Hemoglobin increases leukocyte adhesion and initiates lung microvascular endothelial activation via Toll-like receptor 4 signaling

Abstract

Cell-free hemoglobin is a pathophysiological driver of endothelial injury during sepsis and acute respiratory distress syndrome (ARDS), but the precise mechanisms are not fully understood. We hypothesized that hemoglobin (Hb) increases leukocyte adhesion and endothelial activation in human lung microvascular endothelial cells (HLMVEC). We stimulated primary HLMVEC, or leukocytes isolated from healthy human donors, with Hb (0.5 mg/mL) and found that leukocyte adhesion to lung endothelium in response to Hb is an endothelial-dependent process. Next, we stimulated HLMVEC with Hb over time (1, 3, 6, and 24 h) and found increased transcription and release of inflammatory cytokines (IL-1β, IL-8, and IL-6). In addition, Hb exposure variably upregulated transcription, total protein expression, and cell-surface localization of adhesion molecules E-selectin, P-selectin, intercellular adhesion molecule-1 (ICAM-1), and vascular cell adhesion molecule-1 (VCAM-1). Since VCAM-1 was most upregulated by Hb, we further tested mechanisms for Hb-mediated upregulation of VCAM-1 in HLMVEC. Although upregulation of VCAM-1 was not prevented by hemoglobin scavenger haptoglobin, heme scavenger hemopexin, or inhibition of nod-like receptor protein 3 (NLRP3) signaling, blocking Toll-like receptor 4 (TLR4) with small molecule inhibitor TAK-242 (1 µM) prevented upregulation of VCAM-1 in response to Hb. Consistently, Hb increased nuclear factor-κB (NF-κB) activation and intracellular reactive oxygen species (ROS), which were both prevented by TLR4 inhibition. Together, these data demonstrate that Hb increases leukocyte-endothelial adhesion and activates HLMVEC through TLR4 signaling, indicating a potential mechanism for Hb-mediated pulmonary vascular injury during inflammatory and hemolytic conditions.

INTRODUCTION

Sepsis is a leading cause of morbidity and mortality (25%–30%) with an incidence of 750,000 hospitalizations per year (1, 2). Although not all patients with sepsis develop acute respiratory distress syndrome (ARDS), sepsis is the leading cause of ARDS (3). Despite recent improvements in survival, medical care of these patients remains largely supportive since targeted treatment remains limited to antimicrobials and infection source control. Therefore, an important goal of ongoing sepsis research is to understand the underlying mechanisms of the host response to provide insights into potential treatment modalities.

A key underlying pathophysiological mechanism of ARDS in sepsis is endothelial cell activation and death (4, 5). This initiation of the inflammatory response in endothelial cells involves interactions between pathogen-associated molecular patterns (PAMPs) or damage-associated molecular patterns (DAMPs) with specific pathogen recognition pattern receptors on endothelial cells (6, 7). Following activation, endothelial cells release cytokines and upregulate adhesion molecule expression to trigger leukocyte transmigration to the interstitial space and alveoli (6, 8). Endothelial activation can also lead to increased vascular permeability that contributes to the pulmonary edema formation that is characteristic of ARDS (9, 10).

Leukocyte recruitment to inflamed tissue is a critical step in acute inflammation. Classically, recruitment is described as a stepwise and dynamic process that requires interactions between vascular endothelium and leukocytes. This process is initiated by capture and rolling of leukocytes along the endothelium via reversible binding between leukocyte glycoproteins and endothelial transmembrane glycoproteins called selectins. Following capture and rolling, firm adhesion of leukocytes to the endothelium occurs via binding of leukocyte integrins (CD11/CD18), and interaction with intercellular adhesion molecules [intercellular adhesion molecule-1 (ICAM-1) and vascular cell adhesion molecule-1 (VCAM-1)] leading to arrest of leukocytes on the endothelium before transmigration into the inflamed tissue (10–12). However, regulation of leukocyte adhesion and transmigration may differ in the pulmonary circulation compared with the systemic circulation (11); therefore, evaluation of the specific regulation of leukocyte adhesion molecules in response to hemoglobin (Hb) in pulmonary microvascular endothelial cells is needed.

