Background:Ductal carcinoma in situ (DCIS) is a noninvasive breast lesion with variable risk of progression to invasive breast cancer (IBC). Current transcription and cell marker investigations suggest ECM decreases in later events but are limited in details of ECM proteomic composition, including post-translational modifications. We investigated whether the extracellular matrix (ECM) proteome alters with later breast events of DCIS or IBC. Methods:ECM-targeted mass spectrometry imaging and liquid chromatography-tandem mass spectrometry (LC-MS/MS) were applied to ten tissue microarrays from the Resource of Archival Human Breast Tissue cohort (RAHBT). Primary DCIS specimens (n=136) were analyzed in relation to later events of DCIS (n=40) or IBC(n=30), with a mean follow-up of 192.1 months 95% CI [179.1,205.1]. Statistical modeling, survival analyses, and exploratory machine learning approaches were used to identify ECM peptide signatures associated with later events. Results:Distinct ECM peptide profiles were associated with later events of DCIS or IBC. Fifteen peptides derived from fibrillar collagens (COL1A1, COL1A2, COL3A1) and elastin, showed significantly reduced abundance in patients who developed IBC. Lower expression of specific collagen peptides associated with overall 19.9% 95% CI [17.92, 21.81] decreased disease-free survival for IBC. Lower expression of these peptides was significantly associated with reduced disease-free survival (age-adjusted hazard ratio [HR] = 2.45, 95% CI: 2.33-2.57; P < 0.05). Patient-matched samples of primary DCIS, later DCIS, and later invasive breast cancer further demonstrated reduction in ECM peptide detection. Exploratory predictive modeling from patient-matched samples achieved high performance (AUROC >0.98, accuracy >93%) in distinguishing primary from later events. Following prior work in the RAHBT cohort, reduction of certain collagen peptides was also observed in primary DCIS samples from higher risk patient groups. Conclusions:ECM proteomic remodeling, particularly decreases of specific collagen domains, is strongly associated with later events of DCIS and IBC. These findings highlight ECM proteome as a critical regulator of breast cancer emergence with potential as a prognosticator of risk stratification to guide clinical management of DCIS.

Hepatocellular carcinoma (HCC) mortality rates continue to increase faster than those of other cancer types due to high heterogeneity, which limits diagnosis and treatment. Pathological and molecular subtyping have identified that HCC tumors with poor outcomes are characterized by intratumoral collagenous accumulation. However, the translational and post-translational regulation of tumor collagen, which is critical to the outcome, remains largely unknown. Here, we investigate the spatial extracellular proteome to understand the differences associated with HCC tumors defined by Hoshida transcriptomic subtypes of poor outcome (Subtype 1; S1; n = 12) and better outcome (Subtype 3; S3; n = 24) that show differential stroma-regulated pathways. Collagen-targeted mass spectrometry imaging (MSI) with the same-tissue reference libraries, built from untargeted and targeted LC-MS/MS was used to spatially define the extracellular microenvironment from clinically-characterized, formalin-fixed, paraffin-embedded tissue sections. Collagen α-1(I) chain domains for discoidin-domain receptor and integrin binding showed distinctive spatial distribution within the tumor microenvironment. Hydroxylated proline (HYP)-containing peptides from the triple helical regions of fibrillar collagens distinguished S1 from S3 tumors. Exploratory machine learning on multiple peptides extracted from the tumor regions could distinguish S1 and S3 tumors (with an area under the receiver operating curve of ≥0.98; 95% confidence intervals between 0.976 and 1.00; and accuracies above 94%). An overall finding was that the extracellular microenvironment has a high potential to predict clinically relevant outcomes in HCC.

BackgroundWorldwide, hepatocellular carcinoma (HCC) is the second most lethal cancer, although early-stage HCC is amenable to curative treatment and can facilitate long-term survival. Early detection has proved difficult, as proteomics, transcriptomics, and genomics have been unable to discover suitable biomarkers.MethodsTo find new biomarkers of HCC, we utilized a spatial omics N-glycan imaging method to identify altered glycosylation in cancer tissue (n = 53) and in paired serum of individuals with HCC (n = 23). Glycoproteomics identified the glycoproteins carrying these N-glycan structures, and we utilized an antibody array-based glycan imaging method to examine all the N-glycans associated with the identified glycoproteins. N-glycans from the examined glycoproteins were used to create machine learning algorithms, which were tested in a case-control sample set of 100 patients with cirrhosis and HCC and 101 matched patients with cirrhosis alone.ResultsSpatial glycan imaging identifies thirteen branched, fucosylated, and high mannose glycans as altered in HCC tissue and in matched patient serum. Glycoproteomics identifies over 50 proteins containing these changes, of which sixteen glycoproteins were selected for further testing in an independent patient set. Algorithms using a combination of glycan and glycoproteins accurately differentiate early-stage and all HCC from cirrhosis with AUROC values of 0.88-0.97.ConclusionsIn conclusion, we present the development and application of a new biomarker platform, which can identify effective biomarkers for the early detection of HCC. This platform may also apply to other diseases, in which changes in N-linked glycosylation are known to occur.

