Myeloid Poldip2 Contributes to the Development of Pulmonary Inflammation by Regulating Neutrophil Adhesion in a Murine Model of Acute Respiratory Distress Syndrome.

Background Lung injury, a severe adverse outcome of lipopolysaccharide-induced acute respiratory distress syndrome, is attributed to excessive neutrophil recruitment and effector response. Poldip2 (polymerase δ-interacting protein 2) plays a critical role in regulating endothelial permeability and leukocyte recruitment in acute inflammation. Thus, we hypothesized that myeloid Poldip2 is involved in neutrophil recruitment to inflamed lungs. Methods and Results After characterizing myeloid-specific Poldip2 knockout mice, we showed that at 18 hours post-lipopolysaccharide injection, bronchoalveolar lavage from myeloid Poldip2-deficient mice contained fewer inflammatory cells (8 [4-16] versus 29 [12-57]×104/mL in wild-type mice) and a smaller percentage of neutrophils (30% [28%-34%] versus 38% [33%-41%] in wild-type mice), while the main chemoattractants for neutrophils remained unaffected. In vitro, Poldip2-deficient neutrophils responded as well as wild-type neutrophils to inflammatory stimuli with respect to neutrophil extracellular trap formation, reactive oxygen species production, and induction of cytokines. However, neutrophil adherence to a tumor necrosis factor-α stimulated endothelial monolayer was inhibited by Poldip2 depletion (225 [115-272] wild-type [myePoldip2+/+] versus 133 [62-178] myeloid-specific Poldip2 knockout [myePoldip2-/-] neutrophils) as was transmigration (1.7 [1.3-2.1] versus 1.1 [1.0-1.4] relative to baseline transmigration). To determine the underlying mechanism, we examined the surface expression of β2-integrin, its binding to soluble intercellular adhesion molecule 1, and Pyk2 phosphorylation. Surface expression of β2-integrins was not affected by Poldip2 deletion, whereas β2-integrins and Pyk2 were less activated in Poldip2-deficient neutrophils. Conclusions These results suggest that myeloid Poldip2 is involved in β2-integrin activation during the inflammatory response, which in turn mediates neutrophil-to-endothelium adhesion in lipopolysaccharide-induced acute respiratory distress syndrome.

chemokine (C‐X‐C motif) ligand 1

chemokine (C‐X‐C motif) ligand 2

N‐Formylmethionyl‐leucyl‐phenylalanine

intercellular adhesion molecule 1

mouse lung microvascular endothelial cell

neutrophil extracellular traps

phorbol 12‐myristate 13‐acetate

reactive oxygen species

Clinical Perspective

What Is New?

Myeloid‐specific knockout of Poldip2 (polymerase δ‐interacting protein 2) inhibits leukocyte recruitment but does not alter induction of chemokines in lung tissue during lipopolysaccharide‐induced acute respiratory distress syndrome.

Knockout of Poldip2 in neutrophils does not alter effector function or motility.

Primed neutrophils with Poldip2 deficiency are less adhesive to endothelial cells and undergo less transmigration, likely because of reduced activation of β2‐integrins.

What Are the Clinical Implications?

Impaired neutrophil infiltration in acute respiratory distress syndrome can be protective against lung edema and injury.

Deficiency of myeloid Poldip2 has the potential to decrease the excessive neutrophil extravasation in sepsis‐induced acute respiratory distress syndrome without affecting pathogen killing capacity.

Acute respiratory distress syndrome (ARDS) is a devastating disease characterized by acute hypoxemia and non‐cardiogenic pulmonary edema that often requires mechanical ventilation. ARDS can be caused by a variety of pulmonary (eg, pneumonia) or non‐pulmonary (eg, sepsis, trauma, pancreatitis) insults, and therapeutic strategies are limited to the treatment of the underlying disease in conjunction with mechanical ventilation, which itself can further damage the lungs, causing ventilator‐associated lung injury. , ARDS is characterized by increased pulmonary vascular permeability, lung edema, and loss of aerated lung tissue, as a result of an intense inflammatory response in alveoli with activation of endothelial cells and procoagulant processes, leukocyte and protein infiltration, and innate immune cell–mediated damage of the alveolar‐capillary barriers. Such formation of alveolar protein‐rich exudate is a major factor for hypoxemia and one of the earliest events that define ARDS.

Neutrophils are one of the major immune cell types found in acute lung injury. In addition to their role in the elimination of local pathogens, neutrophils can also act as a source of alveolar injury by releasing proinflammatory mediators, neutrophil extracellular traps (NETs) and reactive oxygen species (ROS). An imbalance between effective immune activation and excessive neutrophil extravasation contributes to worsening lung injury. Neutrophils enter the lung by first adhering to the activated endothelium and then exiting the microvasculature. Firm adhesion of neutrophils to the inflamed site is mediated by β2‐integrin, a leukocyte‐specific integrin that binds to an α subunit and interacts with intercellular adhesion molecule‐1 (ICAM‐1) on the endothelium , during leukocyte recruitment to inflamed sites. As a well‐known regulator of cell‐to‐cell adhesion, β2‐integrin‐dependent neutrophil adhesion in acute lung injury occurs more frequently in Pseudomonas aeruginosa‐, immunoglobulin G immune complex‐ and lipopolysaccharide‐induced ARDS , , compared with other stimuli.

We have recently focused on Poldip2 (polymerase‐δ interacting protein 2) as a novel mediator of ARDS. Poldip2 is an omnipresent protein expressed in most cell types. First discovered as a DNA polymerase‐δ interacting protein and a binding partner for the proliferating cell nuclear antigen, , Poldip2 is also involved in the regulation of a number of physiological processes such as cell cycle progression, mitochondrial function, extracellular matrix deposition, focal adhesion turnover, and cell metabolism. , , , , In our previous studies, we showed that heterozygous Poldip2‐deficient mice exhibited alleviated mortality, reduced vascular permeability, and abrogated inflammatory cytokine induction and leukocyte infiltration in both hypoxia‐induced cerebral ischemia and lipopolysaccharide‐induced acute lung injury. We focused on endothelial and astrocyte Poldip2 in these reports; however, little is known about the function of Poldip2 in immune cells. In this study we used for the first time a myeloid‐specific Poldip2 knockout (myePoldip2‐/‐) murine model to determine whether Poldip2 in myeloid cells, especially neutrophils, plays a role in the development of lipopolysaccharide‐induced ARDS in mice. We hypothesized that loss of myeloid Poldip2 contributes to the protective phenotype of Poldip2 depletion in ARDS and investigated its potential role in neutrophil adhesion and function. We show that a major function of Poldip2 in activated neutrophils is to regulate β2‐integrin‐mediated adhesion to endothelial cells, rather than other effector functions during ARDS.