Previously, we demonstrated that 80% of patients with sepsis have high circulating levels of hemoglobin (Hb), and patients with higher levels of Hb circulating in plasma have higher mortality from sepsis (13). Hb increases paracellular permeability in cultured endothelial cells (14, 15) and increases microvascular permeability in an ex vivo isolated perfused human lung model (15) and a murine model of sepsis (16). Several studies have shown that Hb can act as a damage-associated molecular pattern (DAMP), signaling through Toll-like receptor 4 (TLR4) and nuclear factor-κB to activate endothelium during hemolytic pathologies (17–20) but the specific cellular mechanisms regulating pulmonary endothelial cell activation in sepsis remain poorly understood. We hypothesized that during sepsis, the release of Hb into the circulation directly modulates the lung microvascular endothelium to increase cytokine production and leukocyte adhesion through endothelial activation mediated by TLR4 signaling. In this manuscript, we test this hypothesis using cultured primary human pulmonary microvascular endothelial cells treated with Hb over several time points.

METHODS

HLMVEC Cell Culture

Human lung microvascular endothelial cells (HLMVEC) were purchased from Promocell (Heidelberg, Germany) and grown in Endothelial Cell Growth Medium MV2 (Promocell). Cells were seeded onto 0.1% gelatin-coated plates and incubated in a 5% CO₂ humidified incubator at 37°C for 2–3 days to allow for establishment of tight junctions for use in experimental assays and incubated with endotoxin-free native human hemoglobin (Hb; Cell Sciences, Canton, MA) dissolved in PBS (0.5 mg/mL) or PBS control. For inhibitor studies, HLMVEC were pretreated for 1 h with TAK-242 (1 µM; Calbiochem, Cat. No. 614316) or MCC950 (10 µM; Calbiochem, Cat. No. 5.38120), or treated simultaneously with 2:1 molar ratio of haptoglobin (15.5 µM, purified from human plasma; gift from CSL Behring) or 1:1 molar ratio of hemopexin (7.75 µM, Sigma H9291), or vehicle control (DMSO or PBS), and stimulated with Hb (0.5 mg/mL) or PBS control for 3 h. Cells were used between passages 3–6 and each experiment was completed with cells from at least two separate donors.

Isolation of Peripheral Blood Leukocytes

Leukocytes were isolated from fresh human blood via venipuncture from a healthy donor on the day of the experiment (written informed consent was obtained and approved by Vanderbilt Institutional Review Board, Protocol No. 061088). Isolation was completed using Histopaque 1077/1119 double-gradient density centrifugation (Sigma Aldrich, St Louis, MO), according to the manufacturer’s instructions to obtain purified leukocytes (>95%). The absolute number of leukocytes was adjusted to 1 × 10⁶ cells/well in RPMI media (Millipore Sigma) for each reaction.

Leukocyte Adhesion Assay

HLMVEC were grown to confluence in 0.1% gelatin-coated 24-well plates. On the day of the experiment, endothelial cells (6 h) or leukocytes (1 h) were pretreated individually or in combination with PBS or Hb (0.5 mg/mL). After stimulation, HLMVEC and leukocytes were rinsed three times with warm PBS. Prior to endothelial-leukocyte incubations, HLMVEC were labeled with CellMask Deep Red Stain (Thermo Fisher), and leukocytes were labeled with CellTracker Green (Thermo Fisher) per the manufacturer’s instructions. Leukocytes were incubated with HLMVEC for 1 h. Before fixation (4% PFA for 10 min), nonadherent cells were washed three times with warm PBS. Following fixation, the wells were stained with Hoechst (100 ng/mL) for 15 min. Images were captured using Lionheart FX (BioTek Instruments, Inc., Winooski, VT) automated microscope in 3 × 3 montage at ×4 magnification in DAPI (377,447), CY5 (628,685), and GFP (469,525) channels. Images were analyzed using Gen5 Image+ software version 3.10. Montage images were stitched using Linear Blend of DAPI channel. Cellular analysis to count total number of cells included those between 5 and 50 µm meeting fluorescent threshold based on background. Image statistics were used to measure the total fluorescent area (GFP > 200). Data are from four replicates each from three individual experiments.

