSMA which is a type II transmembrane glycoprotein is expressed on the plasma membrane of the cells and its abundance correlates with aggressive disease progression9, 10 and the protein plays a critical role in PCa cell proliferation.11 Currently, there are multiple PSMA-based applications that are under development and evaluation as PCa diagnostics and therapeutics. Recently, the US Food and Drug Administration-approved Gallium 68 PSMA-11 (Ga 68 PSMA-11) and Pylarify (piflufolastat F 18) as positron emission tomography imaging agents in men with PCa.12, 13
Glycoproteins control numerous critical cellular processes including cell signaling, cell adhesion, protein structure stabilization and trafficking, as well as other processes and activities that promote tumorigenesis. Aberrant protein glycosylation is observed in most cancers including PCa where these changes impact disease progression, severity, and outcome.14-16 Importantly, it has been recently demonstrated that the androgen signaling pathway which is central in PCa initiation and progression is modulated by glycosylation.17-20 PSMA is highly glycosylated with 10 predicted N-linked glycosylation sites at asparagine residues at positions 51, 76, 121, 140, 153, 195, 336, 459, 476, and 638 of its primary amino acid sequence. Its carbohydrate domains contribute up to 30% of its molecular mass.21 These N-linked glycans affect the structure and enzymatic functions of PSMA and it has been shown that PCa cells show differential glycosylation profiles and deglycosylation by glycosidases or by site-directed mutagenesis eliminates its enzymatic activity.22 We and others have performed initial characterization of the composition of N-linked glycans that are conjugated to PSMA isolated from human cells and tissues.23, 24 Although PSMA expression levels correlate with PCa severity, the roles of its conjugated N-linked glycoforms in aggressive disease progression and their potential utility as diagnostic markers, prognostic markers, and therapeutic targets have not been fully characterized.
We have previously used an liquid chromatography tandem mass sp (LC/MS/MS) glycoproteomics approach to characterize PSMA N-linked glycoforms in two human PCa cell lines with different metastatic site localizations. In that study, we uncovered significant qualitative and quantitative changes in PSMA glycoforms between lymph nodes versus bone metastatic cells.23 In this study, we have utilized the same strategy for site-specific N-linked glycan characterization of PSMA glycoforms in seminal plasma from PCa patients stratified into two cohorts based on their International Society of Urological Pathology (ISUP) Gleason Grade scores. Proximal fluids of the prostate like seminal plasma are enriched for prostate-derived proteins compared to blood serum or urine. Potential biomarkers of male reproductive system diseases including PCa are more easily identified and characterized in seminal fluids by proteomics and other techniques. Therefore, seminal plasma has great potential in the discovery and development of protein-based biomarkers and as clinical samples for minimally invasive or noninvasive diagnostics. We hereby present a comprehensive qualitative characterization of the PSMA glycoforms that are observed in seminal plasma from PCa patients with low (ISUP Grade Group 1) to low-intermediate (ISUP Grade Group 2) risk disease.
2 MATERIALS AND METHODS
2.1 Chemicals and reagents
Reagents were obtained from Thermo Fisher Scientific, unless otherwise stated. Sequencing grade modified trypsin (V5111) was purchased from Promega. Anti-human PSMA antibody was purchased from Cell Signaling Technologies Inc. (Catalog #D4S1F; cs-12702) and 3C6 anti-PSMA antibody conjugated sepharose resin that specifically binds to native nondenatured PSMA was a kind gift from C. Berkman and has been previously described.25
2.2 Seminal plasma samples
Deidentified seminal plasma samples were obtained from patients following informed consent under an Institutional Review Board-approved protocol at Eastern Virginia Medical School (IRB # 00-09-EX-0417). Seminal plasma fluids were from men seen at the Eastern Virginia Medical School Department of Urology for PCa screening. Donors were provided with collection kits, and immediately after collection, instant freeze packs were applied to each tube for storage and subsequently transferred to the biorepository. Each sample was thawed and centrifuged to remove sperm before aliquoting and storage. For PSMA glycoform analysis, 32 aliquoted seminal plasma samples (0.25 mL) matched for age, with a mean age of 57.7 years, were selected based on their Pathological Gleason Grade Groups (GG): GG1 ≤ 6 (Gleason Grade 3 + 3) and GG2 = 7 (Gleason Grade 3 + 4), respectively. Protein concentrations were determined in a bicinchoninic acid (BCA) assay before pooling the samples based on normalized protein concentrations. Pooled seminal plasma samples were centrifuged (4°C, 12,000 RPM, 20 min) to remove any cellular contaminants before further analysis. Protein concentrations were determined for the pooled seminal plasma samples using BCA assay and stored at −80°C until further analysis.
