Albumin Nanoparticle Endocytosing Subset of Neutrophils for Precision Therapeutic Targeting of Inflammatory Tissue Injury

Abstract

The complex involvement of neutrophils in inflammatory diseases makes them intriguing but challenging targets for therapeutic intervention. Here, we tested the hypothesis that varying endocytosis capacities would delineate functionally distinct neutrophil subpopulations that could be specifically targeted for therapeutic purposes. By using uniformly sized (∼120 nm in diameter) albumin nanoparticles (ANP) to characterize mouse neutrophils in vivo, we found two subsets of neutrophils, one that readily endocytosed ANP (ANP^high neutrophils) and another that failed to endocytose ANP (ANP^low population). These ANP^high and ANP^low subsets existed side by side simultaneously in bone marrow, peripheral blood, spleen, and lungs, both under basal conditions and after inflammatory challenge. Human peripheral blood neutrophils showed a similar duality. ANP^high and ANP^low neutrophils had distinct cell surface marker expression and transcriptomic profiles, both in naive mice and in mice after endotoxemic challenge. ANP^high and ANP^low neutrophils were functionally distinct in their capacities to kill bacteria and to produce inflammatory mediators. ANP^high neutrophils produced inordinate amounts of reactive oxygen species and inflammatory chemokines and cytokines. Targeting this subset with ANP loaded with the drug piceatannol, a spleen tyrosine kinase (Syk) inhibitor, mitigated the effects of polymicrobial sepsis by reducing tissue inflammation while fully preserving neutrophilic host-defense function.

Keywords: bacterial infection; chemokine receptors; drug carriers; inflammation; nanotechnology; nanotherapeutics; neutrophil heterogeneity.

Figure 1

Differential endocytosis of albumin nanoparticles by PMN. (A) Flow cytometric analysis of mouse lung single cell suspensions. Intravenous injection of albumin nanoparticles (ANP) led to endocytosis of ANP by CD177⁺Ly6G⁺ lung PMN. (B) Transmission electron microscopy (TEM) of lung neutrophil that did not endocytose ANP (ANP^low) and of lung neutrophil that endocytosed ANP (ANP^high). Red arrows indicate ANP in organelles. (C) Flow cytometric analysis of ANP endocytosis by mouse Ly6G⁺ PMN from various tissues. ANP^high PMN were found in bone marrow, peripheral blood, spleen, and lungs at low percentage when compared to PMN that did not endocytose ANP (ANP^low). Systemic challenge (i.v.) with LPS induced the expansion of ANP^high PMN relative to ANP^low PMN, which was most pronounced in lungs (ratio of 9.78). Mice (n = 4 per cohort) were injected with LPS 6 h prior to euthanasia and with unlabeled ANP (control) 30 min prior to euthanasia; with LPS or saline 6 h prior to euthanasia and with AF647-fluorochrome labeled ANP 30 min prior to euthanasia.

Figure 2

Transcriptomic heterogeneity of lung PMN. Transcriptomic profile of lung ANP^high vs ANP^low Ly6G⁺ PMN. (A) Mice were challenged for 6 h by i.p. injection of LPS (12 mg/kg) or saline; 5 h 30 min after challenge, mice were injected, i.v., with 1 dose of fluorochrome-labeled ANP, and euthanized 30 min later; Ly6G⁺ PMN from lungs were sorted according to ANP endocytosis and then mRNA was processed for RNA-Seq. (B) Dendrogram and heat map showing normalized gene expression data of biological replicate samples from saline injected controls, ANP^low (blue), ANP^high (purple), or of LPS-challenged mice, ANP^high (red), ANP^low (green). (C–F) Heatmaps of chemokine receptors or chemokines. Chemokine receptor expression in lung PMN from saline injected mice (C) or LPS-challenged mice (D). Chemokine expression in PMN from (E) saline injected mice or (F) LPS-challenged mice. Significantly higher expression values are shown in red, lower expression values in blue. Representative data from 3 independent experiments.