Patient demographics for A. TMA 1 and B. TMA 2. C. Patient demographics for serum cohort. ALT (Alanine transaminase), AST (aspartate aminotransferase), AFP (Alpha-fetoprotein), and ALP (Alkaline phosphatase). Other liver diseases include nonalcoholic steatohepatitis, hepatitis C with cirrhosis, hepatic adenoma, benign fibrotic gallbladder disease, and diabetes. PSC (Primary Sclerosing cholangitis), HCC (Hepatocellular carcinoma), and OLD (Other liver diseases). Gray shading for missing clinical information. D. Representative N-glycan images of TMA 1 (top) and TMA 2 (bottom) of 2012.717m/z (left) and 1809.646m/z (right). Red boxes select for CCA samples. E. Table details other pathology diagnoses included in TMA 1 with the proposed structure for the N-glycans highly expressed in each. These modifications are characterized based on 1-2 patients.

A. Scatterplots and correlations between N-glycans of interest (1339, 1257, 2158) and clinical information available (ALT, AST, ALK, and AFP). B. Multivariate model-Multiple logistic regression in CCA and PSC serum samples (n=40). Model 1: three N-glycans of interest and clinical information available (left panel). Model 2: Only the three N-glycans of interest (right panel). C. ROC curve of the combination of glycans and clinical information (red-solid line) and only glycans (green-dashed line) (left), classification performance table of the two models in B. p=0.5731, Delong’s test between model 1 and 2. CA19-9, ALT, AST, ALK, and AFP values were log-transformed for plotting and modeling convenience. logALT was removed from the multiple logistic regression analysis due to a high value of Variance inflation factor (VIF).

Peak list of N-glycans detected in tissue (TMA) and serum samples.

A. Relative contribution of each serum-glycan (left) and TMA-glycan (right) in the first and second principal components. The size and color of the circle represent a higher contribution of the glycan to the respective Dim. (Dimension). B. Three N-glycans in TMA and serum were identified after optimization. C. Relative intensity quantification for both TMAs of the respective N-glycan based on small and large duct classification. D. Two common N-glycans were identified between datasets. LOOCV (Leave-One-Out Cross-Validation). AUC (Area Under the Curve). E. Quantification of the relative contribution of N-glycans (left), table of N-glycans, and proposed structure with importance values (top) when removing N-glycan at 1339 m/z from the analysis. F. ROC (Receiving Operator Characteristic) curve classification for serum (left) ROC curve LOOCV (right) of N-glycans in E. G. List combinations of N-glycan as possible biomarkers to differentiate iCCA from PSC. For N-glycans, red triangle, fucose; blue square, N-acetylglucosamine; green circles, mannose; yellow circles, galactose.

Relative intensity quantification of all N-glycans identified in serum (left row) and tissue (right row) analysis. Red font labeling for N-glycans follows the same trend between serum and tissue.

Data description and detailed statistical information regarding model development. Supplementary Table 1. Patient information on samples from the University of Michigan. Supplementary Table 2. Patient information on samples from HALT-C. Supplementary Table 3. Patient information on samples from EDRN. Supplementary Table 4. Patient information on samples from Thomas Jefferson University. Supplementary Table 5. Patient information on samples from the University of Texas Southwestern Medical Center. Supplementary Table 6. Odds ratios for each predictor in univariate and multivariate logistic regressions. Supplementary Table 7. Indices of goodness-of-fit and apparent validation of candidate logistic regression models. Supplementary Table 8. Performance of leave one out cross validation (LOOCV). Supplementary Table 9. Cross Validation of bootstrap method. Supplementary Table 10. AUCs and IDIs from 3-fold Cross-Validation. Supplementary Table 11. Summary statistics of Doylestown model. Supplementary Table 12. Comparison of logistic regression, classification and regression tree (CART) and conditional inference tree (CTREE). Supplementary Figure 1. The distribution of AFP in each data set by cases and controls: Supplementary Figure 2. Quartiles for AFP in the individual patient sets. Supplementary Figure 3. Histogram of predictions and observe occurrence proportion of HCC for the top 4 models.