2.3 Immunoaffinity enrichment and isolation of PSMA from seminal plasma
To reduce sample complexity and volume and enhance overall PSMA relative abundance, concentration-normalized pooled seminal samples, were subjected to centrifugation at 5000xg using 50 K Centricon filters to a final volume of 250 µL. The retentates comprised of proteins with molecular mass greater than 50 kDa including PSMA were resuspended to a final volume of 2000 µL using radioimmunoprecipitation assay (RIPA) buffer supplemented with protease inhibitors before immunoaffinity enrichment. To reduce nonspecific binding of other proteins to the sepharose beads and anti-PSMA immunoglobulin G (IgG), 2000 µg of seminal plasma proteins were precleared by incubation with 100 µL of Protein A/G Agarose beads (Thermo Scientific) mixed with 10 µg of purified control rabbit IgG for 12 h at 4°C with gentle mixing. Protein A/G and antibody complex including nonspecific binding proteins were pelleted and removed after centrifugation. The supernatants were collected and used for sequential immunoprecipitation. In the first stage, the supernatants were incubated with 200 µL of 3C6 anti-PSMA antibody conjugated to cyanogen bromide sepharose for 12 h at 4°C with gentle mixing. Supernant fractions were collected after centrifugation and used for a second round of immunoprecipitation by incubating with 12 µg anti-PSMA antibody (cs-12702; Cell Signaling Technologies, Inc) for 12 h at 4°C with gentle agitation. Thereafter, each sample was incubated with 200 µL packed Protein A/G agarose beads for another 12 h at 4°C with gentle mixing. Both the cyanogen bromide sepharose and Protein/AG beads were washed three times using RIPA buffer before eluting by heating the beads at 95°C in 400 µL Laemmli sample buffer for 10 min. The Laemmli buffer eluted samples from the consecutive immunoprecipitations were combined to a final volume of 800 µL and fractionated using sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) preparative gels (Bio-Rad) before staining with Colloidal Coomassie to visualize the protein bands. The isolation and purity of enriched PSMA were confirmed using standard SDS-PAGE and Western blot analysis.
2.4 Trypsin digestion
Coomassie-stained PSMA bands were cut across the entire length of the preparative gel and subjected to in-gel trypsin digestion as we have previously described.24, 26 Briefly, excised bands were cut into 1 mm2 cubes and destained in 50 mM ammonium bicarbonate in 50% acetonitrile before dehydration using 100% acetonitrile. In situ reduction and alkylation were achieved by incubating the gel pieces in 10 mM dithiothreitol in 50 mM ammonium bicarbonate at 56°C for 1 h followed by 55 mM iodoacetamide for 45 min at room temperature in darkness. The reduced and alkylated proteins were trypsin digested at 37°C for 18 h. Digestion of PSMA with trypsin generates eight single N-linked glycopeptides and a single peptide that contains two N-linked glycosylation sites. Generated tryptic peptides were extracted as previously described,24-26 then dried in a SpeedVac before glycopeptide enrichment using hydrophilic interaction liquid chromatography (HILIC) TopTips (Glygen). A portion of the nonenriched peptides were also analyzed to confirm the identity and total peptide sequence coverage of all PSMA peptides. For HILIC purification, HILIC tips were activated by washing three times with deionized water before equilibration with binding buffer (15 mM ammonium acetate, pH 3.5 in 85% acetonitrile). Dried peptides were reconstituted in a binding buffer and applied and reloaded on the HILIC tips three times to maximize binding. Nonglycopeptides were removed from the HILIC resin by washing three times using the binding buffer. Bound glycopeptides were eluted with 10 μL of water two times and the combined eluents were dried under vacuum and subsequently reconstituted in normalized volumes of 5% acetonitrile and 0.1% formic acid (FA) (buffer A). The peptide concentration of the two samples was determined using a NanoDrop spectrophotometer and the samples were stored at −80°C before further analysis.
2.5 Glycopeptide LC-MS/MS analysis
PSMA glycopeptides resuspended at a concentration of 0.5 µg/µL in 0.1% FA were analyzed by LC/MS/MS on an Orbitrap Eclipse Tribrid Mass Spectrometer (Thermo Fisher) coupled to an EASY-nLC 1200 liquid chromatography system. The instrument set-up and analytic condition were as we have recently described.24 Briefly, 2 µg of peptides from each sample were injected online into a C18 PepMap precolumn (Cat. # 164946; Thermo Fisher Scientific) using 98% buffer A (2% acetonitrile containing 0.1% FA) connected in line with 500 mm × 75 µm C18 PepMap analytic column (Cat. # 164570; Thermo Fisher Scientific) at flow rate of 5 µL/min for 5 min. This was followed by a 140 min separation using a binary gradient buffer system (buffer A, 0.1% FA, and buffer B 0.1% [v/v] FA in 80% acetonitrile) under the following conditions as we have previously described, flow rate of 250 nL/min, isocratic 8% solvent B over 8 min, 8%−12% solvent B over 5 min; 12%–30% solvent B over 100 min; 30%–60% solvent B over 20 min; 60%–98% solvent B over 5 min; isocratic 98% solvent B over 10 min.24 The glycopeptides samples derived from the two pooled seminal plasma cohorts were analyzed in quadruplicate