Multiplex Enzyme-Linked Immunosorbent Assay

Conditioned media was collected from HLMVEC treated with Hb (0.5 mg/mL) at 1, 3, 6, and 24 h from five individual experiments. An electrochemiluminescence (ECL) assay was used to determine concentrations of IL-6, IL-8, and IL-1β according to the manufacturer’s instructions (Meso Scale Diagnostics, Rockville, MD).

Real-Time PCR

Total RNA from HLMVEC was extracted using Qiagen RNeasy Mini Kit per the manufacturer’s instructions. RNA purity and concentration were quantified with Nanodrop Microvolume Spectrophotometer. Total RNA of 1 µg was used per 20 µL of iScript cDNA Synthesis Kit. Gene-specific PCR primers from Bio-Rad were used to analyze mRNA expression with PCR. Qualitative PCR amplifications were performed in a BioRad thermocycler with the following program: 1 step of 2 min at 95°C, 40 cycles of 5 s at 95°C, 30 s at 60°C and 5 s at 72°C and the final step of 5 min at 72°C. Fold increase in mRNA compared with GAPDH was calculated by 2^ΔΔCt as a quantitative estimate as previously described (21).

Western Blots

Cell lysates were prepared from ∼2 × 10⁶ cells using radioimmunoprecipitation assay (RIPA) lysis buffer (Sigma Aldrich) with a protease inhibitor (Roche) per the manufacturer’s instructions. Cell lysates were analyzed to determine protein concentration via bicinchoninic acid (BCA) kit (Thermo Scientific). Samples were reduced with BOLT reducing agent (Novex) at 75°C for 10 min and 20 µg of protein was electrophoresed on BOLT Bis-Tris 4%–12% gels. Protein was then transferred to a nitrocellulose membrane using iBLOT system (Invitrogen). The membrane was blocked with Odyssey TBS blocking buffer for 1 h and incubated with primary antibody [E-selectin (Invitrogen, PA5-86742, 1:1,000), P-selectin (Thermo Fischer, 3H20L10, 1:1,000), ICAM-1 (Abcam, ab2213, 1:500), VCAM-1 (Thermo-Fisher, SA05-04, 1:1,000), phospho-NF-κB p65 (Cell Signaling Technology 3033, 1:1,000)] overnight at 4°C. The following morning the membrane was washed five times with 0.1% Tween/Tris-buffered saline (TBS) wash buffer before incubation with LI-COR donkey anti-rabbit or donkey anti-mouse secondary antibody (1:15,000) for 1 h at room temperature. After a further five washes with 0.1% Tween/TBS wash buffer, images were scanned and analyzed with LI-COR Odyssey Clx. Band intensities were quantified using Image Studio Lite, v5.2.

Immunocytochemistry

HLMVEC were fixed with 4% PFA for 10 min, washed twice with PBS, blocked for 1 h with 4% BSA, and stained with E-selectin (Invitrogen, PA5-86742, 1:100), P-selectin (Thermo Fisher, 3H20L10, 1:200), ICAM-1 (Invitrogen, PA5-82002, 1:100), or VCAM-1 (Thermo Fisher, SA05-04, 1:200) antibody overnight at 4°C. Because the goal was to investigate cell surface-specific staining of adhesion molecules, cells were not permeabilized. The next day, HLMVEC were washed and stained with AF647-conjugated donkey anti-rabbit (1:500) for 1 h at room temperature. Further washes were followed by staining with Hoechst (100 ng/mL) for 45 min. HLMVEC were washed for the final time before imaging with Lionheart FX (BioTek Instruments, Inc., Winooski, VT) automated microscope in 3 × 3 montage at ×20 magnification in DAPI (377,447) and CY5 (628,685) channels. Images were analyzed using Gen5 Image+ software version 3.10 in an automated fashion to remove researcher bias. Cellular analysis to count total number of cells included those between 5 and 50 µm meeting fluorescent threshold based on background. Image statistics were used to measure the total fluorescent intensity (CY5 > 10,000). Data are from three replicates each from three independent experiments.