Hematoxylin & Eosin (H&E) staining of A. Intrahepatic Cholangiocarcinoma (iCCA) tissue, 10x (right) and 40x (left) magnification images from respective areas of the tissue, and B. Hepatocellular Carcinoma (HCC) tissue, 10x magnification images from respective areas of the tissue. Tumor regions are outlined in red, normal areas are outlined in black and fibrotic regions are outlined in blue. C. TMA H&E staining with an outline that specifies the diagnosis for each core for TMA 1 (left) and TMA 2 (right). Mixed carcinoma: HCC and iCCA. Small (yellow) and large (purple) ducts classifications for each TMA.

Supplementary Figure 1 from Novel Changes in Glycosylation of Serum Apo-J in Patients with Hepatocellular Carcinoma

PDF file, 354KB, N-linked glycosylation of total protein from 16 HCC tissue and adjacent liver tissue pairs.

There is an urgent need for the identification of reliable prognostic biomarkers for patients with intrahepatic cholangiocarcinoma (iCCA) and alterations in N-glycosylation have demonstrated an immense potential to be used as diagnostic strategies for many cancers, including hepatocellular carcinoma (HCC). N-glycosylation is one of the most common post-translational modifications known to be altered based on the status of the cell. N-glycan structures on glycoproteins can be modified based on the addition or removal of specific N-glycan residues, some of which have been linked to liver diseases. However, little is known concerning the N-glycan alterations that are associated with iCCA. We characterized the N-glycan modifications quantitatively and qualitatively in three cohorts, consisting of two tissue cohorts: a discovery cohort (n = 104 cases) and a validation cohort (n = 75), and one independent serum cohort consisting of patients with iCCA, HCC, or benign chronic liver disease (n = 67). N-glycan analysis in situ was correlated to tumor regions annotated on histopathology and revealed that bisected fucosylated N-glycan structures were specific to iCCA tumor regions. These same N-glycan modifications were significantly upregulated in iCCA tissue and serum relative to HCC and bile duct disease, including primary sclerosing cholangitis (PSC) (P < 0.0001). N-glycan modifications identified in iCCA tissue and serum were used to generate an algorithm that could be used as a biomarker of iCCA. We demonstrate that this biomarker algorithm quadrupled the sensitivity (at 90% specificity) of iCCA detection as compared with carbohydrate antigen 19-9, the current "gold standard" biomarker of CCA.Significance: This work elucidates the N-glycan alterations that occur directly in iCCA tissue and utilizes this information to discover serum biomarkers that can be used for the noninvasive detection of iCCA.

Supplementary Table S1. Patient Characteristics for patients from the University of California at San Diego; Supplementary Table S2. Fitness of algorithms; Supplementary Table S3. Cross validations of potential models; Supplementary Table S4. Statistical inference of comparing models; Supplementary Figure S1. Scheme of study design;Supplementary Figure S2. Scatter plot; Supplementary Figure S3. AUROC for the individual components analyzed; Supplementary Figure S4. MALDI-TOF analysis of low molecular weight kininogen; Supplementary Figure S5. Glycopeptide analysis of tryptic glycopeptide 44-58; Supplementary Figure S6. Glycopeptide analysis of tryptic glycopeptide 197-208 showing identification of fucosylated glycopeptides; Supplementary Figure S7. Glycopeptide analysis of tryptic glycopeptide 289-300;

Our group has recently developed the GlycoTyper assay which is a streamlined antibody capture slide array approach to directly profile N-glycans of captured serum glycoproteins including immunoglobulin G (IgG). This method needs only a few microliters of serum and utilizes a simplified processing protocol that requires no purification or sugar modifications prior to analysis. In this method, antibody captured glycoproteins are treated with peptide N-glycosidase F (PNGase F) to release N-glycans for detection by MALDI imaging mass spectrometry (IMS). As alterations in N-linked glycans have been reported for IgG from large patient cohorts with fibrosis and cirrhosis, we utilized this novel method to examine the glycosylation of total IgG, as well as IgG1, IgG2, IgG3 and IgG4, which have never been examined before, in a cohort of 106 patients with biopsy confirmed liver fibrosis. Patients were classified as either having no evidence of fibrosis (41 patients with no liver disease or stage 0 fibrosis), early stage fibrosis (10 METAVIR stage 1 and 18 METAVIR stage 2) or late stage fibrosis (6 patients with METAVIR stage 3 fibrosis and 37 patients with METAVIR stage 4 fibrosis (cirrhosis)). Several major alterations in glycosylation were observed that classify patients as having no fibrosis (sensitivity of 92% and a specificity of 90%), early fibrosis (sensitivity of 84% with 90% specificity) or significant fibrosis (sensitivity of 94% with 90% specificity).