Methods

The data that support the findings of this study are available from the corresponding author upon reasonable request. A detailed Materials and Methods section can be found in Data S1 and Table S1.

Animals

Poldip2 gene trap mice on a C57BL/6 background were previously described. Poldip2 myeloid‐specific knockout mice were generated by crossing a LysM‐Cre strain (C57BL/6 background) with our newly created Poldip2 floxed mice (C57BL/6 background). Mice with Poldip2 deficiency in myeloid cells are hereafter designated myePoldip2‐/‐, while their littermates without Poldip2 deficiency are designated myePoldip2+/+. Hematopoiesis was evaluated by assessing complete blood counts from cardiac puncture blood samples. All animal experiments were conducted with the approval of the Institutional Animal Care and Use Committee at Emory University.

Lipopolysaccharide‐Induced ARDS Model

Adult male and female (age 2.5–3.5 months) myePoldip2+/+ and myePoldip2‐/‐ mice were randomly divided into control and lipopolysaccharide groups. Animals in the lipopolysaccharide group received an intraperitoneal injection of lipopolysaccharide (18 mg/kg) diluted in sterile normal saline, while mice in the control group were given sterile normal saline. Eighteen hours after injection, rectal temperature was measured, and mice were euthanized by CO2 asphyxiation for either bronchoalveolar lavage (BAL) or lung tissue collection. No difference between sexes was detected.

Bronchoalveolar Lavage

Mice were first euthanized by CO2 inhalation and tracheas were exposed and cannulated using a 20‐gauge lavage needle. For assessment of total cell counts in BAL, 3 lavages with 1 mL each of PBS containing 2 mmol/L EDTA were injected in the tracheal lavage needle and recovered as previously described. The volume recovered was measured for normalization and then centrifuged at 300g, 4 °C for 10 minutes. Pellets were resuspended in Hanks balanced salt solution without Ca2+ and Mg2+ for either total BAL cell counts or flow cytometry. For total cell counts, pellets were processed with red blood cell (RBC) lysis buffer before counting using an automated cell counter.

Flow Cytometry

For neutrophil identification in BAL, samples were processed with RBC lysis buffer and spun down at 300 g for 5 minutes at room temperature. The cell pellet was resuspended, and cells were blocked with anti‐mouse CD16/32, followed by incubation with anti‐mouse antibodies. For evaluation of integrin surface expression, isolated neutrophils were labeled with eFluor450‐CD11b or FITC‐CD18. For evaluation of L‐selectin shedding, neutrophils were stained with FITC‐CD62L. Data were analyzed with Flowjo software, and leukocytes were gated as CD45+; myeloid cells were further gated as CD11b+ and neutrophils were identified as Ly6G+. For complete blood cell counts, white blood cell differential counting was achieved using a Hemavet 1500 blood analyzer (CDC Technologies, Oxford, CT).

Immunofluorescence, Histology, and Microscopy

Eighteen hours after PBS or lipopolysaccharide treatment, mice were euthanized using CO2 asphyxiation, and lungs were injected with 10% formalin through the trachea. Lungs were then dissected and prepared for immunohistochemistry or immunofluorescence. Briefly, lungs were prepared for hematoxylin and eosin staining according to standard protocols. Pictures were taken under a NanoZoomer‐SQ Digital slide scanner (Hamamatsu) at 20× magnification. For immunostaining, sections were incubated with primary antibody against Ly6G and then stained with Alexa Fluor 568‐conjugated secondary antibody and mounted with VECTASHIELD mounting medium with 4′,6‐diamidino‐2‐phenylindole. Pictures were taken on a Zeiss LSM 800 Airyscan microscope at 20× magnification, and the ratio of Ly6G to 4′,6‐diamidino‐2‐phenylindole positive area was calculated using ImageJ for assessment of neutrophil infiltration. Data were quantified from 3 fields of 1 section per animal, and 5 animals per group.

Bone Marrow Neutrophil Isolation

Primary bone marrow neutrophils were isolated as described and purified by the immunomagnetic negative selection technique, using an Easysep Mouse Neutrophil Enrichment Kit (Stemcell Technologies, #19762). Neutrophil purity was always >85% and was assessed by staining with the fluorophore‐labeled neutrophil marker, Ly6G, followed by flow cytometry analysis.

Cell Culture

Primary rat pulmonary microvascular endothelial cells (Cell Biologics, #RN‐6011) and primary mouse lung microvascular endothelial cells (MLMECs; Cell Biologics, #C57‐6011) were cultured on plates pre‐coated with 0.1% gelatin. Endothelial cell medium was supplemented with 2% fetal bovine serum, endothelial cell growth factors, and antibiotics. Cells were used between passages 4–6.

Static Adhesion Assay

Isolated bone marrow neutrophils were stained with Hoechst and resuspended at 1×106 cells/mL. MLMECs were seeded onto a 24‐well plate and treated with 10 ng/mL tumor necrosis factor‐α (TNF‐α) once a monolayer was formed. After 6 hours of treatment, MLMECs were washed followed by addition of bone marrow neutrophils (2×105 cell per well). Neutrophils and MLMECs were coincubated for 30 minutes and then fixed using 3.7% paraformaldehyde. Fixed cells were washed and immediately imaged with an Olympus IX71 inverted fluorescent microscope using the 4′,6‐diamidino‐2‐phenylindole fluorescence channel at 10× magnification. Three representative pictures were taken per well, and for each condition, triplicate wells were used. Images were analyzed with ImageJ software (National Institutes of Health); stained neutrophil nuclei were counted using the “analyze particles” module to assess firm adhesion to the MLMEC monolayer.