technical replicate injections generating a sum of four spectral data files for each group. To minimize carryover, each sample injection was proceeded by two column washes in the sequence runs using blank buffer A injections. Intact glycopeptides were analyzed using two mass fragmentation methods higher-energy collision dissociation (HCD) and electron-transfer/higher-energy collision dissociation (EThcD) integrated in the MS scans.27-34 The MS acquisition parameters were as follows: electrospray ionization voltage 2.1 kV and capillary temperature 275°C. The MS instrument was run automatically in an automatic data-dependent acquisition acquisition mode switching between MS1 and MS2 scans. MS1 scans were performed over m/z 375–2000 with the wide quadrupole isolation at a resolution of 120,000 (m/z 200), the RF lens at 30%. MS1 acquisitions were measured in the Orbitrap under the following conditions: resolution, 60,000; automatic gain control (AGC), 4e5 and 75 ms injection time. This was followed by oxonium ion triggered (HexNAc, HexHexNAc, Hex, and NeuAc, with m/z 204.09, 366.13, 163.0601, and 274.09, 657.23, respectively) MS2 EThcD MS/MS scans of glycopeptide precursor ions prioritized by the highest charge states and intensity. The peptides were detected in the Orbitrap under the following conditions: 30,000 resolution, 3e5 AGC, and 200 ms injection time with a scan range (m/z) 120–4000. EThcD scans were acquired under optimized conditions as reported previously.24, 33
2.6 Glycomics LC-MS/MS data analysis
Mass spectrometry spectral files were processed using Xcalibur (Thermo Fisher Scientific) before using Byonic (Protein Metrics) for database searches for the identification of glycans that are conjugated to PSMA glycopeptides. Searches were performed against the UniProt human PSMA protein sequence (Accession # Q04609) together with 132 N-glycans from the Byonic human N-glycan database. The search parameters were (i) protease, trypsin with a maximum of two missed cleavages; (ii) fixed modification, cysteine (C) carbamidomethylation. (iii) dynamic modifications, methionine (M) oxidation, asparagine deamidation of (N), and glutamine (Q), N-linked glycan residues were used as variable modifications of asparagine (N) residues. The mass tolerances for the searches were: precursor mass tolerance, 20 ppm; product mass tolerance, 20 ppm. The identification results were filtered at 1% false discovery rate with a confidence threshold set at a Byonic score > 100. The spectra were inspected manually to confirm retention time consistency and for the presence of oxonium and structure-specific ions. In the case of fucosylated and sialylated glycopeptides, monoisotopic peaks and corresponding diagnostic oxonium ion were closely inspected in the case of 2Fuc-1.02 = 1NeuAc to eliminate the false assignment of a sialic acid to the presences of 2 fucoses (NeuAc to 2Fuc) as we have previously described.24
The relative abundance of identified PSMA glycoforms in the two seminal plasma samples was performed using MS1 raw data and Byonic search results using Byologic (Protein Metrics Inc.). All the MS raw files and their corresponding Byonic search results from the quadruplicate technical replicate experiments were uploaded into one project for comparative analysis. The peak area of the extracted ion chromatogram (XIC) of any glycopeptide was integrated and normalized against the sum of XICs of all the PSMA glycopeptides that are identified in each MS run. In these analyses, only the glycoforms that were identified in 50% of the technical replicate experiments were used for subsequent analyses. Site-specific abundance of glycoforms was determined by obtaining the sum of all glycopeptides containing the same glycan at the asparagine residue where the peptide sequences may be different due to trypsin miscleavage or variable modifications. The relative abundance for each of the eight site-specific trypsin-derived glycopeptides that are amenable to analysis was determined at the total glycopeptide and glycosite levels as described previously.24
3 RESULTS
3.1 Identification of PSMA glycoforms in seminal plasma
Human PSMA possesses 10 predicted N-linked glycosylation sites in its primary amino acid sequence. We have previously performed comparative qualitative and quantitative analyses of PSMA glycoforms in LNCaP and MDAPCa2b PCa cell lines with different distant metastatic localization sites.24 Immunoaffinity purification was used to isolate PSMA from cell lysates followed by mass spectrometry analysis of intact glycopeptides. In these studies, we demonstrated that after trypsin digestion, eight of the nine glycopeptides that were generated, have m/z that are amenable for LC/MS/MS analysis and glycoform characterization. One of the peptides possesses the ASN-140 and ASN-153 glycosylation sites with a resulting high m/z that compromises characterization using standard LC/MS/MS-based approaches.24 We were able to characterize the glycoforms that are found in the other eight glycosylation sites of the two cell lines which included ASN-51, ASN-76, ASN-121, ASN-195, ASN-336, ASN-459, ASN-476, and ASN-638. Using this same experimental strategy, we evaluated expression profiles of PSMA glycoforms in patient seminal plasma compared to the two metastatic PCa cell lines, LNCaP and MDAPCa2b.