Intracellular ROS Quantification

HLMVEC were stained with CellROX Deep Red (5 µM; Invitrogen, C10422) for 30 min at 37°C, washed twice with complete media and once with PBS, fixed with 4% PFA for 15 min at room temperature, counterstained with DAPI (2 µg/mL) for 15 min at room temperature, and imaged on Lionheart FX. Each well was imaged at ×20 for nine fields of view. Mean fluorescence intensity (MFI) of Cy5 channel was normalized to cell count (DAPI). Data are from five replicates each from two independent experiments.

Statistical Analysis

Graph construction and statistics were carried out using GraphPad Prism v7. All analysis is shown as means ± SE. Comparisons between groups were made using one-way ANOVA with Dunnett’s or Tukey’s multiple-comparisons test or two-way ANOVA with Sidak’s multiple-comparisons test (α = 0.05).

RESULTS

Hemoglobin Increases Leukocyte Adhesion to Human Lung Microvascular Endothelial Cells

Leukocyte adhesion to human lung microvascular endothelial cells (HLMVEC) was assessed after pretreatment of either HLMVEC alone, leukocytes alone, or both HLMVEC and leukocytes with hemoglobin (Hb) or PBS control. Hb caused a significant increase in adhered leukocytes (GFP fluorescence) compared with control when HLMVEC were exposed to Hb, with or without concomitant leukocyte exposure to Hb, but not when leukocytes alone were exposed to Hb (Fig. 1).

Figure 1.

Hemoglobin (Hb) increases leukocyte adhesion to human lung microvascular endothelial cells (HLMVEC). Hb was used to stimulate either endothelial cells (EC) or leukocytes (polymorphonuclear neutrophils, PMN) alone, or both EC and PMN in combination, before exposing HLMVEC to PMN and assessing leukocyte adhesion. Representative images (Blue = nuclei, Red = EC, Green = PMN; scale bar = 2 mm) (A) and quantification of immunofluorescence staining of PMN (GFP) (B) after adhesion assay (n = 12 replicates) indicate that Hb-mediated PMN adhesion to HLMVEC is dependent on EC. Statistical analysis was performed using two-way ANOVA with Tukey’s multiple comparisons. Graph represents means ± SE.

Hb Increases HLMVEC Transcription and/or Secretion of Cytokines IL-1β, IL-8, and IL-6

Activated endothelial cells upregulate and secrete cytokines including IL-1β, IL-8, and IL-6 to increase leukocyte trafficking to areas of acute inflammation (6, 8). To evaluate whether Hb can activate endothelial cells to release these cytokines, HLMVEC were treated with Hb for 1, 3, 6, and 24 h (Supplemental Fig. S1; see https://doi.org/10.6084/m9.figshare.21692201). Following Hb stimulation, transcription increased for IL-1β at 6 h (Supplemental Fig. S1A), IL-8 at 3 h (Supplemental Fig. S1B), and IL-6 at 24 h (Supplemental Fig. S1C). In cell media, release of IL-1β (Supplemental Fig. S1D) and IL-6 (Supplemental Fig. S1F), but not IL-8 (Supplemental Fig. S1E), increased at 24 h.