μ‐slide Chemotaxis

µ‐Slide Chemotaxis (ibidi GmbH, #80326) was used for the chemotaxis assay. Each slide contains 3 chambers, and each chamber consists of 1 channel for the cells and 2 reservoirs for the chemoattractant or chemoattractant‐free media on either side of the channel. Hoechst 33342‐labeled Poldip2+/+ and Poldip2+/‐ bone marrow neutrophils were seeded into the central channel and exposed to N‐Formylmethionyl‐leucyl‐phenylalanine (fMLP) (50 μmol/L) or medium in the 2 side channels. Negative and positive control experiments were performed in the other 2 chambers of the same μ‐slide. Cell movement was observed under a Leica TCS SP5 II laser scanning confocal microscope (Leica Microsystems CMS GmbH, Wetzlar, Germany). Time lapse videos were acquired using Leica Application Suite Advanced Fluorescence software with a Plan‐Neo 10×0.3NA air objective every minute for a total duration of 2 hours in brightfield mode and 4′,6‐diamidino‐2‐phenylindole mode. Cell trajectories were analyzed using customized Python cell tracker software. Distance threshold was set as 30 μm based on the cell trajectories of negative control experiments. Statistical significance (P<0.05) was calculated from 3 independent experiments (n=3) in which over 60 cells were evaluated in each condition per experiment.

Transmigration Assay

For the transmigration assay, 3.0 μm pore size transwell membrane inserts were used. MLMECs (3×105) were seeded onto the membrane and after forming a monolayer at 24 hours, they were treated with 10 ng/mL TNF‐α or media alone for 6 hours. Hoechst prestained neutrophils (0.5×106) were added to the upper chamber of each insert and 100 nmol/L fMLP or media alone was added to the lower chamber. After 2 hours of incubation at 37°C, non‐migrated cells were removed from the upper surface using a wet cotton swab and migrated cells on the lower surface were fixed in 4% paraformaldehyde for 3 minutes. Inserts were then washed 3 times with PBS and membranes were removed and mounted on slides. Pictures were taken using an Olympus DP71 Digital Microscope at 10× and migrated neutrophils were counted by a masked observer in 6 fields for each well using ImageJ (National Institutes of Health software). Neutrophils that transmigrated into the media were also collected and counted manually using a hemocytometer.

Soluble ICAM‐1 Binding Assay

The soluble ICAM‐1 binding assay was used to assess beta integrin activation in neutrophils. Bone marrow cells isolated from myePoldip2+/+ and myePoldip2‐/‐ mice were exposed to 2 mmol/L EDTA (as negative control), 5 mmol/L MnCl2 or an equal volume of HBSS+, in the presence of 10 μg/mL recombinant mouse ICAM‐1‐Fc chimera (R&D Systems, # 796‐IC‐050) and 10 μg/mL phycoerythrin (PE)‐conjugated anti‐human immunoglobulin G1 Fc (ThermoFisher, #12‐4998‐82). Cells were fixed and then labeled with allophycocyanine‐conjugated anti‐Ly6G to identify neutrophils. ICAM‐1 binding was measured using flow cytometry assessment of PE mean fluorescence intensity in the neutrophil subset.

SYTOX Green Assay

The SYTOX green assay was used to quantify the abundance of extracellular DNA as a surrogate of NET formation as previously described. Poldip2+/+ and Poldip2+/‐ neutrophils were isolated as described above. Fifty thousand cells per well were plated in a 96‐well black clear‐bottom plate and incubated at 37°C for 1 hour. Cells were then stimulated with 324 nmol of phorbol 12‐myristate 13‐acetate or 50 μg/mL of lipopolysaccharide. SYTOX green dye was added to each well and the fluorescence was read every 15 minutes for a total of 90 minutes at 37°C. Four replicates were measured in each of 3 independent experiments. Both the time course curves as well as final fluorescence at 90 minutes were analyzed.

Reactive Oxygen Species Production

ROS production was measured using the cytochrome C assay in bone marrow neutrophils isolated from Poldip2+/+ and Poldip2+/‐ mice. Briefly, cytochrome C (100 nmol/L) was added to all wells and 25 units of superoxide dismutase was added to control wells. Neutrophils were stimulated with 100 nmol/L phorbol 12‐myristate 13‐acetate (PMA). Absorbance was read at 550 nm (absorbance of reduced cytochrome C) and 490 nm (to control for non‐specific absorbance) every minute for 2 hours immediately after application of phorbol 12‐myristate 13‐acetate. The ROS production of 3 independent experiments (2‒6 independent wells each) was calculated.

RNA Extraction and Quantitative Reverse Transcription Polymerase Chain Reaction

Total RNA was purified with Qiazol (Qiagen, #79306) and the RNeasy Plus kit (Qiagen, #74104). Reverse transcription was performed using Protoscript reverse transcriptase (New England Biolabs) with random primers. The resulting cDNA was amplified with previously validated primers (Table). RPL (ribosomal protein L13A) and hypoxanthine guanine phosphoribosyl transferase were used as housekeeping genes as their expression was not affected by lipopolysaccharide treatment. Note that Poldip2 primers can detect messages transcribed from both floxed and Cre‐excised alleles. Amplification was performed in 96‐well plates using Forget‐Me‐Not EvaGreen qPCR Master Mix with low ROX (Biotium, #31045) in a QuantStudio 7 instrument (Invitrogen). Data analysis was performed using the mak3i module of the qpcR software library (version 1.4‐0) , in the R‐environment.