In this study, we have immunoaffinity purified and enriched PSMA from seminal plasma followed by SDS-PAGE separation and sample processing, and trypsin digestion before LC/MS/MS analysis. The isolation and purity of enriched PSMA were confirmed using standard SDS-PAGE and Western blot analysis as shown in Supporting Information S1: Figure 1. Intact N-linked glycopeptides of PSMA were analyzed by HCD and HCD-triggered EThcD MS/MS scans with quadruplicate sample injections. Byonic and Byologic softwares (Versions 5.2.31; Protein Metrics Inc.) were used to extract tandem MS fragmentation data from HCD and EThCD scans from the raw files using the parameters and filters as stated in Section 2.5 and as described previously.24, 35, 36 Glycoform identifications were confirmed as true positives after manual inspection and verification of LC retention times and the presence of peptide fragment ions (b/y and c/z) together with diagnostic oxonium ions to differentiate isobaric mass assignments including NeuAc and 2Fuc.24, 35-37 Positive glycoform identifications were only included for relative abundance analysis upon identification in at least two of the four technical replicate LC/MS/MS experiments. Mass spectrometry data for the identification of a high mannose (Man5) glycoform at the ASN-76 sequon, FLYN*FTQIPHLAGTEQNFQLAK are shown in Figure 1A–C. The glycoforms that are identified in PSMA from pooled seminal plasma from PCa patients with ISUP grades GG1 and GG2 and the relative abundance of the top five most abundant glycoforms at specific sequons are shown in Table 1.
Mass spectrometry identification of a high mannose (Man5) (HexNAc(2)Hex(5) glycoform of the FLYN*FTQIPHLAGTEQNFQLAK peptide from International Society of Urological Pathology GG2 seminal plasma. (A) Isotopic distribution pattern of the parent mass of the intact glycopeptide. (B) Extracted Ion chromatogram of the 4+MS1 precursor ion. (C) Tandem mass spectrometry fragmentation spectrum showing both b and y ions peptide fragment ions and oxonium ions identifying the glycoform. [Color figure can be viewed at wileyonlinelibrary.com]
| Sequon/peptide | GG1 most abundant glycans | % site | GG2 most abundant glycans | % site |
|---|---|---|---|---|
| ASN-51 SSNEATN*ITPK | ND | ND | ND | ND |
| ASN-76 FLYN*FTQIPHLAGTEQNFQLAK | HexNAc(2)Hex(7) | 17 | HexNAc(2)Hex(5) | 27 |
| HexNAc(2)Hex(6) | 15 | HexNAc(4)Hex(5)NeuAc(1) | 11 | |
| HexNAc(4)Hex(5)NeuAc(1) | 11 | HexNAc(3)Hex(6)NeuAc(1) | 7 | |
| HexNAc(2)Hex(5) | 10 | HexNAc(2)Hex(7) | 7 | |
| HexNAc(3)Hex(6) | 7 | HexNAc(4)Hex(5)Fuc(1)NeuAc(1) | 7 | |
| ASN-121 EFGLDSVELAHYDVLLSYPN*K | HexNAc(5)Hex(6)NeuAc(1) | 23 | HexNAc(5)Hex(6)NeuAc(1) | 31 |
| HexNAc(4)Hex(5)Fuc(3)NeuAc(2) | 10 | HexNAc(4)Hex(5)NeuAc(1) | 26 | |
| HexNAc(4)Hex(5)Fuc(2)NeuAc(1) | 9 | HexNAc(5)Hex(6) | 7 | |
| HexNAc(4)Hex(5)NeuAc(2) | 8 | HexNAc(4)Hex(5)NeuAc(2) | 7 | |
| HexNAc(5)Hex(6)NeuAc(2) | 8 | HexNAc(6)Hex(7)Fuc(3) | 4 | |
| ASN-195 IN*CSGK | HexNAc(4)Hex(5)NeuAc(2) | 22 | HexNAc(5)Hex(6)NeuAc(2) | 24 |
| HexNAc(5)Hex(6)NeuAc(2) | 20 | HexNAc(4)Hex(5)Fuc(1)NeuAc(1) | 22 | |
| HexNAc(4)Hex(5)Fuc(1)NeuAc(1) | 20 | HexNAc(4)Hex(5)NeuAc(2) | 17 | |
| HexNAc(5)Hex(6)Fuc(1)NeuAc(2) | 17 | HexNAc(5)Hex(6)Fuc(1)NeuAc(2) | 16 | |
| HexNAc(5)Hex(6)NeuAc(3) | 5 | HexNAc(5)Hex(6)NeuAc(3) | 8 | |
| ASN-336 VPYNVGPGFTGN*FSTQK | HexNAc(6)Hex(7)Fuc(1)NeuAc(3) | 16 | HexNAc(6)Hex(7)Fuc(1)NeuAc(3) | 19 |
| HexNAc(6)Hex(7)Fuc(3)NeuAc(2) | 12 | HexNAc(6)Hex(7)Fuc(3)NeuAc(2) | 13 | |
| HexNAc(6)Hex(7)Fuc(1)NeuAc(2) | 12 | HexNAc(6)Hex(7)Fuc(1)NeuAc(2) | 13 | |
| HexNAc(6)Hex(7)Fuc(3)NeuAc(1) | 10 | HexNAc(6)Hex(7)Fuc(3)NeuAc(1) | 11 | |
| HexNAc(6)Hex(7)Fuc(1)NeuAc(1) | 8 | HexNAc(6)Hex(7)Fuc(1)NeuAc(4) | 6 | |
| ASN-459 GVAYINADSSIEGN*YTLR | HexNAc(2)Hex(7) | 29 | HexNAc(2)Hex(8) | 29 |
| HexNAc(2)Hex(8) | 25 | HexNAc(2)Hex(7) | 28 | |
| HexNAc(2)Hex(6) | 10 | HexNAc(2)Hex(6) | 16 | |
| HexNAc(4)Hex(5)NeuAc(2) | 6 | HexNAc(4)Hex(5)NeuAc(1) | 5 | |
| HexNAc(4)Hex(5)Fuc(1) | 5 | HexNAc(4)Hex(5)Fuc(1) | 5 | |
| ASN-476 VDCTPLMYSLVHN*LTK | HexNAc(2)Hex(5) | 36 | HexNAc(2)Hex(5) | 54 |
| HexNAc(2)Hex(6) | 34 | HexNAc(2)Hex(6) | 41 | |
| HexNAc(2)Hex(7) | 30 | HexNAc(2)Hex(8) | 4 | |
| ASN-638 N*FTEIASK | HexNAc(2)Hex(6) | 65 | HexNAc(2)Hex(8) | 24 |
| HexNAc(2)Hex(8) | 9 | HexNAc(2)Hex(7) | 18 | |
| HexNAc(4)Hex(5) | 7 | HexNAc(4)Hex(5) | 18 | |
| HexNAc(2)Hex(7) | 6 | HexNAc(2)Hex(5) | 11 | |
| HexNAc(4)Hex(5)NeuAc(1) | 4 | HexNAc(4)Hex(5)NeuAc(1) | 8 |
- Abbreviations: ND, not detected; PCa, prostate cancer; PSMA, prostate-specific membrane antigen.
3.2 PSMA from seminal plasma is highly fucosylated and sialylated