Hb Increases HLMVEC Transcription and Cell Surface Expression of E-Selectin, but Not P-Selectin

Following exposure to Hb, transcription of E-selectin increased at 24 h (Fig. 2A). There was no significant change in the total amount of E-selectin protein in cultured cell lysates (Fig. 2, B and C), however, a significant increase in cell surface expression via immunocytochemistry was demonstrated at 3 and 6 h (Fig. 2, D and E) in nonpermeabilized, fixed cells. No notable changes in P-selectin transcription (Fig. 3A), total protein expression (Fig. 3, B and C), or cell surface expression (Fig. 3, D and E) were observed.

Figure 2.

Hemoglobin (Hb) increases transcription and cell surface expression of E-selectin in human lung microvascular endothelial cells (HLMVEC). HLMVEC were stimulated with Hb over 1, 3, 6, and 24 h. A: E-selectin mRNA (n = 7–9 replicates) increased significantly at 24 h. Representative Western blot (B) and quantification of E-selectin (n = 5 replicates) (C) indicate no change of E-selectin protein expression over time. Representative images [Blue = nuclei, Red = E-selectin; scale bar = 300 µm] (D) and quantification of E-selectin immunofluorescence (n = 9 replicates) (E) show increased cell surface expression of E-selectin at 3 and 6 h. Statistical analysis was performed using one-way ANOVA with Dunnett’s multiple comparisons. Graphs represent means ± SE.

Figure 3.

Hemoglobin (Hb) does not increase transcription, protein, or cell surface expression of P-selectin in human lung microvascular endothelial cells (HLMVEC). HLMVEC were stimulated with Hb over 1, 3, 6, and 24 h. A: P-selectin mRNA (n = 7–9 replicates) did not change over time. Representative Western blot (B) and quantification of P-selectin (C) (n = 4 replicates) indicate no change of P-selectin protein expression over time. Representative images [Blue = nuclei, Red = P-selectin; scale bar = 300 µm] (D) and quantification of P-selectin immunofluorescence (n = 9 replicates) (E) show no changes in P-selectin cell surface expression. Statistical analysis was performed using one-way ANOVA with Dunnett’s multiple comparisons. Graphs represent means ± SE.

Hb Upregulates HLMVEC Adhesion Molecules ICAM-1 and VCAM-1

Following exposure to Hb, ICAM-1 transcription was significantly elevated at 3 h (Fig. 4A), but total ICAM-1 protein was not increased at any timepoint (Fig. 4, B and C). Immunofluorescence staining for ICAM-1 demonstrated increased cell surface expression at 24 h (Fig. 4, D and E). Total quantification of ICAM-1 fluorescence intensity remained unchanged over all time points; however, when only non-nuclear staining was quantified, ICAM-1 surface expression increased significantly at 24 h (Fig. 4E).

Figure 4.

Hemoglobin (Hb) increases transcription and cell surface expression of intercellular adhesion molecule-1 (ICAM-1) in human lung microvascular endothelial cells (HLMVEC). HLMVEC were stimulated with Hb over 1, 3, 6, and 24 h. A: ICAM-1 mRNA (n = 7–9 replicates) increased significantly at 3 h. Representative Western blot (B) and quantification of ICAM-1 (n = 8 replicates) (C) indicate no change of ICAM-1 protein expression over time. Representative images [Blue = nuclei, Red = ICAM-1; scale bar = 300 µm] (D) and quantification of ICAM-1 immunofluorescence (n = 9 replicates) (E) indicate increased cell surface expression of ICAM-1 at 24 h. Statistical analysis was performed using one-way ANOVA with Dunnett’s multiple comparisons. Graphs represent means ± SE.

Following exposure to Hb, VCAM-1 transcription and total protein expression were increased significantly at 3 and 6 h (Fig. 5, A–C). Immunofluorescence staining for VCAM-1 demonstrated a significant increase in cell surface expression by 3 h that was sustained for 24 h (Fig. 5, D and E). Because VCAM-1 upregulation was most robustly increased of all the adhesion molecules, further mechanistic studies focused on regulation of VCAM-1 in response to Hb.

Figure 5.

