Proteomics Platform Research Articles

Exploring the role of circulating proteins in multiple myeloma risk: a Mendelian randomization study

Multiple myeloma (MM) is an incurable blood cancer with unclear aetiology. Proteomics is a valuable tool in exploring mechanisms of disease. We investigated the causal relationship between circulating proteins and MM risk, using two of the largest cohorts with proteomics data to-date. We performed bidirectional two-sample Mendelian randomization (MR; forward MR = causal effect estimation of proteins and MM risk; reverse MR = causal effect estimation of MM risk and proteins). Summary statistics for plasma proteins were obtained from genome-wide association studies performed using SomaLogic (N = 35,559; deCODE) and Olink (N = 34,557; UK Biobank; UKB) proteomic platforms and for MM risk from a meta-analysis of UKB and FinnGen (case = 1649; control = 727,247) or FinnGen only (case = 1085; control = 271,463). Cis-SNPs associated with protein levels were used to instrument circulating proteins. We evaluated proteins for the consistency of directions of effect across MR analyses (with 95% confidence intervals not overlapping the null) and corroborating evidence from genetic colocalization. In the forward MR, 994 (SomaLogic) and 1570 (Olink) proteins were instrumentable. 440 proteins were analysed in both deCODE and UKB; 302 (69%) of these showed consistent directions of effect in the forward MR. Seven proteins had 95% confidence intervals (CIs) that did not overlap the null in both forward MR analyses and did not have evidence for an effect in the reverse direction: higher levels of dermatopontin (DPT), beta-crystallin B1 (CRYBB1), interleukin-18-binding protein (IL18BP) and vascular endothelial growth factor receptor 2 (KDR) and lower levels of odorant-binding protein 2b (OBP2B), glutamate-cysteine ligase regulatory subunit (GCLM) and gamma-crystallin D (CRYGD) were implicated in increasing MM risk. Evidence from genetic colocalization did not meet our threshold for a shared causal signal between any of these proteins and MM risk (h4 < 0.8). Our results highlight seven circulating proteins which may be involved in MM risk. Although evidence from genetic colocalization suggests these associations may not be robust to the effects of horizontal pleiotropy, these proteins may be useful markers of MM risk. Future work should explore the utility of these proteins in disease prediction or prevention using proteomic data from patients with MM or precursor conditions.

Protein biomarkers of interstitial lung abnormalities in relatives of patients with pulmonary fibrosis.

First-degree relatives of patients with pulmonary fibrosis (relatives) are at high risk for interstitial lung abnormalities (ILA), highlighting the need for biomarkers for risk prediction. We aimed to identify blood proteins associated with and predictive of ILA among relatives of patients with pulmonary fibrosis. Relatives enrolled in two independent cohorts had protein levels measured using an aptamer-based proteomic platform. ILA was assessed with CT scans per Fleischner Society recommendations. Protein associations with ILA were assessed using regression, and significant proteins were used with clinical variables to detect ILA. Of 237 relatives from two independent cohorts, 26% had ILA. Seven proteins were associated with ILA in the discovery cohort after FDR-adjustment, and all remained significant after adjusting for age, gender, and smoking status. Six of seven were significant in the validation cohort including GDF15, SFTPD, and SFTPB. In a multivariable model, six proteins combined with basic demographics in the discovery cohort had AUC=0.92 (0.88 in validation cohort). LASSO modelling identified three proteins and age as predictors with an AUC=0.89. When applied to the combined cohorts, this simple model would reduce the need for CT imaging in one of every three relatives screened. Peripheral blood proteins are associated with ILA in relatives of patients with pulmonary fibrosis and can be used to detect ILA. Our findings demonstrate the potential utility of blood biomarkers in this high-risk group and suggests molecular targets for future investigation.

The Autonomic Nervous System (ANS)-Immune Network in People Living With HIV.