Figure 3

Kinetics of chemokine and cytokine mRNA expression in lung PMN subsets. Mice were treated with either saline (0 h) or LPS for 3, 6, and 12 h. Fluorochrome labeled ANP were injected 30 min before euthanasia. Ly6G⁺ PMN were sorted into ANP^low and ANP^high. qPCR analysis of ANP^low (green columns) or ANP^high (red columns) PMN. qPCR analysis of (A) Ccl3; (B) Ccl4; (C) Cxcl2; (D) Cxcl3; (E) Il1b; (F) Il15. Representative data from 3 independent experiments. Mean values and SD. n.s., not significant. ** p < 0.01; *** p < 0.001 (Student’s t test).

Figure 4

PMN subset cell-surface-marker expression. (A) Contour plots according to cell density of lung CD45⁺ leukocytes. The three maps were generated by considering 38 cell surface markers (encoding genes listed in Table 1). Unsupervised grouping of individual cells. viSNE visualization of high dimensional single-cell data separated most major leukocyte subtypes (shown in right-hand panel). Unsupervised grouping was confirmed by the markers listed for PMN, T cell B cells natural killer (NK) cells, and dendritic cells (DCs). The contours in each plot (representing cell density) delineated cell heterogeneity. Arrows indicate differences in abundance within the neutrophil compartment. Green arrows indicate ANP^low and red arrows ANP^high PMN. Mice (n = 4 per cohort) were injected with LPS 6 h prior to euthanasia and with AF647-fluorochrome labeled ANP 30 min prior to euthanasia. Lung CD45⁺ leukocytes were left unsorted or sorted by flow cytometry according to their ability to endocytose ANP into ANP^high and ANP^low cells. Cells were then subjected to mass cytometry (CyTOF), and visualized by viSNE, based on the t-Distributed Stochastic Neighbor Embedding (t-SNE) algorithm. (B) Contour plots according to cell density of lung CD177⁺Ly6G⁺ PMN generated by considering 38 cell surface markers (encoding genes listed in Table 1) from mice injected with saline or LPS. The viSNE maps were generated by unsupervised grouping of individual cells. Arrows indicate differences in abundance between neutrophil subsets. Green arrows, ANP^low; red arrows ANP^high. (C–E) Spectrum colored dot plots. Intensities of protein expression of markers are shown on viSNE map as spectrum colored dots (low in blue, high in red). (F, G) Spectrum colored dot plots. Intensities of protein expression of markers are shown on viSNE map as spectrum colored dots (low in blue, high in red). (H) Contour plots according to cell density of lung CD177⁺Ly6G⁺ PMN generated by considering 5 cell surface markers (antigens listed in Table 2) from mice injected with saline or LPS. The 5 viSNE maps were generated by unsupervised grouping of individual cells. Arrows indicate differences in abundance between neutrophil subsets. Green arrows, ANP^low; red arrows ANP^high.

Figure 4

PMN subset cell-surface-marker expression. (A) Contour plots according to cell density of lung CD45⁺ leukocytes. The three maps were generated by considering 38 cell surface markers (encoding genes listed in Table 1). Unsupervised grouping of individual cells. viSNE visualization of high dimensional single-cell data separated most major leukocyte subtypes (shown in right-hand panel). Unsupervised grouping was confirmed by the markers listed for PMN, T cell B cells natural killer (NK) cells, and dendritic cells (DCs). The contours in each plot (representing cell density) delineated cell heterogeneity. Arrows indicate differences in abundance within the neutrophil compartment. Green arrows indicate ANP^low and red arrows ANP^high PMN. Mice (n = 4 per cohort) were injected with LPS 6 h prior to euthanasia and with AF647-fluorochrome labeled ANP 30 min prior to euthanasia. Lung CD45⁺ leukocytes were left unsorted or sorted by flow cytometry according to their ability to endocytose ANP into ANP^high and ANP^low cells. Cells were then subjected to mass cytometry (CyTOF), and visualized by viSNE, based on the t-Distributed Stochastic Neighbor Embedding (t-SNE) algorithm. (B) Contour plots according to cell density of lung CD177⁺Ly6G⁺ PMN generated by considering 38 cell surface markers (encoding genes listed in Table 1) from mice injected with saline or LPS. The viSNE maps were generated by unsupervised grouping of individual cells. Arrows indicate differences in abundance between neutrophil subsets. Green arrows, ANP^low; red arrows ANP^high. (C–E) Spectrum colored dot plots. Intensities of protein expression of markers are shown on viSNE map as spectrum colored dots (low in blue, high in red). (F, G) Spectrum colored dot plots. Intensities of protein expression of markers are shown on viSNE map as spectrum colored dots (low in blue, high in red). (H) Contour plots according to cell density of lung CD177⁺Ly6G⁺ PMN generated by considering 5 cell surface markers (antigens listed in Table 2) from mice injected with saline or LPS. The 5 viSNE maps were generated by unsupervised grouping of individual cells. Arrows indicate differences in abundance between neutrophil subsets. Green arrows, ANP^low; red arrows ANP^high.