Our previous study using PCa cell lines24 and another study using recombinantly expressed PSMA isolated from insect S2 cells and HEK293T17 cells, and a limited number of prostate tissue samples,23 have both shown site-specific expression profiles and spatial distribution of specific glycoforms in the protein. Here, we present glycosylation profiles that are detected in PSMA isolated from the seminal plasma of PCa patients. As previously observed, for the ASN-51 sequon, there was a significant reduction in glycosylation levels at this sequon and with the filters applied for searches and identification using Byonic, no glycoforms were detected at this particular sequon of PSMA isolated from seminal plasma from ISUP GG1/GG2 patients. This is consistent with our previous data and results from other investigators that demonstrate reduced glycosylation at this sequon using PSMA isolated from different sources.23, 24 Similar to our previous results, three PSMA sequons, namely, ASN-459, ASN-476, and ASN-638 are predominantly conjugated with high mannose glycans. For ASN-459, the total relative abundance of high mannose structures in the top five most abundant glycans in ISUP GG1 and GG2 samples are 85% and 87%, respectively. The most abundant glycoforms at this sequon are Man7 and Man8 and the mass spectrometry results including isotope distribution plot, extracted ion chromatogram, and MS2 fragmentation data for the Man8 glycoform are shown in Supporting Information S1: Figure 2. For ASN-476, all the identified glycans are high mannose (100%) in all patient samples. The glycoforms that are identified at the sequon include Man5, Man6, Man7, and Man8, and Supporting Information S1: Figure 3 shows a representative isotope distribution plot, extracted ion chromatogram, and MS2 fragmentation data identifying a Man6 glycan conjugated to ASN-476. For ASN-638, the relative abundance of high mannose glycans is 95% and 90%, respectively, for PSMA from GG1 and GG2 patient samples. The most abundant glycoforms are also Man5, Man6, Man7, and Man8. Supporting Information S1: Figure 4 shows a representative isotope distribution plot, extracted ion chromatogram, and MS2 fragmentation data identifying a Man6 glycan conjugated to ASN-638. The sequon ASN-76 has a reduced abundance of high mannose structures compared to these three other sites, with relative abundance of 70% and 56%, respectively, in ISUP GG1 versus GG2 patient samples. For this particular sequon, in the ISUP GG1 patient pool, a sialylated complex structure with composition HexNAc(4)Hex(5)NeuAc(1) is the second most abundant at 18% and a hybrid structure with composition HexNAc(3)Hex(6) at an abundance of 12%. In ISUP GG2 patient pool, 46% of the remaining glycans are sialylated complex glycans with a single glycan that is also fucosylated. The other three sequons ASN-121, ASN-195, and ASN-336 are predominantly conjugated with complex sialylated glycans with varying levels of fucosylation. In the sequon ASN-121, for the GG1 group sample, all the top abundant glycan species (100%) are sialylated with a corresponding 33% fucosylation. In contrast, we observed differences in the abundance and composition of the glycans at the same sequon in the ISUP GG2 samples, where 84% of the top abundant glycan species are sialylated and these species are not fucosylated. The remaining glycans are comprised of a triantennary complex structure that is not sialylated nor fucosylated (HexNAc(5)Hex(6)) at 9% relative abundance and a triantennary complex structure with three fucoses (HexNAc(6)Hex(7)Fuc(3)) at 5% relative abundance. The most abundant glycoform at this sequon in both the GG1 and GG2 group samples is a triantennary monosialylated glycan with composition HexNAc(4)Hex(5)NeuAc(1). The mass spectrometry results including isotope distribution plot, extracted ion chromatogram, and MS2 fragmentation are shown in Figure 2A–C. In the sequon ASN-195, and in the two samples GG1 and GG2, all of the top abundant glycans are complex sialylated species (100%) with a corresponding 43% fucosylation and complete overlap of identified glycans in the two groups. The most abundant glycoform at this sequon in both patient group samples is a triantennary disialylated glycan with composition HexNAc(4)Hex(5)NeuAc(2). The mass spectrometry results including isotope distribution plot, extracted ion chromatogram, and MS2 fragmentation are shown in Figure 3A–C. Similarly, in ASN-336, all of the top five abundant glycoforms are complex species that are both sialylated and fucosylated. There is a significant overlap in the composition of the glycoforms that are identified at this site between the two patient groups. Four glycans are identified in both groups and two are uniquely detected in either group, HexNAc(6)Hex(7)Fuc(1)NeuAc(1) in GG1 and HexNAc(6)Hex(7)Fuc(1)NeuAc(4) in GG2 sample. The most abundant glycoform at this sequon in both the GG1 and GG2 samples is a tetraantennary fucosylated tri-sialylated