Purpose Pre-clinical studies have demonstrated direct influences of the autonomic nervous system (ANS) on the immune system. However, it remains unknown if connections between the peripheral ANS and immune system exist in humans and contribute to the development of chronic inflammatory disease. This study had three aims: 1.) To examine the relationship between IL-6 and the parasympathetic/vagal component of baroreflex sensitivity (BRS-V) in people with HIV; 2.) To determine if the subtype and severity of HIV-autonomic neuropathy (AN) would predict distinct immunotypes; 3.) To compare the burden of non-AIDS-related co-morbidities between immunotypes. Methods 79 adult people with well-controlled HIV underwent a standard battery of autonomic function tests summarized as the Composite Autonomic Severity Score and vagal and adrenergic baroreflex sensitivity (BRS-V and BRS-A). Levels of immune biomarkers were measured in all participants using the Target 96 Inflammation Panel on the Olink proteomics platform and immunotypes were identified using unbiased, non-negative matrix factorization. Mass cytometry (CyTOF) was completed on a subset of participants with and without autonomic neuropathy (N = 10). Results First, we found reduced BRS-V predicted higher levels of IL-6 (p = 0.002). Second, a pro-inflammatory immunotype defined by elevations in type 1 cytokines (IL-6, IL-17) and increased numbers of CD8 + T-cells was associated with autonomic neuropathy characterized by deficits in sympathetic nervous system activity (aOR = 4.7, p = 0.017). This pro-inflammatory immunotype was older with a greater burden of co-morbidities Conclusion Deficits in the parasympathetic/cardiovagal and the sympathetic nervous system are associated with inflammation and disease burden in people living with HIV. Future longitudinal research is needed to examine causality.

Cardiovascular Risk Biomarkers in Women with and Without Polycystic Ovary Syndrome.

Polycystic ovary syndrome (PCOS) is a prevalent metabolic disorder with an increased risk for cardiovascular disease (CVD) that is enhanced by obesity. This study sought to determine whether a panel of cardiovascular risk proteins (CVRPs) would be dysregulated in overweight/obese PCOS patients, highlighting potential biomarkers for CVD in PCOS. In this exploratory cross-sectional study, plasma levels of 54 CVRPs were analyzed in women with PCOS (n = 147) and controls (n = 97). CVRPs were measured using the SOMAscan proteomic platform (version 3.1), with significant proteins identified through linear models, regression analysis, and receiver operating characteristic (ROC) analysis. Analysis on BMI-matched subsets of the cohort were undertaken. Functional enrichment and protein-protein interaction analyses elucidated the pathways involved. Eleven CVRPs were dysregulated in PCOS (whole set, without matching for body mass index (BMI) or insulin resistance (IR)): leptin, Interleukin-1 receptor antagonist protein (IL-1Ra), polymeric immunoglobulin receptor (PIGR), interleukin-18 receptor (IL-18Ra), C-C motif chemokine 3 (MIP-1a), and angiopoietin-1 (ANGPT1) were upregulated whilst advanced glycosylation end product-specific receptor, soluble (sRAGE), bone morphogenetic protein 6 (BMP6); growth/differentiation factor 2 (GDF2), superoxide dismutase [Mn] mitochondrial (MnSOD), and SLAM family member 5 (SLAF5) were downregulated versus the controls. In BMI-matched (overweight/obese, BMI ≥ 26 kg/m2) subset analysis, six CVRPs were common to the whole set: ANGPT1 and IL-1Ra were upregulated; and sRAGE, BMP6, GDF2, and Mn-SOD were downregulated. In addition, lymphotactin (XCL1) was upregulated and placenta growth factor (PIGF), alpha-L-iduronidase (IDUA), angiopoietin-1 receptor, and soluble (sTie-2) and macrophage metalloelastase (MMP12) were downregulated. A subset analysis of BMI-matched plus insulin resistance (IR)-matched women revealed only upregulation of tissue factor (TF) and renin in PCOS, potentially serving as biomarkers for cardiovascular risk in overweight/obese women with PCOS. A combination of upregulated obesity-related CVRPs (ANGPT1/IL/1Ra/XCL1) and downregulated cardioprotective proteins (sRAGE/BMP6/Mn-SOD/GDF2) in overweight/obese PCOS women may contribute to the increased risk for CVD. TF and renin upregulation observed in the BMI- and IR-matched limited sample PCOS subgroup indicates their potential risk of CVD.

The Plasma Proteome and Risk of Future Venous Thromboembolism-Results from the HUNT Study.

This study aimed to identify novel plasma proteins associated with first-lifetime venous thromboembolism (VTE) and molecular pathways involved in VTE pathogenesis. A case-cohort comprising incident VTE cases (n = 294) and a randomly sampled age- and sex-weighted subcohort (n = 1,066) was derived from the Trøndelag Health Study (HUNT3, n = 50,800). Blood samples were collected and stored at cohort inclusion (2006-2008), and participants were followed up to 5 years. Proteome-wide analyses was performed using the 7k SomaScan® proteomics platform, and weighted Cox-regression models adjusted for age, sex, and sample batch were conducted, with the Bonferroni method applied to account for multiple testing. Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses were applied on the top-ranked 200 proteins associated with VTE. Out of 7,288 human proteins, 7 proteins were significantly associated with higher VTE risk with p-value <6.9 × 10-6 (hazard ratios per 1 standard deviation increase in protein levels ranging from 1.39 to 1.86). Except for coagulation factor VIII and tumor necrosis factor soluble receptor II, these proteins were novel associations and included collagen alpha-3(VI):BPTI/Kunitz inhibitor, histo-blood group ABO system transferase, peroxidasin, human epididymis protein 4, and regulator of G protein signaling 3. KEGG analyses of the top-ranked 200 proteins revealed significant pathway enrichment of nine proteins in the complement (mainly lectin pathway) and coagulation (mainly intrinsic pathway) cascades. Our proteome-wide analysis led to discovery of five novel protein candidates associated with 5-year risk of future VTE. KEGG analyses supported an interplay between the complement and coagulation pathways in the pathogenesis of VTE.

