2  ▏Nature Communications

      Circulating causal protein networks linked to future risk of myocardial infarction

      Multi-omic landscape of human gliomas from diagnosis to treatment and recurrence

      Mass spectrometry-based profiling of single-cell histone post-translational modifications to dissect chromatin heterogeneity

      Plasma proteomics to explore benefit from anti-PD-(L)1 in advanced solid tumors: a prospective pilot study

      Single cell proteomic analysis defines discrete neutrophil functional states in human glioblastoma

7  ▏Nature Communication

      Effects of an anti-lipogenic low-carbohydrate high polyunsaturated fat diet or a healthy Nordic diet versus usual care on liver fat and cardiometabolic disorders in type 2 diabetes or prediabetes: a randomized controlled trial (NAFLDiet)

Reliable tools for the early identification of Crohn’s disease remain scarce. A multicenter team led by Guangdong Provincial People’s Hospital analyzed 2,736 plasma proteins in 39,634 participants from the UK Biobank and identified 44 proteins associated with the onset of Crohn’s disease. Among them, CD274, CHI3L1, REG1B, ITGAV, PRSS8, ITGA11, GDF15, DEFA1_DEFA1B and IL6 ranked highest in protein importance. A machine learning model built on these nine proteins achieved high predictive performance for Crohn’s disease in a geographically distinct UKB test cohort (n = 13,262; AUC 0.76), outperforming the clinical risk model. The model was externally validated in the EPIC-Norfolk cohort (n = 2,944; AUC 0.73) and showed high discriminative ability in a cross-sectional cohort from southern China (n = 74; AUC 0.79). In the UKB test cohort, combining protein data with clinical data raised predictive performance to AUC 0.78, with predictions reaching as far back as 16 years before diagnosis. Within the same cohort, individuals classified as high risk by the protein model were 4.23 times more likely to develop Crohn’s disease than those classified as low risk. These findings highlight proteomics-based models as a promising approach for predicting Crohn’s disease up to 16 years before diagnosis, opening the door to early screening and intervention.

▼

Differences in blood protein levels have been linked to a range of complex diseases, including cardiovascular disease. These associations reflect a complex interplay of local and systemic factors in the development of cardiovascular disease and underscore the need for a comprehensive, systems-level understanding of its etiology. A multicenter team including the University of Iceland developed a causal network inference framework using data from one of the largest serum proteomic studies to date, comprising measurements of 7,523 serum proteins in the prospective, population-based Age, Gene/Environment Susceptibility–Reykjavik Study (AGES) cohort of 5,376 older adults. Using cis-acting protein quantitative trait loci (pQTLs) as instrumental variables within a causal inference framework—an approach designed to reduce hidden confounding—the team identified 185 high-confidence causal serum protein subnetworks acting on 5,611 targets. Several of these subnetworks, many of which form a hierarchical framework of directional relationships, were significantly associated with multiple cardiometabolic traits and with the future risk of myocardial infarction and its long-term complication, heart failure.

▼

Scan the QR code to read the full article

Multi-omic landscape of human gliomas from diagnosis to treatment and recurrence

Gliomas are among the most lethal cancers, with limited treatment options. To uncover the hallmarks of therapeutic escape and the landscape of the tumor microenvironment (TME), the research team of Sean C. Bendall at Stanford University School of Medicine applied spatial proteomic, transcriptomic and glycomic analyses to 670 lesions from 310 adult and pediatric patients. Single-cell analysis showed that B7H3-positive tumor cells are widespread in glioblastoma (GBM) and pleomorphic xanthoastrocytoma; however, in most gliomas—including pediatric cases—targetable tumor antigens were expressed in fewer than 50% of tumor cells, which may explain the failure of some clinical trials. Analysis of paired isocitrate dehydrogenase (IDH)-mutant glioma samples revealed that recurrence is driven by tumor–immune spatial reorganization, shifting from niches rich in T cells and vascular-associated myeloid cells to tumors dominated by microglia and CD206+ macrophages. Integrated multi-omics analysis identified N-glycosylation as the best classifier for distinguishing tumor grade, while the immune transcriptome best predicted survival in glioblastoma. Together, this work provides a framework and baseline for targeted therapy, classification, prognostic prediction and understanding of TME composition across disease stages.

▼

Single-cell proteomics can reliably quantify cellular heterogeneity, yet quantitative analysis of post-translational modifications—such as those on histones—has remained elusive. A multicenter collaboration including the Broad Institute of MIT and Harvard developed a mass spectrometry-based method for profiling single-cell histone post-translational modifications (sc-hPTM). The method identifies single and combinatorial histone marks in an unbiased manner (67 peptide forms in total), covering nearly all frequently studied histone modifications, with reproducibility comparable to conventional bulk experiments. Treating cells with the histone deacetylase inhibitor sodium butyrate, the method was able to (i) distinguish treated from untreated cells; (ii) identify cell subpopulations responding heterogeneously to treatment; and (iii) reveal differential co-regulation of histone modifications in the context of drug treatment. The sc-hPTM method makes it possible to study chromatin heterogeneity comprehensively at single-cell resolution and advances our understanding of the "histone code."