Table 1

Mouse Primer Sequences for RT‐qPCR Assays

Annealing At 55 ℃	Primer Sequences 5ʹ‐3ʹ
Gene	Forward	Reverse
RPL	ATGACAAGAAAAAGCGGATG	CTTTTCTGCCTGTTTCCGTA
HPRT	GCTGACCTGCTGGATTACAT	GGTCCTTTTCACCAGCAAGCT
IL‐1β	ACCAAGCAACGACAAAATAC	CACTTTGCTCTTGACTTCTATC
IL‐6	CTACCCCAATTTCCAATGCT	ACCACAGTGAGGAATGTCCA
TNF‐α	CTATGTCTCAGCCTCTTCTC	GGCCATTTGGGAACTTCTCA
CCL‐2/MCP‐1	CAAGATGATCCCAATGAGTAG	CAGATTTACGGGTCAACTTC
CXCL‐1	AAAGATGCTAAAAGGTGTCC	GTATAGTGTTGTCAGAAGCC
CXCL‐2	GTTGACTTCAAGAACATCCAG	CTTTCTCTTTGGTTCTTCCG
Poldip2	GAGACCACCGAGAACATCCG	GTGGGAATTCTGGGCTTCCCTC′

CCL‐2 (also known as MCP‐1 [monocyte chemoattractant protein 1]) indicates chemokine (C‐C motif) ligand 2; CXCL‐1, chemokine (C‐X‐C motif) ligand 1; CXCL‐2, chemokine (C‐X‐C motif) ligand 2; HPRT, hypoxanthine guanine phosphoribosyl transferase; IL‐1β, interleukin‐1β; IL‐6, interleukin‐6; Poldip2, polymerase (DNA‐directed) δ‐interacting protein 2; RPL, ribosomal protein L13A; RT‐qPCR, quantitative reverse transcription polymerase chain reaction; and TNF‐α, tumor necrosis factor‐α.

Enzyme‐Linked Immunosorbent Assay

Isolated bone marrow neutrophils (3×106) were stimulated with lipopolysaccharide (1 μg/mL) or sterile Hanks Balanced Salt Solution with calcium and magnesium (HBSS+) for 8 hours and supernatant was then collected for TNF‐α, interleukin‐1β, and interleukin‐6 production measurement. Commercial ELISA kits (R&D Systems) were used according to the manufacturer’s instructions. Five independent experiments were performed, and 2 replicates were measured in each experiment.

Western Blotting

Whole cell lysate was prepared from isolated neutrophils using a lysis buffer described in our previous study. For Pyk2 phosphorylation detection, bone marrow neutrophils were plated onto ICAM‐1 coated 6‐well plates and stimulated with TNF‐α or media alone for the specified time. Neutrophil lysates were then prepared by adding 2× Laemlli buffer and boiling for 5 minutes. Samples were stored at −20 °C until gel loading. Proteins were separated by SDS‐PAGE, transferred to nitrocellulose membranes, and assessed by blotting with primary antibodies against Poldip2, Pyk2, phosphorylated Pyk2, β‐actin and vinculin. Blots were incubated with horseradish peroxidase‐conjugated secondary antibodies. Detection was performed using enhanced chemiluminescence and densitometry was performed using ImageJ.

Statistical Analysis

Analyses were performed using results from 3 to 10 independent experiments with GraphPad Prism software version 8. Data are represented as medians with their 95% CIs and analyzed using non‐parametric methods, either Mann Whitney or Kruskal‐Wallis with Dunnett multiple comparisons. A threshold of P<0.05 was considered significant. Micrographs showing average numbers of adherent cells were selected as representative.

Results

Characterization of Myeloid‐Specific Poldip2 Knockout Mice

To study the effect of myeloid ablation of Poldip2 in lipopolysaccharide‐induced ARDS in vivo, we established myeloid‐specific Poldip2 knockout (myePoldip2‐/‐) mice by crossing a LysM‐Cre strain with Poldip2 floxed mice. To verify the efficiency and specificity of Poldip2 deletion, we measured the expression of Poldip2 in different organs and cells by quantitative real‐time polymerase chain reaction. While Poldip2 mRNA was decreased by 50% in purified neutrophils from myePoldip2‐/‐ mice compared with their littermates, Poldip2 levels remained unchanged in hearts and lungs isolated from myePoldip2‐/‐ compared with myePoldip2+/+ mice (Figure 1A). Reduction of Poldip2 protein expression in purified neutrophils was further confirmed by Western blotting (Figure 1B). To detect any potential effect of Poldip2 myeloid‐specific deletion on the composition of circulating leukocytes, we performed complete blood cell counts using a blood analyzer, and neither total leukocyte counts nor white blood cell differential counts were affected (Figure 1C). These results suggest that Poldip2 was specifically knocked out in myeloid cells from myePoldip2‐/‐ mice, and that knockout efficiency was similar to Poldip2 gene trap heterozygotes (Poldip2+/‐) (Figure S1). Moreover, since no change in the circulating leukocyte population was noted, myePoldip2‐/‐ mice are expected to possess the same numbers of defensive immune cells as wild type mice in response to external insult.

Figure 1

Characterization of myeloid‐specific Poldip2 (polymerase [DNA‐directed] δ‐interacting protein 2) knockout mice.

A, Poldip2 mRNA expression (relative to myePoldip2+/+, normalized with RPL and HPRT) in purified bone marrow neutrophils, perfused hearts and lungs, measured by quantitative reverse transcription polymerase chain reaction. Data represent medians with 95% CIs (n=5); **P<0.01 (Mann‐Whitney test). B, Poldip2 protein expression (relative to wild‐type and normalized with vinculin) in myePoldip2+/+ and myePoldip2‐/‐ neutrophils. Data represent medians with 95% CIs (n=8); **P<0.01 (Mann‐Whitney test). C, Differential counts of circulating leukocytes. Blood collected from myePoldip2+/+ and myePoldip2‐/‐ mice was analyzed for absolute number of white blood cells, lymphocytes, monocytes, and granulocytes. Data represent medians with 95% CIs (n=5). No difference was observed between genotypes. GRAN indicates granulocytes; HPRT, hypoxanthine guanine phosphoribosyl transferase; LYM, lymphocytes; MONO, monocytes; myePoldip2+/+, wild type; myePoldip2‐/‐, myeloid‐specific Poldip2 knockout; Poldip2, polymerase (DNA‐directed) δ‐interacting protein 2; RPL, ribosomal protein L13A; and WBC, white blood cells.