glycan with composition HexNAc(6)Hex(7)Fuc(1)NeuAc(3). The mass spectrometry results including isotope distribution plot, extracted ion chromatogram, and MS2 fragmentation are shown in Figure 4A–C. The identified glycan is core fucosylated as confirmed by the presence of the diagnostic ions of the 1+ and 2+ charged peptide+HexNAc+Fuc (m/z 2162.0288 and 1081.5180 respectively) and the 2+ charged peptide+2HexNAc+Fuc (m/z 1183.0577). Further inspection of the MS2 spectra reveals the presence of α2-3 linkages of terminal sialic acids with the diagnostic fragment at m/z 454.1 albeit at a low intensity. The corresponding oxonium ions with formulae HexNAc, HexHexNAc, NeuAc, NeuAc-H2O, HexNAcHexNeuAc with m/z 204.09, 366.13, m/z 292.10, m/z 274.09, and m/z 657.23, respectively, are identified in the MS2 spectra with high intensities. This would indirectly imply that a significant proportion of the observed sialylation comprises of α2-3 linked moieties. A summary of the five most abundant glycans that are identified in the seven PSMA N-linked glycosylation sites that we have characterized in these seminal plasma samples is provided in Supporting Information S1: Table 1.
Mass spectrometry identification of a complex triantennary mono-sialylated HexNAc(5)Hex(6)NeuAc(1) glycoform of the EFGLDSVELAHYDVLLSYPN*K peptide from GG2 seminal plasma. (A) Isotopic distribution pattern of the parent mass of the intact glycopeptide. (B) Extracted Ion chromatogram of the 3+MS1 precursor ion. (C) Tandem mass spectrometry fragmentation spectrum showing both b and y ions peptide fragment ions and oxonium ions identifying the glycoform. The corresponding Symbol Nomenclature For Glycans monosaccharide symbols and nomenclature are provided. [Color figure can be viewed at wileyonlinelibrary.com]
Mass spectrometry identification of a complex triantennary disialylated HexNAc(5)Hex(6)NeuAc(1) glycoform of the IN*CSGK peptide from GG2 seminal plasma. (A) Isotopic distribution pattern of the parent mass of the intact glycopeptide. (B) Extracted Ion chromatogram of the 3+MS1 precursor ion. (C) Tandem mass spectrometry fragmentation spectrum showing both b and y ions peptide fragment ions and oxonium ions identifying the glycoform. The corresponding Symbol Nomenclature For Glycans monosaccharide symbols and nomenclature are provided. [Color figure can be viewed at wileyonlinelibrary.com]
Mass spectrometry identification of a complex tetraantennary core fucosylated tri-sialylated HexNAc(6)Hex(7)NeuAc(1) glycoform of the HexNAc(6)Hex(7)Fuc(1)NeuAc(3) peptide from GG2 seminal plasma. (A) Isotopic distribution pattern of the parent mass of the intact glycopeptide. (B) Extracted Ion chromatogram of the 4+MS1 precursor ion. (C) Tandem MS fragmentation spectrum showing both b and y ions peptide fragment ions and oxonium ions identifying the glycoform that includes intact peptide and core fucosylated HexNAc+Fuc sugar moiety with m/z 2162.0288 (+1). The corresponding Symbol Nomenclature For Glycans monosaccharide symbols and nomenclature are provided. [Color figure can be viewed at wileyonlinelibrary.com]
A total of 54 unique glycans were identified under these experimental and analytical conditions. The structural formulae of these glycans using the corresponding Symbol Nomenclature For Glycans monosaccharide symbols are provided in Supporting Information S1: Table 2. Forty-nine glycans were identified in the ISUP GG1 pooled samples and 54 in ISUP GG2 pooled samples. Forty-six glycans are shared between the two sample groups demonstrating an 85% overlap in their distribution. In the GG1 sample, 63% of the identified glycans are fucosylated and 67% possess sialic acid, whereas in the GG2 sample, 59% are fucosylated and 71% are sialylated. After taking all the glycoforms that are identified at the seven sequons into account, we observed a significant overlap in PSMA glycoforms in seminal plasma from GG1 and GG2 PCa patients. At ASN-76, 85% of the glycoforms are identified in both cohorts, the remaining 15% are exclusively found in GG1 patients and the differences are attributed to varying levels of fucosylation and sialylation. The same situation is observed in ASN-121 with a 58% overlap and 25% and 17% uniquely present in the GG1 and GG2 samples respectively. In the case of ASN-195, the overlap is 90% and the mono-sialylated biantennary glycan with structure HexNAc(4)Hex(5)NeuAc(1) was found only in the GG1 cohort. The identification of the disialylated version of the glycoform with the structure HexNAc(4)Hex(5)NeuAc(2) in both GG1 and GG2 samples would imply that the monosialylated glycoform completely undergoes additional sialylation into the disialylated form in GG2. In the case of ASN-336, there is an overlap of 91% with a tetra-antennary difucosylated glycoform with the structure