Proteomic Sensors for Quantitative, Multiplexed and Spatial Monitoring of Kinase Signaling.

Understanding kinase action requires precise quantitative measurements of their activity in vivo . In addition, the ability to capture spatial information of kinase activity is crucial to deconvolute complex signaling networks, interrogate multifaceted kinase actions, and assess drug effects or genetic perturbations. Here we developed a proteomic kinase activity sensor platform (ProKAS) for the analysis of kinase signaling using mass spectrometry. ProKAS is based on a tandem array of peptide sensors with amino acid barcodes that allow multiplexed analysis for spatial, kinetic, and screening applications. We engineered a ProKAS module to simultaneously monitor the activities of the DNA damage response kinases ATR, ATM, and CHK1 in response to genotoxic drugs, while also uncovering differences between these signaling responses in the nucleus, cytosol, and replication factories. Furthermore, we developed an in silico approach for the rational design of specific substrate peptides expandable to other kinases. Overall, ProKAS is a novel versatile system for systematically and spatially probing kinase action in cells.

Quantitative Top-down Proteomics Revealed Kinase Inhibitor-Induced Proteoform-Level Changes in Cancer Cells.

Quantitative analysis of proteins and their post-translational modifications (PTMs) in complex biological samples is critical to understanding cellular biology as well as disease detection and treatment. Top-down proteomics methods provide a "bird's eye" view of the proteome by directly detecting and quantifying intact proteoforms. Here, we developed a high-throughput quantitative top-down proteomics platform to probe intact proteoform and phosphoproteoform abundance changes in HeLa cells as a result of treatment with staurosporine (STS), a broad-spectrum kinase inhibitor. In total, we identified and quantified 1187 proteoforms from 215 proteoform families. Among them, 55 proteoforms from 37 proteoform families were significantly changed upon STS treatment. These proteoforms were primarily related to catabolic, metabolic, and apoptotic pathways that are expected to be impacted as a result of kinase inhibition. In addition, we manually evaluated 25 proteoform families that expressed one or more phosphorylated proteoforms. We observed that phosphorylated proteoforms in the same proteoform family, such as eukaryotic initiation factor 4E binding protein 1 (4EBP1), were differentially regulated relative to the unphosphorylated proteoforms. Combining relative profiling of proteoforms within these proteoform families with individual proteoform profiling results in a more comprehensive picture of STS treatment-induced proteoform abundance changes that cannot be achieved using bottom-up methods.

New proteomic signatures in Alzheimer Disease identification and Mild Cognitive Impairment Classification

AbstractBackgroundEmerging technologies and novel biomarker tools are transforming the field of Alzheimer’s disease, allowing for a more in‐depth exploration of biological mechanisms underpinning the disease aetio‐pathogenesis. In this context, there is growing recognition of the potential of plasma proteomics for AD risk assessment and disease characterization. On the other hand, differences between proteomics platforms introduce uncertainties regarding cross‐platform applicability.MethodElastic net analysis was applied to 190 analytes measured by the Luminex xMAP platform in plasma from 566 participants [58 cognitively normal (CN), 396 with mild cognitive impairment (MCI) and 58 with dementia due to AD], from the Alzheimer’s Disease Neuroimaging Initiative (ADNI). Validation was conducted on 1303 participants (113 CN, 795 MCI, 395 AD) from the Spanish ACE study, which was analysed by the SOMAscan 7k proteomic platform. MCI participants were classified into MCI stable or decliners according to their follow‐up diagnosis (ADNI: 158 vs. 217, Ace: 449 vs. 346). Signatures derived from our elastic net analysis and other methods applied in prior ADNI proteomic signature studies were compared, with our approach aimed at reducing dimensionality of signatures of AD vs. CN or MCI subgroup classification.ResultWe identified a signature including 11 plasma analytes and sex, age, and APOEε4 gene status, for distinguishing AD from CN, with 94% accuracy on ADNI (sensitivity 88%, specificity 94%) and 95% on Ace (specificity 96%, sensitivity 93%). Another 19‐analyte signature with the above confounders was used for classifying MCI stable and decliners, with a 77% accuracy on ADNI (sensitivity 85%, specificity 40%) and 51% on Ace (sensitivity 68%, specificity 36%). These two signatures share one common analyte, Vitronectin, a protein found to appear in degenerative diseases related abnormal deposits, such as amyloid.ConclusionCompared with prior proteomic signature studies, all based on ADNI, our findings attained higher specificity and sensitivity, while utilizing a smaller panel for AD classification. They also confirm the reliability and consistency of a group of plasma proteomic biomarkers across two different platforms.