Myeloid Specific Deletion of Poldip2 Protects Against Neutrophil Infiltration in Lipopolysaccharide‐Induced ARDS

As reported in our previous research, whole body heterozygous deletion of Poldip2 decreased lipopolysaccharide‐induced mortality and immune cell recruitment into the lung. Here, we sought to explore the contribution of myeloid Poldip2 in pulmonary leukocyte infiltration after lipopolysaccharide‐induced ARDS. The development of lipopolysaccharide‐induced sepsis was first confirmed by measuring rectal temperature before euthanasia. Both myePoldip2+/+ and myePoldip2‐/‐ mice experienced equivalent body temperature drops after lipopolysaccharide treatment indicating a similar overall response to lipopolysaccharide (Figure 2A). Leukocyte recruitment into the lung was then determined by counting cells in BAL fluid recovered 18 hours after lipopolysaccharide injection, a time point where significant accumulation of neutrophils occurs. Consistent with our results in Poldip2 gene trap heterozygotes, lipopolysaccharide induced a remarkable increase of cell infiltration in myePoldip2+/+ mice, which was not observed in myePoldip2‐/‐ mice (Figure 2B). Pulmonary infiltrates in response to lipopolysaccharide in wild‐type mice, but not in myePoldip2‐/‐ mice were further confirmed by hematoxylin and eosin staining (Figure 2C). Given that our previous study indicated that the invading myeloid cell population at 18 hours following lipopolysaccharide‐induced ARDS was mostly neutrophils, we performed flow cytometry to investigate the proportion of neutrophils in BAL leukocytes. As shown in Figure 2D and 2E, the percentage of infiltrated neutrophils was significantly reduced in myePoldip2‐/‐ mice after 18 hours lipopolysaccharide treatment compared with myePoldip2+/+ mice. Neutrophil infiltration into lungs was further examined by staining lungs with Ly6G, a neutrophil‐specific marker. Consistent with our flow cytometry results, our immunofluorescence data suggest that the number of lung‐sequestered neutrophils was significantly increased in myePoldip2+/+ mice after lipopolysaccharide and abrogated in myePoldip2‐/‐ mice (Figure 2F and 2G).

Figure 2

Myeloid‐specific Poldip2 (polymerase [DNA‐directed] δ‐interacting protein 2) knockdown reduces neutrophil pulmonary infiltration in lipopolysaccharide‐induced acute respiratory distress syndrome.

A, Core body temperature measured 18 hours after injection of PBS or lipopolysaccharide (18 mg/kg). Similar drops in temperature show that both genotypes responded equally to lipopolysaccharide. Data represent medians with 95% CIs (n=5); **P<0.01 compared with PBS of the same genotype (Mann Whitney tests). B, Total cell counts were measured in bronchoalveolar lavage collected 18 hours after PBS or lipopolysaccharide injection. Data represent medians with 95% CIs (n=10); ***P<0.001 compared with myePoldip2+/+ PBS group, # P<0.05 compared with myePoldip2+/+ lipopolysaccharide group (Kruskal‐Wallis with Dunn multiple comparisons test). C, MyePoldip2+/+ and myePoldip2‐/‐ mice were given intraperitoneal PBS or lipopolysaccharide as described above, and lungs were harvested for hematoxylin and eosin staining. Representative pictures show decreased interstitial edema and cell infiltration in lungs after lipopolysaccharide administration in myePoldip2‐/‐ compared with myePoldip2+/+ (n=5). D and E, Following administration of lipopolysaccharide as above, bronchoalveolar lavage cells were labeled with CD45, CD11b, and Ly6G antibodies for neutrophil identification by flow cytometry. Panel D shows the gating strategy in which leukocytes were first gated as CD45+ (left) and neutrophils were further gated as CD11b+Ly6G+ (right). The percentages of neutrophils in bronchoalveolar lavage leukocytes after lipopolysaccharide injection were quantified (E). Data represent medians with 95% CIs (n=6 in myePoldip2+/+ and n=5 in myePoldip2‐/‐); **P<0.01 (Mann‐Whitney test). F and G. Lungs were harvested 18 hours post PBS or lipopolysaccharide injection for immunofluorescence staining. Panel F shows neutrophil pulmonary infiltration after staining with the neutrophil‐specific marker Ly6G, which was quantified in panel G. Data represent medians with 95% CIs (n=5); **P<0.01 compared with myePoldip2+/+ PBS group, # P<0.05 compared with myePoldip2+/+ lipopolysaccharide group (Mann‐Whitney tests). BAL indicates bronchoalveolar lavage; CD11b, integrin alpha M; CD45, leukocyte common antigen; DAPI, 4′,6‐diamidino‐2‐phenylindole; Ly6G; lymphocyte antigen 6 complex, locus G; myePoldip2+/+, wild type; myePoldip2‐/‐, myeloid‐specific Poldip2 knockout; Poldip2, polymerase (DNA‐directed) δ‐interacting protein 2; and SSC‐H, side scatter‐height.

CXCL‐1/CXCL‐2 Induction In Vivo is Independent of Myeloid Poldip2 Deletion

Because leukocyte recruitment to inflamed sites is dependent upon chemokine production by endothelium, , we investigated the induction of major chemoattractants for neutrophils in response to lipopolysaccharide treatment in vivo by measuring their mRNA levels in lung tissue. As indicated in Figure S2, C‐X‐C motif ligand 1 and 2 (CXCL‐1 and CXCL‐2) were dramatically increased 18 hours post‐lipopolysaccharide injection, in both myePoldip2+/+ mice and myePoldip2‐/‐ mice, but no significant difference was noted between the 2 genotypes. These findings indicate that CXCL‐1/CXCL‐2 production in myePoldip2‐/‐ mice is normal, suggesting that the primary driving force behind neutrophil recruitment to inflamed lung is not affected by depletion of Poldip2 in myeloid cells, but there is still a possibility that other chemokines in the milieu may contribute.