HexNAc(6)Hex(7)Fuc(2) being identified exclusively in GG1 and two others glycoforms with structures HexNAc(3)Hex(4)Fuc(1)NeuAc(1) and HexNAc(5)Hex(6)NeuAc(3) that are identified exclusively in GG2 samples. At this sequon, the HexNAc(6)Hex(7)Fuc(2) difucosylated glycoform is potentially an intermediate glycoform in the synthetic pathway that leads to the formation of the trifucosylated glycoform that is identified in both sample cohorts. A similar situation is observed at ASN-459 where the overlap is also 91% with the identification of potential triantennary glycan structures with structures HexNAc(5)Hex(6)Fuc(2)NeuAc(2) and HexNAc(5)Hex(6)Fuc(3)NeuAc(2) in the GG1 and GG2 samples, respectively. The ASN-476 sequon is interesting, only high mannose glycoforms were detected at this sequon with Man5 and Man6 glycans being detected in both GG1 and GG2 samples, in contrast, Man7 was exclusively identified in GG1 and Man8 exclusively in GG2. For the last sequon, ASN-638, the overlap is 62% where seven glycoforms that include Man6, a hybrid glycan, and a combination of six sialylated and fucosylated tri- and tetraantennary structures with different levels of sialylation and fucosylation are exclusively identified in GG1 sample. A similar situation is seen in the GG2 sample where a Man5 and three difucosylated and monosialylated bi-, tri-, and tetraantennary glycans are identified. The overall distribution and abundance of the glycoforms that are identified in the sequons that we have characterized in the two sample groups are summarized in Supporting Information S1: Figure 5. Our future objective is to perform extensive comparative quantitative analyses to determine the expression profiles of the glycoforms at these different sequons using individual disease-stratified post-digital rectal examination (DRE) urine samples from PCa patients.
4 DISCUSSION
Currently, PSMA and PSMA targeting molecules are being used in the clinic or are under development and evaluation as PCa diagnostics, prognostics, and therapeutics. The molecule is a particularly attractive target for these applications because of its localization, orientation, and surface topology on prostate cell membranes. These molecular characteristics provide the basis for utilizing PSMA as an attractive target for imaging, immunotherapy, and radiotherapy. To date, all theranostic applications have targeted the peptidomic domains of the protein. However, PSMA is highly glycosylated and almost 30% of the molecular mass of the protein is contributed by its conjugated carbohydrate domains.21 The composition of N-linked glycans that are conjugated to PSMA are poorly characterized and, thus, the role of these glycoforms in disease progression and their potential utility as biomarkers is unknown. As we and others have shown, PSMA is differentially glycosylated across cell lines, and deglycosylation using glycosidases or site-directed mutagenesis completely abolishes PSMA folate hydrolase and NAALADase enzymatic activity.21, 22
In this study, we demonstrate that PSMA is detectable in seminal plasma, and for the first time characterize and describe PSMA glycoforms in seminal plasma obtained from patients with GG1 and GG2 (Gleason Score 3 + 3 and 3 + 4) PCa. This is the first report describing the characterization of PSMA glycoforms in seminal plasma using mass spectrometry-based glycoproteomics. Our comprehensive analysis revealed that in PSMA isolated from seminal plasma, high mannose glycan structures are predominantly conjugated to three sequons namely ASN-459, ASN-476, and ASN-638 in both ISUP GG1 and GG2. These sequons are located at the basal domain of the protein that topologically faces the plasma membrane in cells (Protein Data Bank Code 2C6C38). The ASN-76 sequon which is located in the intermediate domain of the protein shows a mixture of high mannose glycans and complex glycans. The sequons ASN-121, ASN-195, and ASN-336 that are located and exposed at the apical domain of the protein predominantly possess complex sialylated and fucosylated N-linked glycans. In PCa cells, these sequons and their conjugated glycans are localized and exposed at the apical domain of PSMA on their plasma membrane. It is plausible that these glycans may have functional relevance in mediating and modulating interactions with carbohydrate-binding ligands with potential implications in disease progression. The identification of α2-3 linkages of terminal sialic acids glycoforms at ASN-336 is particularly noteworthy as this structure has been considered a hallmark of cancer and may have utility as a prognostic biomarker of PCa.39-43 In our prior study of PSMA glycoforms in metastatic LNCaP and MDAPCa2b PCa cell lines, we observed similar profiles with respect to high mannose expression at ASN-459, ASN-476 and ASN-638 