▼

Scan the QR code to read the full article

Anti-PD-(L)1 therapy is now widely used across many cancers, yet only a small proportion of patients derive benefit, and tools for patient selection remain inadequate. This study asked whether plasma proteomics could be used to explore the therapeutic effect of anti-PD-(L)1 antibodies. The team led by Jussi Koivunen at the University of Oulu and Oulu University Hospital in Finland enrolled 56 patients with various solid tumors receiving anti-PD-(L)1 therapy and performed proteomic profiling of pre-treatment plasma samples using a proximity extension assay panel covering 92 immuno-oncology-related biomarkers. Associations between protein expression and radiological outcome and survival were examined, and models predicting objective response rate, disease control rate and overall survival were evaluated. All analyses were conducted within a pooled exploratory framework. Several proteins, including ADGRG1 and ARG1, were associated with radiological response; ARG1 yielded areas under the ROC curve of 0.68 (ORR) and 0.53 (DCR). Machine learning models were constructed to predict radiological response and long-term survival. For overall survival, a six-protein Cox proportional hazards model discriminated risk groups (hazard ratio 4.8; 95% CI 2.1–10.9), while larger protein panels improved discrimination but were prone to overfitting and require further validation. Unsupervised clustering showed no dominant separation by tumor type. Findings from this small, prospective, heterogeneous cohort indicate that plasma proteomics shows promising but preliminary results in predicting benefit from anti-PD-(L)1 therapy across patients with advanced solid tumors. The DCR and ORR model findings should be considered exploratory and require external validation at both tumor-specific and pan-cancer levels before clinical application.

▼

Scan the QR code to read the full article

Single cell proteomic analysis defines discrete neutrophil functional states in human glioblastoma

Neutrophils are essential innate immune cells that have been shown to infiltrate glioblastoma, yet the molecular understanding of their functional states within the tumor niche remains incomplete. Because neutrophils display marked discordance between mRNA and protein abundance, a multicenter collaboration including the Centre for Inflammation Research at the University of Edinburgh's Institute for Regeneration and Repair and the Vienna Biocenter of the Austrian Academy of Sciences developed an ultrasensitive micro-scale bulk and single-cell proteomics (SCP) workflow to investigate the heterogeneity of neutrophils in peripheral blood and tumor-associated neutrophils (TANs) from patients with glioblastoma. Micro-scale population analysis achieved deeper protein coverage of circulating immature and mature neutrophils and TAN populations, defined signatures of maturity, and showed that TANs resemble mature circulating neutrophils. Analysis of the SCP data enabled detection of more than 1,100 proteins from individual TANs, allowing detailed characterization of neutrophil subsets in glioblastoma. This approach revealed pathogenic (pro-tumor) and anti-tumor cell clusters and uncovered cell states that single-cell RNA sequencing (scRNAseq) was unable to detect, opening new opportunities for selectively targeting pro-tumor neutrophil states.

▼

Scan the QR code to read the full article

Effects of an anti-lipogenic low-carbohydrate high polyunsaturated fat diet or a healthy Nordic diet versus usual care on liver fat and cardiometabolic disorders in type 2 diabetes or prediabetes: a randomized controlled trial (NAFLDiet)

Polyunsaturated fatty acids (PUFA) have been shown to reduce liver fat compared with saturated fat, but the effects of a novel "anti-lipogenic" diet in which PUFA replaces carbohydrates (LCPUFA), or of a low-fat, whole-grain-rich healthy Nordic diet (HND), remain unclear. This study investigated the effects of 12 months of these dietary interventions on liver fat (the primary outcome) and related glycemic and lipid disorders in individuals with prediabetes or type 2 diabetes (T2D), compared with usual care (UC). A three-arm, parallel, ad libitum randomized trial was completed in December 2022. Outcome assessors and care providers were blinded to participants' dietary allocation. A collaborative research team from Uppsala University in Sweden, conducting the study at Uppsala University Hospital, randomized 150 men and women with prediabetes or T2D (55%) in a 1:1:1 ratio, stratified by sex and T2D status, and used general linear models to estimate intention-to-treat effects. Compared with the UC group (n = 43), both the LCPUFA group (n = 54) and the HND group (n = 51) reduced liver fat, by −1.46% (95% CI −2.42 to −0.51) and −1.76% (95% CI −2.96 to −0.57), respectively. No difference was observed between LCPUFA and HND in reducing liver fat. HND was more effective than the other diets in reducing body weight and glycated hemoglobin, whereas no difference was observed between LCPUFA and UC on these measures. Compared with UC, both the HND and LCPUFA diets similarly reduced low-density lipoprotein cholesterol, but only HND lowered triglycerides, inflammatory markers and liver enzymes. A total of four serious adverse events occurred, distributed across the groups. Compared with usual care, a plant-based, ad libitum LCPUFA diet and HND were equally effective in reducing liver fat and LDL cholesterol. Even without deliberate energy restriction, HND additionally improved body weight, glycemic control, liver biochemistry, triglycerides and inflammation, suggesting that HND is a clinically feasible dietary approach for managing T2D and metabolic dysfunction-associated steatotic liver disease (MASLD).

▼

Scan the QR code to read the full article