Neutrophils Exhibit Normal Effector Function Ex Vivo Despite Poldip2 Reduction

Since our data show that myeloid‐specific deletion of Poldip2 decreased lipopolysaccharide‐induced neutrophil infiltration into the lungs, we next asked if Poldip2 deletion affects neutrophil function per se. Specifically, we looked at induction of inflammatory cytokines and chemokines, as well as pathogen elimination using freshly isolated bone marrow‐derived neutrophils. Considering that myePoldip2‐/‐ neutrophils have a similar knockout efficiency as Poldip2+/‐ mice (compare Figure S1 to Figure 1), we utilized Poldip2+/‐ neutrophils as a more readily available resource for in vitro experiments. To examine the induction of inflammatory markers in response to proinflammatory stimuli, neutrophils were stimulated with lipopolysaccharide for 2 hours or 8 hours before assessing inflammatory marker mRNA or protein levels. As shown in Figure 3A through 3D, inflammation‐related cytokines such as TNF‐α and interleukin‐1β, and neutrophil‐secreted chemokine CXCL‐2 (which has been shown to be essential for diapedesis) were increased to the same degree after lipopolysaccharide stimulation in both Poldip2+/+ and Poldip2+/‐ neutrophils, with the exception of interleukin‐6, which was increased to a slightly lesser extent in Poldip2+/‐. The concentration of neutrophil‐released TNF‐α, interleukin‐1β and interleukin‐6 in the supernatant were further measured using ELISA (Figure S3) and none of them were different between genotypes. To ascertain the ability to eliminate pathogens, we measured NETs and ROS production in Poldip2+/+ and Poldip2+/‐ neutrophils in response to either PMA or lipopolysaccharide (Figure 3E and 3F and Figure S4). Poldip2+/‐ neutrophils functioned as well as Poldip2+/+ neutrophils in terms of releasing pathogen killing mediators. Taken together, these data indicate that neutrophil effector functions were not disrupted by Poldip2 reduction.

Figure 3

Poldip2 (polymerase [DNA‐directed] δ‐interacting protein 2) knockdown has little effect on neutrophil effector function.

A through D. Isolated bone marrow neutrophils were stimulated with 1 μg/mL lipopolysaccharide for 2 hours, and mRNA expression of inflammatory markers was measured by quantitative reverse transcription polymerase chain reaction. Lipopolysaccharide induced a significant increase of TNF‐α, IL‐1β, IL‐6 and CXCL‐2 mRNA in both Poldip2+/+ and Poldip2+/‐ neutrophils. No significant difference between genotypes was noted, except that IL‐6 was slightly less upregulated in Poldip2+/‐ compared with Poldip2+/+ neutrophils. Bars represent medians with 95% CIs (n=4 for TNF‐α, IL‐1β, and xIL‐6 and n=5 for CXCL‐2); *P<0.05, **P<0.01, compared with respective PBS group; # P<0.05, compared with Poldip2+/+ lipopolysaccharide (Mann‐Whitney tests). E, Isolated bone marrow neutrophils were stimulated with either PMA (324 nM) or lipopolysaccharide (50 μg/mL) before measuring neutrophil extracellular trap formation with the SYTOX green assay. Lipopolysaccharide induced a significant increase of neutrophil extracellular trap (NET) formation in both genotypes. Bars represent final fluorescence intensity at 90 minutes relative to control as medians with 95% CIs (n=3); *P<0.05, **P<0.01, compared with respective control groups (Kruskal‐Wallis with Dunn multiple comparisons test). No significant difference of NETs formation was observed between Poldip2+/+ and Poldip2+/‐ neutrophils. F, Time course of superoxide production. Neutrophils were stimulated with PMA (324 nM) and absorbance was measured every minute for 2 hours. The concentration of superoxide was calculated as described in Methods. Data represent mean ± SEM (n=3). No significant difference in superoxide formation was observed at any time point between Poldip2+/+ and Poldip2+/‐ neutrophils. CXCL‐2 indicates chemokine (C‐X‐C motif) ligand 2; IL‐1β interleukin‐1β; IL‐6, interleukin‐6; NETosis, NET formation; PMA, phorbol 12‐myristate 13‐acetate; Poldip2 polymerase (DNA‐directed) δ‐interacting protein 2; and TNF‐α, tumor necrosis factor‐α.

Neutrophil Motility Remains Unaffected After Poldip2 Knockdown

Because mediators released by neutrophils remained intact, we next investigated whether Poldip2 would affect the ability of neutrophils to respond to chemotactic stimuli. Using a live cell imaging system, neutrophils isolated from Poldip2+/+ and Poldip2+/‐ mice were placed in a channel with the general chemoattractant fMLP on 1 side, and their movement was recorded for motility analysis. As shown in Figure S5A, neutrophil speed was not affected, and total distance traveled was about the same in both genotypes (Figure S5B and Figure S6). Directional movement induced by fMLP was assessed by calculating the forward movement index, defined as the fraction of distance traveled toward fMLP. In accord with the locomotion data, neutrophil directional movement was independent of Poldip2 knockdown (Figure S5C). Altogether, these results indicated that neutrophil motility is preserved after Poldip2 depletion and that the diminished neutrophil recruitment in inflamed lung must be due to other factors.

Poldip2 Mediates Neutrophil Adhesion to Pulmonary Endothelial Monolayer in Part Via Activation of β2 Integrins

Leukocyte recruitment into inflamed tissue proceeds in a cascading fashion, and a pivotal step during this process is firm adhesion of neutrophils to stimulated endothelium. , Thus, we next sought to examine neutrophil adhesion to a pulmonary endothelial monolayer. Rat pulmonary microvascular endothelial cells were stimulated with TNF‐α, one of the main effectors of lipopolysaccharide, for 6 hours before seeding neutrophils. After a 30 minute coincubation, firmly adhered neutrophils were quantified. Downregulation of Poldip2 in neutrophils isolated from myePoldip2‐/‐ mice significantly attenuated their adhesion to TNF‐α‒stimulated pulmonary endothelial monolayers (Figure S7). This finding was further confirmed using monolayers of MLMECs (Figure 4A and 4B). Once neutrophils adhere to endothelial cells, they migrate through the monolayer. To mimic the inflammatory microenvironment, we used a transwell system in which the insert was coated with MLMECs and stimulated with TNF‐α and fMLP in the lower chamber to act as a chemoattractant. In accordance with our adhesion data, myePoldip2+/+ neutrophils exhibited enhanced transmigration under inflammatory conditions, which was diminished in myePoldip2‐/‐ neutrophils (Figure 4C through 4E).

Figure 4

Reduced adhesion and transendothelial migration in myePoldip2‐/‐ neutrophils.