Hemoglobin (Hb) increases transcription, protein, and cell surface expression of vascular cell adhesion molecule-1 (VCAM-1) in human lung microvascular endothelial cells (HLMVEC). HLMVEC were stimulated with Hb over 1, 3, 6, and 24 h. A: VCAM-1 mRNA (n = 7–9 replicates) increased significantly at 3 and 6 h. Representative Western blot (B) and quantification of VCAM-1 (n = 7 replicates) (C) indicate significantly increased total VCAM-1 protein expression at 3 and 6 h. Representative images [Blue = nuclei, Red = VCAM-1; scale bar = 300 µm] (D) and quantification of VCAM-1 immunofluorescence (n = 9 replicates) (E) shows increased cell surface expression of VCAM-1 at 3, 6, and 24 h. Statistical analysis was performed using one-way ANOVA with Dunnett’s multiple comparisons. Graphs represent means ± SE.

Upregulation of VCAM-1 by Hb is Mediated through Toll-like Receptor 4 Signaling and Increased Reactive Oxygen Species

Based on previous reports that Toll-like receptor 4 (TLR4) signaling plays a role in mediating endothelial activation and upregulation of adhesion molecules in response to hemoglobin or heme (22, 23), we hypothesized that the upregulation of VCAM-1 in response to Hb stimulation observed in our study may depend on TLR4 signaling. To test this, we assessed upregulation of VCAM-1 by Hb after 3 h in the presence or absence of TLR4 inhibitor, TAK-242 (1 µM, 1-h pretreatment). Indeed, inhibition of TLR4 prevented increased VCAM-1 transcription (Fig. 6A) and total protein expression of VCAM-1 (Fig. 6, B and C). This was concordant with Hb-mediated increased activation of NF-κB (downstream of TLR4) that was also prevented by TLR4 inhibition (Fig. 6, B and D). In addition, increased generation of intracellular reactive oxygen species (ROS) induced by Hb stimulation was also prevented by inhibition of TLR4 (Fig. 6, E and F).

Figure 6.

Inhibition of Toll-like receptor 4 (TLR4) with TAK-242 (TAK) prevents Hb-mediated upregulation of vascular cell adhesion molecule-1 (VCAM-1), NF-κB activation, and reactive oxygen species (ROS) generation. HLMVEC were stimulated with Hb for 3 h with or without 1-h pretreatment with TAK-242 (1 µM). A: increased VCAM-1 mRNA (n = 3 or 4 replicates) in response to Hb was prevented by TLR4 inhibition. Representative Western blot (B) and quantification of VCAM-1 (n = 4 or 5 replicates) (C) and phospho-NF-κB (n = 3 or 4 replicates) (D) indicate VCAM-1 protein expression and NF-κB activation were prevented by TLR4 inhibition. Representative images [blue = nuclei, red = CellRox; scale bar = 300 µm] (E) and quantification of CellRox (AF647; MFI/cell, n = 5 replicates) (F) indicate Hb-mediated increase in ROS was prevented by TLR4 inhibition. Statistical analysis was performed using one-way ANOVA with Tukey’s multiple comparisons. Graphs represent means ± SE.

We also assessed whether nod-like receptor protein 3 (NLRP3) signaling (which can also be downstream of TLR4) played a role in upregulation of VCAM-1 in HLMVEC stimulated with Hb; however, inhibition of NLRP3 with MCC950 (10 µM, 1-h pretreatment) did not prevent VCAM-1 transcription or total protein expression, nor increased ROS (Supplemental Fig. S2; see https://doi.org/10.6084/m9.figshare.21692273).

To determine whether endogenous scavengers for Hb or the heme moiety would abrogate Hb-mediated upregulation of adhesion molecules, we assessed whether blocking Hb with haptoglobin or hemopexin (endogenous scavenger of hemoglobin or heme, respectively) would prevent upregulation of VCAM-1; however, neither haptoglobin nor hemopexin prevented VCAM-1 transcription or total protein expression (Supplemental Fig. S3; see https://doi.org/10.6084/m9.figshare.21692276).

