Fucosylated haptoglobin drives inflammation via Mincle in sepsis: an observational study

Haptoglobin (Hp) scavenges cell-free hemoglobin and correlates with the prognosis of human sepsis, a life-threatening systemic inflammatory condition. Despite extensive research on Hp glycosylation as a glyco-biomarker for cancers, understanding glycosylated modifications of Hp in sepsis patients (SPs) remains limited. Our study reveals elevated levels of terminal fucosylation at Asn207 and Asn211 of Hp in SP plasma, along with heightened inflammatory responses, compared to healthy controls (trial registration NCT05911711). Fucosylated (Fu)-Hp purified from SPs upregulates inflammatory cytokines and chemokines, along with NLRP3 inflammasome activation. Single-cell RNA sequencing identifies a distinct macrophage-like cell population with increased expressions of inflammatory mediators and FUT4 in response to Fu-Hp. Additionally, Mincle, a C-type lectin receptor, interacts with Fu-Hp to amplify the inflammatory responses and signaling. Moreover, the Hp fucosylation (AAL) level significantly correlates with the levels of inflammatory cytokines in sepsis patients, suggesting that Fu-Hp is clinically relevant. Finally, Fu-Hp treatment significantly enhances the levels of inflammatory cytokines in the plasma and various tissues of mice. Together, our findings reveal a role of Fu-Hp, derived from sepsis patients, in driving inflammation, and suggest that targeting Fu-Hp could serve as a promising intervention for combating sepsis. Trial registration NCT05911711 Overall design: RNA sequencing was performed on PBMCs from HC and SP to examine the different gene expression and ontology pathway analysis. Briefly, RNA-seq libraries were prepared using a SureSelect RNA Direct_Human Library Construction Kit (Agilent Technologies). Total RNA was extracted from PBMCs using TRIzol® Reagent (Thermo Fisher, 15596026) according to the manufacturer's instructions. Transcriptome library sequencing was carried out using the 101-bp paired-end mode with Illumina NovaSeq 6000. The data integrity of the raw sequences was evaluated with FastQC v0.11.7 (http://www.bioinformatics.babraham.ac.uk/projects/fastqc/). The raw read data were trimmed using Trimmomatic (version 0.38) and the following criteria: first trimming: base quality < 3, window size = 4, and mean quality = 15; second trimming: min length = 36 bp. The trimmed reads were mapped against the reference genome GRCh38 using HISAT2 version 2.1.0 (https://ccb.jhu.edu/software/hisat2/index.shtml) and Bowtie2 aligner. After the read mapping, StringTie (version 2.1.3b) was used for transcript assembly. The expression profile was calculated for each sample and transcript/gene as read count, FPKM (Fragment per Kilobase of transcript per Million mapped reads) and TPM (Transcripts Per Kilobase Million). DEG (Differentially Expressed Genes) analysis was performed on a comparison pair (Sepsis patients_vs_Healthy control) using DESeq2. The statistical method used was calculated based on fold change, nbinomWaldTest using DESeq2, and Hierarchical Clustering. Using each sample's normalized value, the high expression similarities were grouped together. The significant results are selected on conditions of the absolute value of |log2(fold change)| >= 2 & nbinomWaldTest raw p-value < 0.05. Normalized gene expression values were utilized to generate a heatmap using the pheatmap package (version 1.0.12) and a volcano plot employing the ggplot2 package (version 3.4.4) within the R programming environment (version 4.1.3). Pathway enrichment analysis was conducted using the clusterProfiler package (version 4.10.0) in R and the Gene Set Enrichment Analysis (GSEA) software (version 4.3.2).