A through B, Adhesion of Hoechst‐labeled neutrophils to a TNF‐α‒stimulated mouse lung microvascular endothelial cell monolayer. TNF‐α‐induced adhesion was abrogated in neutrophils isolated from bone marrow of myePoldip2‐/‐, compared with myePoldip2+/+ mice. Bars represent medians with 95% CIs (n=5); **P<0.01 compared with myePoldip2+/+ control; # P<0.05 compared with myePoldip2+/+ TNF‐α (Mann‐Whitney tests, 1‐tailed). C, Representative images of neutrophils at the bottom of transwell inserts membrane following transmigration assays. D and E, Quantification of transmigrated neutrophils on the bottom of membrane (D) and in the lower chamber media (E). MyePoldip2+/+ neutrophils showed enhanced transmigration when the endothelial monolayer was stimulated with TNF‐α and 100nM fMLP was added in the lower chamber, which was not observed in myePoldip2‐/‐ neutrophils. Data are presented as fold change relative to myePoldip2+/+ baseline transmigration and bars represent medians with 95% CIs (n=5); **P<0.01 compared with myePoldip2+/+ without TNF‐α and fMLP, # P<0.05, ## P<0.01 compared with myePoldip2+/+ with TNF‐α and fMLP (Mann‐Whitney tests, 1‐tailed). fMLP indicates N‐formylmethionyl‐leucyl‐phenylalanine; MLMEC, mouse lung microvascular endothelial cell; myePoldip2+/+, wild type; myePoldip2‐/‐, myeloid‐specific Poldip2 knockout; Poldip2, polymerase (DNA‐directed) δ‐interacting protein 2; and TNF‐α, tumor necrosis factor α.

To study the underlying mechanism of impaired adhesion and transmigration, we measured the surface expression of β2 integrin, which mediates neutrophil firm adhesion. As expected from other studies, TNF‐α triggered upregulation of both the α‐ and β‐subunits of β2 integrin surface expression (Figure 5A); however, the increased expression of integrins was similar in myePoldip2‐/‐ and myePoldip2+/+ neutrophils. In contrast, when we assessed the activation of β2 integrin by using a soluble ICAM‐1 binding assay, we found that TNF‐α induced ICAM‐1 binding in the presence of Mn2+ was significantly lower in myePoldip2‐/‐ neutrophils than in myePoldip2+/+ neutrophils (Figure 5B and 5C). The specificity of this assay for β2 integrin was confirmed by the lack of effect of anti‐Mac‐1 antibody on soluble ICAM‐1 binding (Figure S8). To further confirm the activation of β2 integrin and to provide insight into how Poldip2 affected cell adhesion, we examined phosphorylation of proline‐rich tyrosine kinase 2 (Pyk2), a member of the focal adhesion kinase family that is known to associate with the β2 integrin cytoplasmic tail upon stimulation and is rapidly phosphorylated and activated in an adhesion‐dependent manner. , , As shown in Figures 5D and 5E, Pyk2 phosphorylation was reduced in Poldip2 deficient neutrophils compared with wild‐type neutrophils. These data suggest that Poldip2 is involved in the activation of β2 integrins without disturbing integrin surface expression.

Figure 5

Poldip2 (polymerase [DNA‐directed] δ‐interacting protein 2) deficiency impairs β2‐integrin activation and Pyk2 phosphorylation in neutrophils.

A, Surface expression of β2 integrin αL subunit (left), αM subunit (middle) and β subunit (right) in bone marrow neutrophils isolated from myePoldip2+/+ and myePoldip2‐/‐ mice. Data represent mean fluorescence intensity fold change relative to control as medians with 95% CIs (n=5), *P<0.05, **P<0.01, compared with respective control group (Mann‐Whitney tests). No significant difference was noted between genotypes. B and C, Assessment of β2 integrin activation using the soluble intercellular adhesion molecule‐1 binding assay. Isolated bone marrow neutrophils were stimulated using 20 ng/mL TNF‐α for 10 minutes and incubated with soluble intercellular adhesion molecule‐1‐Fc in the presence of 5mM Mn2+. Data represent fluorescence intensity fold change as medians with 95% CIs (n=5); **P<0.01 compared with myePoldip2+/+ control group, # P<0.05 compared with myePoldip2+/+ Mn2+ + TNF‐α group (Mann‐Whitney tests). D and E, Pyk2 phosphorylation of adhered neutrophils. Isolated bone marrow neutrophils were added to intercellular adhesion molecule‐1 coated plates and either left unstimulated or stimulated with 20 ng/mL TNF‐α for 10 or 30 minutes. Data represent the ratio of phosphorylated Pyk2 (p‐Pyk2) to total Pyk2 as medians with 95% CIs (n=4 for all groups except for myePoldip2‐/‐ control, in which n=3), *P<0.05 compared with myePoldip2+/+ control group, # P<0.05 compared with myePoldip2+/+ TNF‐α 10 minute group (Mann‐Whitney tests). ICAM‐1 indicates intercellular adhesion molecule 1; MFI, mean fluorescence intensity; myePoldip2+/+, wild type; myePoldip2‐/‐, myeloid‐specific Poldip2 knockout; Poldip2, polymerase (DNA‐directed) δ‐interacting protein 2; Pyk2, protein tyrosine kinase 2 beta; and TNF‐α, tumor necrosis factor α.

Discussion

Our previous work showed that Poldip2+/‐ mice are less susceptible to injury triggered by acute inflammation. , However, such studies did not distinguish the role of Poldip2 in immune cells from that in endothelial cells in vivo. Here, the generation of myeloid‐specific knockout mice allowed us to address the function of myeloid Poldip2 during the development of lipopolysaccharide‐induced ARDS. Our findings suggest that myeloid Poldip2 is required for neutrophil recruitment in sepsis‐induced ARDS, at least in part because of its role in regulating β2‐integrin activation and adhesion.

The recruitment of neutrophils into the lungs is one of the hallmarks of ARDS. Considering the destructive potential of neutrophil effectors such as NETs and ROS, delicate control of neutrophil recruitment is required to avoid severe tissue damage. Clinical data and animal models have revealed the importance of neutrophils in acute lung injury. The proportion of neutrophils in the BAL fluid correlates with ARDS severity and outcome in patients, , and depletion of neutrophils in mice can attenuate the severity of lung injury. Therefore, our finding that Poldip2 levels can control neutrophil adhesion and ultimately accumulation in the lung and BAL is of utmost importance.