Large‐scale proteomic and genomic analysis identify plasma proteins influencing human brain structure and Alzheimer’s disease risk

AbstractBackgroundRecent advances in proteomics technology enable us to comprehensively examine the impact of the proteome on Alzheimer’s disease (AD) and intermediate phenotypes, such as brain imaging measures. Here, by leveraging large‐scale proteomic, brain imaging, and AD genome‐wide association studies (GWAS), we performed Mendelian randomization (MR) and colocalization analysis to investigate the impact of plasma protein levels on brain imaging measures and the link between plasma protein levels, brain imaging, and AD.MethodWe leverage the protein quantitative trait loci (pQTL) datasets from the UK Biobank Pharma Proteomics Project (UKB‐PPP; 2,625 proteins by Olink assay) and deCODE genetics (Ferkingstad et al. 2021; 4,472 proteins by SomaLogic assay). UKB‐PPP and the deCODE study represent the largest plasma proteomics GWAS to date based on two different proteomics platforms, where the total sample sizes are over 30,000 for each. For the brain imaging and AD genetic datasets, we used the brain imaging GWAS from UKB (453 T1 MRI and DTI imaging; Smith et al. 2021), and AD GWAS from Bellenguez et al. 2022.ResultAfter extensive data processing and quality control, we performed two‐sample MR and identified significant associations in 170 protein‐DTI imaging pairs and 38 protein‐MRI pairs for 60 unique proteins from UKB‐PPP and 41 unique proteins from deCODE. Colocalization analysis further prioritized proteins related to AD implicated in previous studies. We further extended the association between plasma protein and brain imaging to AD by performing MR and colocalization analyses using pQTL and AD GWAS. As an example, we showed that plasma progranulin (P28799) level showed a significant effect on AD risk (P = 6.79×10−12), while supported by associations between plasma progranulin level and caudate volumes (right and left T1 DTI, P = 8.41×10−6).ConclusionIn summary, we presented the largest to date study for the impact of plasma protein levels on brain imaging measures and a novel approach to identify biomarkers for AD. Overall, the integration of large‐scale MRI, proteomics, AD phenotype, and genome‐wide data holds great promise for enhancing our understanding of AD, and ultimately facilitates biomarker discovery and therapeutic development.

Proteogenomics in cerebrospinal fluid and plasma reveals new biological fingerprint of cerebral small vessel disease

AbstractBackgroundCerebral small vessel disease (cSVD) is a leading cause of stroke and dementia. Its underlying mechanisms remain elusive and specific mechanism‐based drugs are lacking.MethodWe integrated more than 2,800 CSF and 4,600 plasma pQTL, derived from the largest proteomic studies so far (SOMAscan 7k and 4k; in up to 35,559 individuals), and the two most prevalent MRI‐markers of cSVD (MRI‐cSVD, white matter hyperintensities and perivascular spaces burden; in up to 48,454 individuals) in a Mendelian Randomization (MR) framework to identify causal and druggable targets for cSVD. Identified association were followed‐up using a multipronged approach: across fluids, proteomics platforms (Olink 3072, N=8,590) and lifespan (N=1,748), using both MR and individual‐level data.ResultWe found 51 proteins associated with MRI‐cSVD of which 46 in CSF and 9 in plasma. Among available significant CSF‐ and plasma‐proteins, 32% and 31% replicated in cross‐fluid and cross‐platform follow‐up, and 47% were associated with stroke and/or dementia at least at nominal significance. We found converging evidence that protein‐cSVD associations are enriched in extracellular matrix and immune response pathways. Immunity‐related proteins already showed association with MRI‐cSVD already in young adults in their twenties. Furthermore, we provide genetic support for drug repositioning opportunities for cSVD, including compounds crossing the blood brain barrier.ConclusionTogether, these findings provide a novel proteogenomic signature of cSVD and pave the way for novel therapeutic developments.