sequons respectively. However, there were significant qualitative and quantitative differences in the expression profiles of glycoforms at some of the other sequons.24 As an example, in ASN-336, there were statistically significant differences in fucosylation and sialylation of glycans at this sequon with increased fucosylation in LNCaP versus MDAPCa2b and conversely an increase in sialylation in MDAPCa2b versus diminished sialylation in LNCaP cells. LNCaP cells have a lymph node metastatic homing preferences whereas MDAPCa2b are bone metastatic. This would imply that differential glycosylation may be associated with metastasis to specific niches. In this study, relatively minor qualitative glycosylation differences are observed between pooled GG1 and GG2 seminal plasma samples from PCa patients. This may plausibly be due to relatively low differences in disease severity between the low- and low-intermediate-risk group patients. As previously postulated, the observed differential expression with high mannose glycoforms at ASN-459, ASN-476, and ASN-638 in the basal domain of PSMA, and predominantly complex structures glycoforms ASN-121, ASN-195, and ASN-336 that are located at the apical domain may be due to their differential access to enzymes in the N-glycosylation pathway.23 We have further compared the glycoforms that are identified in this study using seminal plasma fluids from GG1 and GG2 PCa patients versus our previous study focusing on LNCaP and MDAPCa2b PCa cell lines.24 Supporting Information S1: Figure 6 provides a schematic of PSMA showing the cytoplasmic, helical, and extracellular domains of the PSMA, the enzymatic functional epitopes of the PSMA including protease activity, NAALADase activity, peptidase_28 and a transferrin-like receptor dimer epitope at the C-terminus of the protein. The 10 N-linked glycosylation sites of PSMA are included in the diagram and the effects of loss of glycosylation at these sites through N → A mutations as previously reported.21 Importantly, the loss of glycosylation is associated with a reduction of enzymatic activity at different levels in a glycosite-dependent manner. It is important to note that loss of glycosylation at ASN-638 abolishes enzyme activity.21 Taken together, the following conclusions can be made, the glycosylation profiles that are observed in seminal fluid are more closely aligned with those of MDAPCa2b by possessing a higher degree and abundance of sialylated glycoforms compared to LNCaP cells at ASN-121, ASN-195, and ASN-336. At these three sites, the most abundant glycoforms in LNCaP cells possess both core and terminal fucosylation24 as summarized in Supporting Information S1: Figure 6.
The main weakness of our study lies in the observed differences between the low and low-intermediate risk groups. Specifically, we used pooled samples that likely mask individual variability, and our overall numbers only allow for exploratory conclusions. Additional studies will be required to separately analyze individual seminal plasma samples from a broader spectrum of disease risk, an effort that will benefit from our current conclusions. The objective will be to determine whether PSMA glycoforms have utility in discriminating different stages of PCa and in the development of diagnostic and prognostic markers. In addition, under the current experimental conditions using trypsin as protease, it is not possible to identify glycoforms at two sequons ASN-140 and ASN-153. The characterization of the glycoforms at these sequons would require the utilization of different protease/s with different site specificities and the ability to generate glycopeptides that can be analyzed using standard LC/MS conditions. We provide novel descriptive and comparative glycosylation data for PSMA in seminal plasma from PCa patients GG1 and GG2. These studies form the basis for the discovery and development of PSMA glycoform-based biomarkers for PCa detection, diagnosis, and prognosis using prostate proximal fluids including post-DRE urines obtained from men using minimally invasive methods.
ACKNOWLEDGMENTS
This work was supported by funds from the US Department of Defense (DOD)/Congressionally Directed Medical Research Programs through the Prostate Cancer Research Program Award Number W81XWH-18-1-0228 PC170843 to Julius O. Nyalwidhe. Opinions, interpretations, conclusions, and recommendations are those of the authors and are not necessarily endorsed by the DOD (US Army Medical Research Acquisition Activity). This work was supported by the Eastern Virginia Medical School Biorepository and we acknowledge the excellent technical support and assistance provided by Mary Ann Clements and Brian P. Main.
CONFLICT OF INTEREST STATEMENT
The authors declare no conflict of interest.
REFERENCES
Comments
Post a Comment