Lipopolysaccharide, a component of the gram‐negative bacteria wall, can mimic sepsis and sequentially results indirectly in ARDS, making it a useful experimental model to study contributing mechanisms. Consistent with previous reports, we observed an obvious influx of neutrophils into the lung 18 hours post‐intraperitoneal injection of lipopolysaccharide. Strikingly, we found a substantial inhibition of neutrophil sequestration and transalveolar migration in inflamed lungs in myePoldip2‐/‐ mice, supporting our hypothesis that myeloid Poldip2 is involved in the regulation of leukocyte recruitment to inflamed sites. Using this injury model, we previously found a reduction in CXCL‐1 and CXCL‐2 upregulation in Poldip2+/‐ mice. However, deletion of Poldip2 only in myeloid cells led to no inhibition of these chemokines (Figure S2), suggesting that Poldip2 regulation of CXCL‐1 and CXCL‐2 release occurs primarily in other cell types, such as pulmonary endothelial cells and alveolar epithelial cells. Moreover, Poldip2‐deficient neutrophils were able to respond normally to a chemotactic stimulus and produce equal amount of ROS and NETs. These somewhat unexpected findings led us to conclude that neutrophil Poldip2 is not involved in the ability of neutrophils to eliminate pathogens per se, but rather that it impairs the response to recruitment signals despite the presence of sufficient CXCL‐1 and CXCL‐2 in the milieu. It should be noted that besides these primary chemokines, there are many other chemokines that potentially contribute to leukocyte recruitment and could be affected by myeloid Poldip2 depletion, a possibility that remains to be explored. Moreover, given the fact that cytokine synthesis and secretion is a dynamic process, and the response of neutrophils to lipopolysaccharide varies over time, additional time points should be studied to fully characterize the kinetics of the neutrophil response to lipopolysaccharide in myePoldip2‐/‐ mice.

Leukocyte recruitment to injured tissue is an orchestrated response consisting of multiple stages and is controlled by a variety of factors. The process of neutrophil egression from the vasculature to inflamed sites begins with their attraction by chemokine gradients, and is followed by rolling, adhesion, and transmigration at local inflamed sites. L‐selectin expressed by leukocytes is essential in governing leukocyte slow rolling on endothelium, and activation of neutrophils can lead to L‐selectin shedding. In accord with the literature, we observed a directional movement in the presence of chemoattractant fMLP as well as a downregulation of surface L‐selectin (Figure S9) on neutrophils after TNF‐α stimulation. However, we found that neither neutrophil motility and orientation towards fMLP nor L‐selectin shedding were affected by the deficiency of Poldip2, suggesting that the impact of Poldip2 occurs primarily in subsequent stages.

Adhesion is another critical and rate‐limiting step in leukocyte extravasation. Inflamed endothelial cells express adhesion molecules such as ICAM‐1 to strengthen leukocyte adhesion as a consequence of their enhanced interaction. As reported in other studies, we observed an increase in neutrophil adhesion to TNF‐α‐stimulated pulmonary endothelial cells, but adhesion was markedly impaired in Poldip2‐deficient neutrophils. Interestingly, studies have shown that neutrophils adhered to the endothelium affect the endothelial cytoskeleton, inducing remodeling of tight junctions , and thereby increase endothelial permeability and lead to amplified neutrophil transmigration, which would be consistent with our previous observation that Poldip2 depletion reduces lung permeability. β2‐integrin on leukocytes is the counter‐ligand to ICAM‐1 on the endothelium and is pivotal in regulating leukocyte‐to‐endothelium adhesion. , The strength of adhesion is determined by the amount of β2‐integrin expressed on the leukocyte surface and the status of its activation. While we found no change in surface expression of β2‐integrins, using the soluble ICAM‐1 binding assay and measuring phosphorylation of the downstream effector, Pyk2, we show clear evidence of diminished β2‐integrin activation in Poldip2‐deficient neutrophils.

The mechanism by which Poldip2 regulates β2‐integrin activation remains unclear. It is well‐established that integrin adhesiveness is determined by both conformational change (affinity) triggered by intracellular signals and cluster formation on membrane (avidity) initiated by extracellular ligand binding. β2‐integrins on leukocytes tend to be largely inactive at rest, and the switch from low‐affinity to high‐affinity conformation is required for ligand binding in primed neutrophils. Talin and kindlin are of great importance in mediating β2‐integrin affinity by binding to the cytoplasmic tail of the β2 subunit. , , Notably, both of these molecules are cytoskeleton‐associated proteins and our previous study in vascular smooth muscle cells emphasized the significance of Poldip2 in regulating cytoskeletal dynamics, indicating a potential role for Poldip2 regulating inside‐out β2‐integrin activation. Focal adhesion kinase is another essential protein in mediating integrin signaling and is required for ICAM‐1 mediated neutrophil adhesion. Intriguingly, our team found that Poldip2 is involved with focal adhesion kinase activation in both endothelial cells and smooth muscle cells, , which potentially implicates a similar pathway in leukocytes. Our current study extends these findings to include Pyk2, a focal adhesion kinase family member. Finally, Poldip2 depletion was shown to upregulate the expression of β1‐integrin in smooth muscle cells as a result of elevated activation of the PI3K/Akt/mTOR signaling pathway, suggesting that while regulation of integrin signaling may be a common mechanism for Poldip2‐mediated effects, the precise pathways affecting different integrins may be cell type‐specific. Further work will be required to define how Poldip2 regulates β2‐integrin activity in neutrophils.

In summary, our work indicates that Poldip2 mediates β2‐integrin‐dependent adhesion at least in part by impacting β2‐integrin activation during neutrophil recruitment to inflamed lungs but is unrelated to their effector response to inflammatory stimuli. Together with our previous studies delineating a role for Poldip2 in endothelial permeability and inflammation, these findings suggest that Poldip2 may play a central role in the response to lung injury.

Sources of Funding

This study was supported in part by the Emory Flow Cytometry Core, one of the Emory Integrated Core Facilities and is subsidized by the Emory University School of Medicine. This work was supported by National Institutes of Health grants and HL152167 and HL095070 to Kathy K. Griendling and HL151133 to Elena Dolmatova.

Disclosures

None.

Data S1

Table S1

Figures S1–S9

Click here for additional data file.