Figure 5

ANP^high PMN transfer inflammation. (A) Timeline of adoptive transfer. Donor mice were challenged with a lethal dose of LPS and injected with two doses of ANP labeled with the stable fluorochrome AF647. ANP^high PMN (8 × 10⁵) or, as controls, an equal number of ANP^low lung Ly6G⁺ PMN were adoptively transferred by i.v. injection into syngeneic recipient mice. (B) Flow cytometric analysis of lung cells from mice that received ANP^low or ANP^high donor cells. Dot blot. Percentages of Ly6G⁺ANP⁺ cells (red) were significantly greater in mice that received ANP^high donor cells as compared to mice that received ANP^low donor cells. p < 0.001 (Student’s t test). (C) Flow cytometric analysis of lung cells from mice that received ANP^low or ANP^high donor cells. Dot blot. Percentages of ROS⁺CD11b⁺ (red) were significantly greater in mice that received ANP^high donor cells as compared to mice that received ANP^low donor cells. (D) Concentrations of IL-1β in lung tissue extracts from mice that have received ANP^low or ANP^high donor cells. (E) Concentrations of CXCL2 in lung tissue extracts from mice that have received ANP^low or ANP^high donor cells. Squares represent values from individual mice and lines indicate mean values + SD *p < 0.05 (Student’s t test). Representative data from 3 independent experiments are shown.

Figure 6

Functional heterogeneity of lung PMN. (A) Timeline of assays. (B) E. coli phagocytosis. PMN were incubated with E. coli bacteria (50 cfu or 100 cfu) for 1 h. E. coli-specific fluorescence is shown. (C) Killing of intracellular E. coli bacteria. Single cell suspensions of lung unsorted Ly6G⁺ PMN (blue) or sorted according to endocytosis of ANP (low, green; or high, red). PMN were incubated with E. coli bacteria (50 cfu or 100 cfu) for 1 h, and then washed and incubated for additional 3 h to evaluate bacterial killing. E. coli-specific fluorescence at 4 h relative to E. coli-specific fluorescence at 1 h, corresponding to bacterial killing. Average (n = 3) of fluorescence detected at 1 h = 100%; % killing = 100 – percentage of fluorescence detected at 4 h post start of incubation. Markers represent results from individual mice*p < 0.005, **p < 0.002, ***p < 0.0002. (D) mRNA expression of hydrogen voltage gated channel 1 (Hvcn1) and peptidyl arginine deiminase 4 (Padi4) in ANP^low and ANP^high lung PMN after LPS-challenge. (E) ROS production by ANP^high is not impaired. Mice were injected i.p. with LPS and 2 h 30 min later with ANP (i.v.) and euthanized 30 min thereafter. Single cell suspensions of the lungs Ly6G⁺ PMN were sorted according to endocytosis of ANP. Cells were incubated and stimulated with DMSO or phorbol ester PMA. Control, unsorted PMN from naive mice. Squares represent results from individual mice. *p < 0.005, **p < 0.002, ***p < 0.0001.

