Herbert, B. R. | Search bioRxiv

bioRxiv2026

De-N-glycosylation of in vivo and in vitro adipogenic stem cell products unmasks differential expression of CD36 glycoprotein in human adipogenesis

Adipogenesis is the process of adipose-derived stem cells (ADSCs) responding to extracellular signals from the stem cell niche to differentiate into adipocytes (fat cells) and may be studied in vitro using a cocktail of chemicals that promote adipogenic differentiation to produce differentiated ADSCs (dADSCs). The global membrane N- and O-glycosylation changes of this process have been previously analysed and compared to native adipocytes as a benchmark for a true adipocyte profile, and revealed that bisecting GlcNAc type N-glycans are characteristic of adipogenesis. As stem cell differentiation has been widely reported to result in cellular protein changes, the same cells (ADSCs, dADSCs and mature adipocytes) were characterised for their membrane proteome here using label-free quantitative shotgun proteomics analysis. The membrane proteome displayed more differences in protein numbers between the cell types compared to the previously reported N-glycome which had shown high identical glycomes between stem cells and in vitro dADSCs, suggesting that the proteome is more dynamic during in vitro adipogenesis. Following the global shotgun proteomics analysis, a more targeted approach of carrying out proteomic analysis of de-N-glycosylated peptides of gel-separated proteins unearthed new glycoproteins not detected in the shotgun proteomic analysis. This approach identified the adipogenic marker, CD36, to be under-represented in the shotgun proteome analysis, but as the dominant (glyco)protein in the adipocyte membrane proteome that was also up-regulated at the mRNA transcript level in both the in vitro differentiated ADSCs (7.1-fold increase) and mature adipocytes (102.9-fold increase). A comparison of CD36 sequence coverage in the global shotgun analysis with the de-N-glycosylated CD36 revealed a 41% increase when N-glycans were removed prior to trypsin digestion, explaining its observed increased abundance and highlights the crucial need for de-N-glycosylation of proteins in proteomics experiments for increased identification of glycoproteins. The systems glycobiology approach by the integration of previously reported glycomics data and the proteomics and transcriptomics analyses in this work extended the investigation of membrane protein glycosylation changes in adipose-derived stem cell differentiation. The work provides a framework for future glycoproteomics-based investigations into the differentiation of stem cells into adipocytes, and will allow their related pathologies and potential therapeutic applications to be discovered. GRAPHICAL ABSTRACT O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=121 SRC="FIGDIR/small/722121v1_ufig1.gif" ALT="Figure 1"> View larger version (44K): org.highwire.dtl.DTLVardef@189a786org.highwire.dtl.DTLVardef@5563b8org.highwire.dtl.DTLVardef@5cb5borg.highwire.dtl.DTLVardef@69e11f_HPS_FORMAT_FIGEXP M_FIG C_FIG

cell biology↗

bioRxiv2024

Assessment of a high-throughput mass spectrometry method to accelerate biomarker discovery in clinical cancer cohorts using volumetric absorptive microsampling (VAMS) devices.

Identification of biomarkers of early-stage disease typically requires analysis of very large cohorts which can only be reasonably achieved using high-throughput methods. We have developed and optimised novel methods for whole blood analysis using volumetric absorptive microsampling devices to produce over 3000 protein identifications by LCMS on a Q Exactive HF-X Orbitrap. These methods were tested using a set of whole blood samples from lung cancer patients and matching healthy controls finding 455 differentially expressed proteins, using a mid-throughput method enabling analysis of 18 samples per day. To increase throughput for larger clinical cohorts, a 60-sample per day method was tested on a Sciex ZenoTOF 7600. The high-throughput method produced 1.5-fold fewer protein identifications and a higher overall % CV compared to the mid-throughput method. Despite the lower numbers, it produced a set of 36 disease-relevant and discriminatory differentially expressed proteins that, using a machine learning model, could differentiate between the disease and control samples with an area under the ROC curve (AUC) of 88.9% using random forest algorithms. These data support the use of high-throughput mass spectrometry methods to screen large cohorts for diagnostic biomarkers that can then be followed up with more targeted analyses.

biochemistry↗