DISCUSSION

In this study, we demonstrate that direct stimulation with Hb induces leukocyte-endothelial adhesion and activation of HLMVEC, characterized by secretion of inflammatory cytokines and upregulation of adhesion molecules to the cell surface, especially VCAM-1. VCAM-1 upregulation and concomitant increases in NF-κB activation and intracellular ROS generation were prevented by TLR4 inhibition. These data indicate that Hb plays a role in modulating the pulmonary microvascular endothelium to recruit leukocytes to areas of injury during inflammatory conditions through TLR4 signaling.

Stimulation of HLMVEC with Hb increased leukocyte adhesion to the endothelium. Moreover, we separated the potential effect of Hb on leukocyte activation by stimulating freshly isolated human leukocytes or primary HLMVEC with Hb separately or in combination. We found that only Hb stimulation of HLMVEC increased leukocyte adhesion to the endothelium, indicating that activation of the endothelium drives the adhesion of leukocytes to the HLMVEC monolayer.

Prior studies have reported that Hb can modulate adhesion molecule expression in other endothelial cell types. Treatment of human umbilical vein endothelial cells (HUVECs) with Hb increases expression of E-selectin, P-selectin, ICAM-1, and V-CAM-1 in vitro (23–25). In a model of murine sickle cell disease, Hb infusion increased leukocyte adhesion to the endothelium and led to an influx of leukocytes within liver, spleen, and pancreas (23, 24), but the effect of Hb on leukocyte influx in the lungs was not studied. In nonpulmonary organs, antibody blockade of endothelial cell adhesion molecules inhibited leukocyte adhesion to endothelium following Hb treatment (23). The current study demonstrates that Hb activates lung microvascular endothelium and increases adhesion molecule expression; however, our results suggest that the pattern of expression of these molecules may be different in the pulmonary circulation.

It is well known that leukocyte adhesion and migration within the pulmonary microvasculature is unique compared with other organ systems. Leukocyte behavior in the lung differs from other systems in several ways. First, while leukocytes freely circulate through other organs, in the lungs, there is a pre-existing pool of marginated neutrophils. In addition, the site of leukocyte transmigration differs, occurring within the pulmonary capillary system rather than the postcapillary venules (11). Furthermore, the hemodynamics and physical network of pulmonary capillaries lead to different requirements for adhesion molecules. The pulmonary capillaries are a complex structure of branching segments, whereas the diameters of most segments are less than half the size of a spherical neutrophil that makes rolling unfeasible (26–28). Selectin-mediated rolling has been shown to occur within the pulmonary venules, but only makes a small contribution to the total pool of marginated neutrophils (26). This may explain why we observed a significant change in E-selectin expression following exposure to Hb but did not observe a change in P-selectin expression or in total E-selectin protein levels.

As with selectin expression, the role of adhesion molecules within the lung microvascular endothelium differs from other vascular beds that largely rely on endothelial ICAM-1 binding of β2-integrins on leukocytes. In the pulmonary vasculature, ICAM-1 has a level of constitutive expression that is 30-fold higher compared with other endothelial beds; this has been postulated to be responsible for pulmonary leukocyte margination within the capillary beds at baseline (11, 29). In agreement with what others have shown, we found high baseline total expression of ICAM-1 that did not change with Hb exposure; however, Hb exposure increased cell surface localization of ICAM-1. Concomitantly, VCAM-1 gene expression, total protein expression, and cell surface localization increased significantly beginning at 3 h of Hb stimulation. This finding indicates that modulation of VCAM-1 may be the primary driver of leukocyte adhesion in HLMVEC in response to Hb.