Plasma Proteomic Determinants of Small Vessel Disease of the Brain: the Multi‐Ethnic Study of Atherosclerosis

AbstractBackgroundThe identification of novel blood‐based biomarkers of small vessel disease of the brain (SVD) may improve pathophysiologic understanding and inform the development of new therapeutic strategies for prevention. We evaluated plasma proteomic associations of white matter fractional anisotropy (WMFA), white matter hyperintensity (WMH) volume, enlarged perivascular space (ePVS) volume, and the presence of microbleeds (MB) on brain magnetic resonance imaging (MRI) in the population‐based Multi‐Ethnic Study of Atherosclerosis (MESA).MethodsEligible MESA participants had 2941 plasma proteins measured from stored blood samples (collected in 2016‐2018) using the antibody‐based Olink proteomics platform, and completed brain MRI scans in 2018‐2019. Participants with quality control exclusion of protein measurements, missing covariate data, and poor quality or missing MRI outcome variables were excluded. The cross‐sectional association between the abundance of each plasma protein (normalized protein expression – a relative protein quantification unit measured on a log2 scale) was modeled separately with WMFA, WMH volume, total ePVS volume, and the presence of MBs using multivariable linear or modified Poisson regression, adjusting for demographic variables, estimated glomerular filtration rate (eGFR), and SVD risk factors. The Benjamini‐Hochberg procedure was used to control the false discovery rate (FDR) &lt;0.05 to account for multiple hypothesis testing. For proteins independently associated with the SVD markers on MRI, penalized regression with least absolute shrinkage and selection operator (LASSO) was used to create a parsimonious proteomic model.ResultsEligible participants (total N=709) had a mean age of 73 years, 53% were women, 25% were Black, 17% were Chinese, 19% were Hispanic or Latino, and 39% were White (Table 1). After adjustment for demographics, eGFR, and SVD risk factors, 769 plasma proteins were associated with WMFA (FDR &lt;0.05) (Figure). LASSO regression identified a 37‐protein model predictive of WMFA (Table 2). We did not find plasma proteins to be independently associated with WMH volume, ePVS volume, or the presence of MBs.ConclusionMultiple circulating proteins – implicated in central nervous system myelination, lipid metabolism, angiogenesis, coagulation, cellular adhesion and migration, appetite regulation, energy homeostasis, systemic inflammation, and immune regulation – were independently associated with WMFA in a multi‐ethnic cohort of older adults.

Lipopolysaccharide Imprinted Polymers for Specific Recognition of Bacterial Outer Membrane Vesicles.

Outer membrane vesicles (OMVs) secreted by bacteria are emerging diagnostic markers for bacterial infection or disease detection due to their carriage of various signaling molecules. However, actual biological samples of patients are extremely complex, and applying OMVs to clinical diagnosis remains a major challenge. One of the major challenges is that there are still great difficulties in the enrichment of OMVs including tedious steps and lower concentration. And some commonly used exosome enrichment methods, such as ultracentrifugation, still have some shortcomings. Herein, we introduce lipopolysaccharide (LPS) molecularly imprinted polymer (MIP) for efficient capturing and analyzing OMVs, enabling a novel approach to bacterial disease diagnosis based on biorecognition materials. LPS, as a unique structure of Gram-negative bacteria, also widely expressed on the surface of OMVs, which will form cyclic hydrogen bonds with functional monomers of MIP with affinity interactions. The prepared MIP efficiently can isolate OMVs from 100 μL of culture broth via specific affinity LPS in less than 40 min with a recovery rate of over 95%. Moreover, MIP exhibits good reusability, with almost identical enrichment performance after 5 repeated cycles, contributing to reducing experimental costs in both time and economy. The captured OMVs can be detected using Western blotting with target protein antibodies or in combination with proteomic analysis, providing a proteomic biomarker platform for early bacteria disease diagnosis.

Circulating inflammatory proteins are elevated in type 1 and type 2 diabetes and associated to complications.

The presence of low-grade inflammation has been reported in people with type 2 diabetes and related to the development of (macro)vascular complications. Whether systemic inflammation is present in type 1 diabetes and linked to long-term complications remains unknown. We used a targeted proteomics approach to compare inflammation in people with type 1 diabetes and type 2 diabetes with control subjects and linked these proteins to diabetes related characteristics and complications. We included 233 participants with type 1 diabetes, 387 participants with type 2 diabetes and 150 healthy controls. Plasma was collected and used to determine high sensitive C-reactive proteins (hs-CRP) and an additional 92 inflammatory proteins using the Olink proteomics platform. Compared to healthy controls, 41 circulating inflammatory proteins were higher in type 1 diabetes (FDR < 0.05) and 64 inflammatory proteins in type 2 diabetes (FDR < 0.05) (including CXCL5, IL-15RA, MCP-4 and AXIN1 for both groups). HbA1c levels were positively associated with 21 inflammatory proteins (including CDCP1, FGF-21, HGF and IL-18R1) in type 1 diabetes (FDR < 0.05), whereas a positive association existed between body mass index (BMI) and 26 inflammatory proteins (including IL6, IL17C, FGF-23 and CSF-1) in type 2 diabetes. Inflammatory proteins associated with the presences, of complications, particularly nephropathy, were similar in both type 1 and type 2 diabetes. FlT3L and EN-RAGE were associated with the development of cardiovascular disease (CVD) in type 2 diabetes. Both type 1 diabetes and type 2 diabetes are associated with increased circulating inflammatory protein concentrations, but the increase is more pronounced in type 2 diabetes. These results suggest both differences in drivers of inflammation between type 1 diabetes and type 2 diabetes as well as potential similarities in pathways involved in the development of diabetes-associated complications.