We have previously identified alterations in glycosylation on serum proteins from patients with HCC and developed plate-based assays using lectins to detect the change in glycosylation. However, heterophilic antibodies, which increase with non-malignant liver disease, compromised these assays. To address this, we developed a method of polyethylene glycol (PEG) precipitation that removed the contaminating IgG and IgM but allowed for the lectin detection of the relevant glycoprotein. We found that this PEG-precipitated material itself could differentiate between cirrhosis and HCC. In the analysis of three training cohorts and one validation cohort, consisting of 571 patients, PEG-IgG had AUC values that ranged from 0.713 to 0.810. In the validation cohort, which contained samples from patients at a time of 1–6 months prior to HCC detection or 7+ months prior to detection, the AUC of this marker remained consistent (0.813 and 0.846, respectively). When this marker was incorporated into a biomarker algorithm that also consisted of AFP and fucosylated kininogen, the AUROC increased to 0.816–0.883 in the training cohort and was 0.909 in the external validation cohort. Biomarker performance was also examined though the analysis of partial ROC curves, at false positive values less than 10% (90-ROC), ≤20% (80-ROC) or ≤30% (70-ROC), which highlighted the algorithm’s improvement over the individual markers at clinically relevant specificity values.

Hepatocellular carcinoma (HCC) is the most common form of liver cancer and the fifth most common cancer overall. Late-stage therapeutic options are limited, while resection or ablation of small tumors can lead to overall survival rates of greater than 60 months. Therefore, early detection of HCC is crucial for patient survival. Currently, there are only three widely used biomarkers for HCC: α-fetoprotein (AFP), core fucosylated AFP (AFP-L3), and des-gamma-carboxy prothrombin (DCP). All three of these markers have shown some value in the detection of HCC but with limited sensitivity. While serum is hepatic in nature, the tissue origin of these biomarkers is not determinable based on serum analysis alone, despite the ability of AFP to detect later-stage cancers. Therefore, further tissue analysis is needed for improved detection. Here, multi-omic approaches of HCC tissue are discussed, beginning with large-scale analyses to identify larger biocommunication networks predominant in HCC progression and moving toward smaller and more specific analyses. In the large-scale studies, the data suggests dysregulation in many major pathways, specifically the β-catenin/WNT and RAS pathways, and in specific sub-types, changes in specific genes such as TP53, TERT, and CTNNB1. Proteomics and glycomics are of special interest due to the glycosylation changes observed with AFP in HCC cases. From studying these glycomic and proteomic profiles of HCC serum and tissue, many groups have identified increased fucosylation and branching that are related to presence and progression of HCC. New techniques such as MALDI mass spectrometry glycan imaging have been used to identify specific glycan changes in cancer tissue. Further studies are necessary to accurately pinpoint the location of these modified glycoproteins to tumor-specific regions due to the heterogeneous composition of HCC tissue and to identify all the proteins that are modified and could act as potential biomarkers for HCC.

(1) Glycoproteins account for ~80% of proteins located at the cell surface and in the extracellular matrix. A growing body of evidence indicates that α-L-fucose protein modifications contribute to breast cancer progression and metastatic disease. (2) Using a combination of techniques, including matrix-assisted laser desorption/ionization imaging mass spectrometry (MALDI-IMS) based in cell and on tissue imaging and glycan sequencing using exoglycosidase analysis coupled to hydrophilic interaction ultra-high performance liquid chromatography (HILIC UPLC), we establish that a core-fucosylated tetra-antennary glycan containing a single N-acetyllactosamine (F(6)A4G4Lac1) is associated with poor clinical outcomes in breast cancer, including lymph node metastasis, recurrent disease, and reduced survival. (3) This study is the first to identify a single N-glycan, F(6)A4G4Lac1, as having a correlation with poor clinical outcomes in breast cancer.

A new platform for N-glycoprotein analysis from serum that combines matrix-assisted laser desorption/ ionization mass spectrometry imaging (MALDI MSI) work- flows with antibody slide arrays is described. Antibody panel based (APB) N-glycan imaging allows for the specific capture of N-glycoproteins by antibodies on glass slides and N-glycan analysis in a protein-specific and multiplexed manner. Development of this technique has focused on characterizing two abundant and well-studied human serum glycoproteins, alpha-1-antitrypsin and immunoglobulin G. Using purified standard solutions and 1 mu L samples of human serum, both glycoproteins can be immunocaptured and followed by enzymatic release of N-glycans. N-Glycans are detected with a MALDI FT-ICR mass spectrometer in a concentration-dependent manner while maintaining specificity of capture. Importantly, the N-glycans detected via slide-based antibody capture were identical to that of direct analysis of the spotted standards. As a proof of concept, this workflow was applied to patient serum samples from individuals with liver cirrhosis to accurately detect a characteristic increase in an IgG N-glycan. This novel approach to protein-specific N-glycan analysis from an antibody panel can be further expanded to include any glycoprotein for which a validated antibody exists. Additionally, this platform can be adapted for analysis of any biofluid or biological sample that can be analyzed by antibody arrays.