Many signaling pathways have been implicated in the endothelial response to Hb (30). Others have demonstrated that in nonpulmonary human endothelium or in murine pulmonary vascular beds, endothelial activation induced by Hb is dependent on TLR4 signaling and activation of NF-κB to upregulate adhesion molecules and release of chemokines and cytokines (17, 22, 23). Our findings agree with these studies, as we demonstrate that upregulation of VCAM-1, activation of NF-κB, and generation of ROS are induced by Hb and prevented by TLR4 blockade. Interestingly, though NLRP3 inhibition slightly lowered ROS generation, it was not sufficient to block upregulation of VCAM-1. In addition, the endogenous scavengers of hemoglobin (haptoglobin) or heme (hemopexin) were not sufficient to prevent upregulation of VCAM-1 in response to Hb stimulation. It is unclear why these scavengers were unable to block the endothelial upregulation of VCAM-1. Although haptoglobin has successfully blocked the toxic effects of Hb in numerous physiological models, the specific biological processes involved in the ability of haptoglobin to protect endothelial damage have not been completely distinguished. One possibility for how haptoglobin protects endothelium from Hb-mediated damage is the ability of haptoglobin to bind hemoglobin and protect it from oxidation; our study used an already oxidized form of hemoglobin, which might explain why haptoglobin was ineffective. The inability of hemopexin to block this effect may imply that VCAM-1 upregulation is not dependent on heme being released from the full hemoglobin molecule.

Our study has some limitations. First, we studied endothelial cells under static conditions. How these molecules are regulated in response to Hb under flow conditions is a topic for future study. Furthermore, though we did identify specific signaling mechanisms involved in the regulation of VCAM-1, it is possible that other pathways mediate the upregulation of the other adhesion molecules we identified in our study such as E-selectin and ICAM-1. In addition, we did not investigate specific effects of Hb stimulation of leukocytes. These potential signaling mechanisms and how they contribute to regulation of endothelial activation and leukocyte adhesion will be important to investigate in future studies. Also, the hemoglobin used in this study was sourced commercially. Primarily methemoglobin, this lyophilized Hb has a higher potential for risk of dimerization, more rapid loss of heme, or damage. Future studies will utilize cell-free hemoglobin freshly isolated from red blood cells. Moreover, although these cell culture studies provide important mechanistic insight, these findings should be investigated in a physiological in vivo model of sepsis.

In summary, endothelial activation and injury are central to the pathogenesis of sepsis and sepsis-induced ARDS. Understanding the mechanisms by which this occurs will help to elucidate future pathways for treatment of patients. This study demonstrates that Hb plays a role in leukocyte adhesion and endothelial activation through release of cytokines and upregulation of adhesion molecules, especially VCAM-1, via TLR4 signaling, downstream activation of NF-κB, and ROS generation in primary HLMVEC.

DATA AVAILABILITY

Data will be made available upon reasonable request.

SUPPLEMENTAL DATA

Supplemental Fig. S1: https://doi.org/10.6084/m9.figshare.21692201.

Supplemental Fig. S2: https://doi.org/10.6084/m9.figshare.21692273.

Supplemental Fig. S3: https://doi.org/10.6084/m9.figshare.21692276.

GRANTS

This work was supported by NIH, K24HL103836 and R01HL158906 (to L. B. Ware), R01HL135849 (to L. B. Ware and J. A. Bastarache), R35HL150783 (to J. A. Bastarache), T32HL094296 (to J. E. Meegan), and Francis Family Foundation Parker B. Francis Fellowship (to J. E. Meegan).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

A.K.C., L.B.W., J.A.B., and J.E.M. conceived and designed research; A.K.C., T.T., K.J.R., J.S.D., L.E.B., and J.E.M. performed experiments; A.K.C., T.T., K.J.R., J.S.D., L.E.B., J.A.B., and J.E.M. analyzed data; A.K.C., K.J.R., L.B.W., J.A.B., and J.E.M. interpreted results of experiments; A.K.C. and J.E.M. prepared figures; A.K.C., K.J.R., and J.E.M. drafted manuscript; A.K.C., L.B.W., J.A.B., and J.E.M. edited and revised manuscript; A.K.C., T.T., K.J.R., J.S.D., L.E.B., L.B.W., J.A.B., and J.E.M. approved final version of manuscript.