Abstract C028: Decreased heterogeneity in immune interactions in oral cavity squamous cell carcinoma patients with recurrence

Abstract There have been improvements in clinical risk stratifications for oral cavity squamous cell carcinoma (OSCC) with the AJCC 8th edition staging system. Other studies suggest the addition of a metric of specific immune interactions in and around the tumor may be helpful to further stratify patients, with mixed findings. The advent of spatial imaging has allowed a global view for assessing tumor features in the context of all surrounding cell interactions. We hypothesize that the lack of architectural information could contribute to varied clinical outcomes in previous studies, such as tumor recurrence. This highlights how spatial technologies could aid in understanding the role of tissue architecture and its influence on biology and medicine. To study tissue remodeling spatially, we focused on multiplexed-ion beam imaging (MIBI), a spatial proteomics platform to visualize 40 immuno-onco proteins simultaneously. Our banked patient cohort consists of 20 recurrent and 49 non-recurrent OSCC primary surgery samples. Clinical metadata capturing pathologic stage, alcohol use, smoking history, survival data, and more were also collected. With our pathologist, we curated 290 field-of-views (FOVs) of matched primary tumors, uninvolved lymph nodes, and metastatic lymph nodes (if any). Downstream computational analysis using existing MIBI codebases augmented with in-house spatial analytics has revealed a robust spatial dataset. After using machine-learning algorithms to cell segment our images, we validated these assigned cells and marker staining to further annotate cell populations. When we analyzed direct cell interactions, there was a decrease in heterogeneity in immune interactions within recurrent patients compared to non-recurrent patients. To ascertain how the loss of these interactions affected recurrence prognosis, we investigated their prevalence within the tumor border vicinity. We hypothesize that the proximity and cell-cell interactions near the border will promote a tumor-suppressive environment. Utilizing centroid-to-centroid distance measurements, we distilled the top 30 cell-cell interactions within a 100-micron distance from the tumor border. Cell pairs involving macrophage tissue-resident memory cells and CD8 T memory resident cells were more abundant in recurrent patients, while pairs involving CD4 T regulatory and CD8 cells were found in non-recurrent patients. Reassuringly, cell pairings involving myeloid cells were upregulated in the recurrent cohort as expected. Interestingly, cancer-associated fibroblast cell pairings were upregulated indiscriminately in both patient sub- cohorts. This suggests that the impact of certain known immunosuppressive changes may contribute in varying amounts to disease control. Our data demonstrates differences in the architecture of tumors with known divergent outcomes, which suggests global tissue-level changes could be a predictive factor. Clinically, interrogating these interactions of interest could guide treatment escalation or de-escalation in the future. Citation Format: Connie J. Zhou, Jane Lee, Rebecca Jaszczak, Steven R. Long, Patrick K. Ha, Matthew H. Spitzer. Decreased heterogeneity in immune interactions in oral cavity squamous cell carcinoma patients with recurrence [abstract]. In: Proceedings of the AACR Special Conference in Cancer Research: Tumor-body Interactions: The Roles of Micro- and Macroenvironment in Cancer; 2024 Nov 17-20; Boston, MA. Philadelphia (PA): AACR; Cancer Res 2024;84(22_Suppl):Abstract nr C028.

Decoding Spatial Tissue Architecture: A Scalable Bayesian Topic Model for Multiplexed Imaging Analysis.