Figure 7

Therapeutic targeting ANP^high PMN. (A) Flow cytometric analysis of peripheral blood, lung, or liver Ly6G⁺ PMN. % of cells with high ANP-specific fluorescence 6 h after surgery. Squares represent values from individual mice and lines indicate mean values + SD (B,C) Kaplan–Meier survival curves. (B) i.v. injections of PANP or ANP given 2 and 4 h after CLP. (C) i.v. injections of PANP or ANP given at 1 and 2 h after i.p. challenge with LPS [30 mg/kg]. Representative data 10 mice per treatment group. *p < 0.05, *** p < 0.0001. (D,E) Number and velocity of lung Ly6G⁺ PMN determined by two-photon in vivo microscopy of lungs. Velocity in μm/s. i.p. injections of LPS [30 mg/kg], 6 h prior to, and of ANP or PANP 3 and 4 h prior to start of imaging. (D) Fluorescence (a.u.) of Ly6G⁺ PMN. n = 5 for ANP, n = 6 for PANP. (E) Quantitative analysis of PMN velocity (n = 40 for ANP, n = 49 for PANP). PMN velocity determined for at least 1 min in the field of view. Error bars indicate SD * p < 0.01*, **p < 0.005, *** p < 0.001.

Figure 8

Targeting ANP^high PMN improves tolerance of polymicrobial infection. (A) Flow cytometric analysis of intracellular ROS in lung PMN. Mice were treated with 2 consecutive i.v. injections of PANP or ANP 1 and 2 h after i.p. challenge with LPS, and ROS was measured 6 h after challenge. Histograms representing ROS in all Ly6G⁺ PMN, ANP^high, or PANP^high (P/ANP^high). ROS production was measured using DHR-123. PMN ROS is significantly reduced by PANP treatment (from average 50.2% ROS⁺ cells to 4.7%). Representative data from a minimum of 3 mice per treatment group are shown. (B) Bacterial load (cfu) in peripheral blood, lungs, livers, or spleens of mice post-CLP. PANP treatment, given 2 and 4 h after CLP, did not increase bacterial burden 18 h after surgery compared to ANP-vehicle treated controls. n.s., not significant. (C) IL-1β and CXCL2 in lung protein lysate of mice 18 h after CLP. Two i.v. injections of PANP or ANP were given 2 and 4 h after CLP. Squares represent values from individual mice, and lines indicate mean values + SD *p < 0.05. (D) Nitrotyrosine formation. Photomicrographs of lung and liver sections from septicemic mice treated with ANP-vehicle or PANP. Paraffin embedded sections were stained with Ab to nitrotyrosine (red) and with DAPI to visualize nuclei (blue). Polymicrobial sepsis was induced by CLP; mice were treated with PANP or ANP 2 and 4 h after challenge and were sacrificed for tissue processing and staining 18 h after challenge. Bar measures 40 μm, lung, or 20 μm, liver. (E) Quantification of nitrotyrosine formation (ratio of nitrotyrosine⁺ cells per nucleus). Error bars indicate SD ***p < 0.001. (F) Lung wet-weight to dry weight ratio of mice instilled, i.t., with live P. aeruginosa (10⁷ cfu) or after CLP to induce polymicrobial sepsis or after i.p. injections of LPS. Two consecutive i.v. injections of PANP or ANP were given 2 and 4 h after challenge. Lung wet to dry weight ratio measured 6 h after challenge. Squares represent values from individual mice and lines indicate mean values. *p < 0.05. (G) Serum markers of tissue damage measured 18 h after CLP and two consecutive injections of PANP or ANP given 2 and 4 h after CLP. Lactate dehydrogenase (LDH) activity was significantly reduced in peripheral blood sera obtained from PANP-treated mice compared ANP-vehicle treated controls. Hepatocyte-specific sorbitol dehydrogenase (SDH) activity was significantly reduced by PANP treatment when compared to ANP treated controls. Squares represent values from individual mice, and lines indicate mean values + SD *p < 0.05.