Deciphering the role of endothelial granulocyte macrophage-CSF in chronic inflammation associated with HIV

Abstract

People with HIV (PWH) experience endothelial dysfunction (ED) that is aggravated by chronic inflammation and microbial translocation across a damaged gut barrier. Although this paradigm is well-described, downstream pathways that terminate in endothelial dysfunction are only partially understood. This study found increased expression of granulocyte macrophage colony stimulating factor (GM-CSF), toll-like receptor-4 (TLR4), and myeloperoxidase in the aortic endothelium of PWH compared to those without HIV. Bacteria-derived lipopolysaccharide (LPS) heightened glucose uptake and induced GM-CSF expression in primary human endothelial cells. Exposure to sodium-glucose cotransporter-2 (SGLT2) inhibitors reduced glucose uptake, GM-CSF release, and ED in LPS-activated endothelial cells ex vivo, and PWH treated with SGLT2 inhibitors for diabetes had significantly lower plasma GM-CSF levels than non-diabetic PWH not on this medication. The findings suggest that microbial products trigger glucose uptake and GM-CSF expression in the endothelium, contributing to localized inflammation in PWH. Modifying this altered state could offer therapeutic benefits.

Keywords: Cell biology; Immunology; Molecular biology; Virology.

Graphical abstract

Figure 1

Dual gene expression analysis and GM-CSF protein quantification in the vascular endothelium of PWH (A) Representative combined FISH and immunofluorescence staining in the vascular endothelium and subendothelial tissues obtained an HIV negative (top) and an PWH (bottom) donors for CSF2 mRNA (red) and GM-CSF protein (green); the endothelial lining is shown by the space between dashed lines; each circle represents the average fluorescence intensity data from at least 5 images acquired from a tissue sample. Right panel: Violin plots show the quantitative assessment of CSF2 mRNA (upper: MFI-red) and GM-CSF protein (lower: MFI-green) expression in HIV Negative donors (n = 12), and PWH (n = 10), (B) Positive correlations of CSF2 mRNA and GM-CSF protein expression through combined FISH and immunofluorescence staining in the vascular endothelium obtained from negative controls (n = 12) and PWH (n = 10), (A: ∗p < 0.05). (C) Measurement of GM-CSF levels in plasma samples obtained from HIV Neg. (n = 17) and PWH (n = 17) using ELISA., (∗p < 0.05). All data are represented as mean ± SEM.

Figure 2

Myeloperoxidase levels in the endothelium and sub-endothelium of PWH (A) Epi-fluorescent microscopy images depict myeloperoxidase (MPO, red) positive staining in the vascular endothelium and subendothelial tissues obtained from HIV Negative controls and PWH. (B) Violin plots showing quantified MFI values of MPO expression in the vascular endothelium and subendothelial tissues of HIV Neg and PWH. Each circle represents the average fluorescence intensity data from at least 5 images acquired from a tissue sample. (HIV Neg., n = 12; and PWH, n = 10). (∗p < 0.05). Data are represented as mean ± SEM.

Figure 3

Quantitative analysis of toll-like receptor 4 (TLR4) gene and receptor expression in vascular endothelium of PWH (A) TLR4 mRNA copy number per 100k pixel in the vascular endothelium and subendothelial tissues of uninfected controls (n = 12) and PWH (n = 10). (∗∗∗p < 0.0005). (B) TLR4 receptor expression in the vascular endothelium comparing uninfected controls and PWH (n = 22), (∗p < 0.05). All data are represented as mean ± SEM.

Figure 4

LPS-induced GM-CSF expression and release by human aortic endothelial cells (A) Immunofluorescence microscopy image shows intracellular staining of GM-CSF (red), eNOS (green), and DAPI (blue) in primary human aortic endothelial cells (ECs). Time-dependent changes in intracellular GM-CSF and eNOS levels are quantified using the intracellular GM-CSF-to-DAPI MFI ratio and intracellular eNOS-to-DAPI MFI ratio. The image represents endothelial responses to LPS at indicated time points from one of the three identical experiments. (B) (Top image) Detection of GM-CFS by Western Blot (see Figure S3C for the complete WB data) in EC supernatant and in purified, lysed EVs from cells stimulated overnight with LPS (100 ng/mL) or incubated without stimulation (supernatant 2). (Bottom image) ELISA data (n = 3) illustrating a dose-dependent increase in GM-CSF concentration in cell culture supernatants or in purified and lysed EVs, derived from culture supernatants following overnight incubation of ECs with or without exposure to LPS at varying concentrations. Data shown here (A, B) are from one of the 3 independent experiments, each circle represents average MFI values from one image with ∼10 ECs, n = ∼100 cells/experiment (∗p < 0.05). (C) GM-CSF detection in endothelial cell culture supernatants after overnight incubation with the indicated TLR agonists (n = 3, ∗p < 0.05). All data are represented as mean ± SEM.

Figure 5

Single-cell RNA sequencing (RNAseq) reveals dynamic gene expression patterns in human aortic endothelial cells (ECs) ECs were cultured in medium alone or medium supplemented with LPS (100 ng/mL) or Dapa (1μg/mL) for 24 h. An additional 7days incubation period was included for estradiol (10 ng/mL) only (D5-E2-D7). (A). Unbiased clustering and UMAP representation of scRNAseq data analyzed with a resolution setting of 0.12, displaying distinct clusters of human aortic endothelial cells on a UMAP plot. (B) HeatMap of treatment-dependent differentially expressed genes illustrates the differential expression of genes in ECs across various treatment conditions, highlighting treatment-dependent clusters. (C and D) Violin plots display the expression profiles of selected genes (KLF2, eNOS, ICAM-1, VCAM-1, CSF-1, CSF2, and CSF3) within clusters from LPS treated cells. Expression levels are represented as probability distributions across clusters on the y axis.

Figure 6

Modulation of LPS-induced GM-CSF expression by human aortic endothelial cells (EC) through pharmacological interventions Endothelial cells were cultured in medium alone or medium supplemented with LPS (100 ng/mL), LPS + Dapa (1μg/mL) or LPS + TLR-4 Inhibitor (TLR4-C34) (10μg/mL). (A) Representative microscopy and fluorescent images of NBD glucose uptake under the indicated conditions (left) and violin plots show the quantification of NBD glucose uptake (MFI, right) in EC receiving indicated treatments (data from one of the 3 independent experiments, each circle represents MFI value from one EC, n = ∼100 cells/experiment, ∗∗∗p < 0.0005). (B) Immunofluorescence microscopy images (left) and violin plots showing quantified image data (MFI/DAPI ratio, right) illustrating KLF2 protein expression in indicated treatment groups (data from 100 ECs per condition; each circle represents average MFI values from one image with 5–10 ECs). (C) ELISA data illustrate GM-CSF protein concentrations in supernatants of endothelial cell cultures after 24 h of incubation under specified conditions (n = 3; ∗p < 0.05). (D) Violin plots show ELISA data of GM-CSF levels in plasma samples from PWH, with (n = 7) and without (n = 10) receiving Dapa treatment (∗∗∗p < 0.0005). All data are represented as mean ± SEM.