Recent progress in multiplexed tissue imaging is advancing the study of tumor microenvironments to enhance our understanding of treatment response and disease progression. Cellular neighborhood analysis is a popular computational approach for these complex image data. Despite its popularity, there are significant challenges, including high computational demands that limit feasibility for largescale applications and the lack of a principled strategy for integrative analysis across images. This absence hampers the precise and consistent identification of spatial features and tracking of their dynamics over disease progression. To overcome these challenges, we introduce SpatialTopic , a spatial topic model designed to decode high-level spatial architecture across multiplexed tissue images. SpatialTopic integrates both cell type and spatial information within a topic modelling framework, originally developed for natural language processing and adapted for computer vision. Spatial information is incorporated into the flexible design of documents, representing densely overlapping regions in images. We employ an efficient collapsed Gibbs sampling algorithm for model inference. We benchmarked the performance against five state-of-the-art algorithms through various case studies using different single-cell spatial transcriptomic and proteomic imaging platforms across different tissue types. We show that SpatialTopic is highly scalable on large-scale image datasets with millions of cells, along with high precision and interpretability. Our findings demonstrate that SpatialTopic consistently identifies biologically and clinically significant spatial "topics" such as tertiary lymphoid structures (TLSs) and tracks dynamic changes in spatial features over disease progression. Its computational efficiency and broad applicability across various molecular imaging platforms will enhance the analysis of large-scale tissue imaging datasets.

Protein Biomarkers of Adverse Clinical Features and Events in Sarcomeric Hypertrophic Cardiomyopathy.

Hypertrophic cardiomyopathy (HCM) is a heterogeneous condition that can lead to atrial fibrillation, heart failure, and sudden cardiac death in many individuals but mild clinical impact in others. The mechanisms underlying this phenotypic heterogeneity are not well defined. The aim of this study was to use plasma proteomic profiling to help illuminate biomarkers that reflect or inform the heterogeneity observed in HCM. The Olink antibody-based proteomic platform was used to measure plasma proteins in patients with genotype positive (sarcomeric) HCM participating in the HCM Registry. We assessed associations between plasma protein levels with clinical features, cardiac magnetic resonance imaging metrics, and the development of atrial fibrillation. We measured 275 proteins in 701 patients with sarcomeric HCM. There were associations between late gadolinium enhancement with proteins reflecting neurohormonal activation (NT-proBNP [N-terminal pro-B-type natriuretic peptide] and ACE2 [angiotensin-converting enzyme 2]). Metrics of left ventricular remodeling had novel associations with proteins involved in vascular development and homeostasis (vascular endothelial growth factor-D and TM [thrombomodulin]). Assessing clinical features, the European Society of Cardiology sudden cardiac death risk score was inversely associated with SCF (stem cell factor). Incident atrial fibrillation was associated with mediators of inflammation and fibrosis (MMP2 [matrix metalloproteinase 2] and SPON1 [spondin 1]). Proteomic profiling of sarcomeric HCM identified biomarkers associated with adverse imaging and clinical phenotypes. These circulating proteins are part of both established pathways, including neurohormonal activation and fibrosis, and less familiar pathways, including endothelial function and inflammatory proteins less well characterized in HCM. These findings highlight the value of plasma profiling to identify biomarkers of risk and to gain further insights into the pathophysiology of HCM.

Classification of Predictors of Rapid Development of Kidney Failure and Short-Term Changes in Concentration of Circulating Proteins.

Limited knowledge exists regarding short-term changes/increases in concentrations of circulating proteins (referred here as deltas) and rapid development of kidney failure (rapid KF) in diabetes mellitus. Concentrations of 452 circulating proteins were measured by OLINK proteomics platform at baseline and after a median interval of 3-4 years in 106 individuals with type 1 and 77 with type 2 diabetes in two case-control studies. During 10-year follow-up, 31 and 26 individuals, respectively, developed rapid KF. Deltas for 40 proteins predicted rapid KF in both studies. All were better predictors than delta urine albumin-creatinine ratio, and half were better than delta glomerular filtration rate. Comparing the delta proteins with 46 circulating proteins of which elevated baseline concentrations were predictors of rapid KF risk in our previous study, 61 unique proteins were identified. Among these proteins, 21 were good predictors of rapid KF only when measured at baseline (predictors of initiation), 15 were good predictors when measured as deltas (predictors of progression) and 25 were good predictors when both baseline and delta concentrations were used (predictors of initiation and progression). An index score, developed for the latter 25 proteins, provided superior prediction of rapid KF. A subset of these latter proteins was associated with apoptotic processes/tumor necrosis factor (TNF) receptor signaling pathways. Development of rapid KF in diabetes was preceded by elevated concentrations of multiple circulating proteins both at baseline and during short follow-up. Comparing baseline and short-term changes in concentrations of circulating proteins classified predictors of rapid KF risk into those associated with initiation, progression, or both. Predictors of both initiation & progression flagged apoptosis processes and TNF receptor signaling pathways. Multi-protein prognostic algorithms using proteins associated with both initiation and progression improved prediction of rapid KF risk beyond clinical variables.

Circulating proteins linked to apoptosis processes and fast development of end-stage kidney disease in diabetes.

Many circulating proteins are associated with risk of ESKD, but their source and the biological pathways/disease processes they represent are unclear. Using OLINK proteomics platform, concentrations of 455 proteins were measured in plasma specimens obtained at baseline from 399 individuals with diabetes. Elevated concentrations of 46 circulating proteins were associated (P < 1 × 10-5) with development of ESKD (n = 143) during 7-15 years of follow-up. Twenty of these proteins enriched apoptosis/TNF receptor signaling pathways. A subset of 20 proteins (5-7 proteins), summarized as an apoptosis score, together with clinical variables accurately predicted risk of ESKD. Expression of genes encoding the 46 proteins in peripheral WBCs showed no difference between cells from individuals who did or did not develop ESKD. In contrast, plasma concentration of many of the 46 proteins differed by this outcome. In single-nucleus RNA-Seq analysis of kidney biopsies, the majority of genes encoding for the 20 apoptosis/TNF receptor proteins were overexpressed in injured versus healthy proximal tubule cells. Expression of these 20 genes also correlated with the overall index of apoptosis in these cells. Elevated levels of circulating proteins flagging apoptotic processes/TNF receptor signaling pathways - and likely originating from kidney cells, including injured/apoptotic proximal tubular cells - preceded the development of ESKD.

Protein biomarker signatures of preeclampsia - a longitudinal 5000-multiplex proteomics study

We aimed to explore novel biomarker candidates and biomarker signatures of late-onset preeclampsia (LOPE) by profiling samples collected in a longitudinal discovery cohort with a high-throughput proteomics platform. Using the Somalogic 5000-plex platform, we analyzed proteins in plasma samples collected at three visits (gestational weeks (GW) 12–19, 20–26 and 28–34 in 35 women with LOPE (birth ≥ 34 GW) and 70 healthy pregnant women). To identify biomarker signatures, we combined Elastic Net with Stability Selection for stable variable selection and validated their predictive performance in a validation cohort. The biomarker signature with the highest predictive performance (AUC 0.88 (95% CI 0.85–0.97)) was identified in the last trimester of pregnancy (GW 28–34) and included the Fatty acid amid hydrolase 2 (FAAH2), HtrA serine peptidase 1 (HTRA1) and Interleukin-17 receptor C (IL17RC) together with sFLT1 and maternal age, BMI and nulliparity. Our biomarker signature showed increased or similar predictive performance to the sFLT1/PGF-ratio within our data set, and we were able to validate the biomarker signature in a validation cohort (AUC ≥ 0.90). Further validation of these candidates should be performed using another protein quantification platform in an independent cohort where the negative and positive predictive values can be validly calculated.

Plasma proteome-wide analysis of cerebral small vessel disease identifies novel biomarkers and disease pathways.

Cerebral small vessel disease (SVD), as defined by neuroimaging characteristics such as white matter hyperintensities (WMHs), cerebral microhemorrhages (CMHs), and lacunar infarcts, is highly prevalent and has been associated with dementia risk and other clinical sequelae. Although conditions such as hypertension are known to contribute to SVD, little is known about the diverse set of subclinical biological processes and molecular mediators that may also influence the development and progression of SVD. To better understand the mechanisms underlying SVD and to identify novel SVD biomarkers, we used a large-scale proteomic platform to relate 4,877 plasma proteins to MRI-defined SVD characteristics within 1,508 participants of the Atherosclerosis Risk in Communities (ARIC) Study cohort. Our proteome-wide analysis of older adults (mean age: 76) identified 13 WMH-associated plasma proteins involved in synaptic function, endothelial integrity, and angiogenesis, two of which remained associated with late-life WMH volume when measured nearly 20 years earlier, during midlife. We replicated the relationship between 9 candidate proteins and WMH volume in one or more external cohorts; we found that 11 of the 13 proteins were associated with risk for future dementia; and we leveraged publicly available proteomic data from brain tissue to demonstrate that a subset of WMH-associated proteins was differentially expressed in the context of cerebral atherosclerosis, pathologically-defined Alzheimer's disease, and cognitive decline. Bidirectional two-sample Mendelian randomization analyses examined the causal relationships between candidate proteins and WMH volume, while pathway and network analyses identified discrete biological processes (lipid/cholesterol metabolism, NF-kB signaling, hemostasis) associated with distinct forms of SVD. Finally, we synthesized these findings to identify two plasma proteins, oligodendrocyte myelin glycoprotein (OMG) and neuronal pentraxin receptor (NPTXR), as top candidate biomarkers for elevated WMH volume and its clinical manifestations.