Neutrophil-Endothelial Interactions of Murine Cells is not a Good Predictor of Their Interactions in Human Cells

Abstract

All drugs recently developed in rodent models to treat inflammatory disease have failed in clinical trials. We therefore used our novel biomimetic microfluidic assay (bMFA) to determine whether the response of murine cells to inflammatory activation or anti-inflammatory treatment is predictive of the response in human cells. Under physiologically relevant flow conditions, permeability and transendothelial electrical resistance (TEER) of human or mouse lung microvascular endothelial cells (HLMVEC or MLMVEC), and neutrophil-endothelial cell interaction was measured. The differential impact of a PKCδ-TAT peptide inhibitor (PKCδ-i) was also quantified. Permeability of HLMVEC and MLMVEC was similar under control conditions but TNF-α and PKCδ-i had a significantly higher impact on permeability of HLMVEC. TEER across HLMVEC was significantly higher than MLMVEC, but PKCδ-i returned TEER to background levels only in human cells. The kinetics of fMLP-mediated neutrophil migration was significantly different between the two species and PKCδ-i was significantly more effective in attenuating human neutrophil migration. However, human and mouse neutrophil adhesion patterns to microvascular endothelium were not significantly different. Surprisingly, while ICAM-1 was significantly upregulated on activated HLMVEC, it was not significantly upregulated on activated MLMVEC. Responses to activation and anti-inflammatory treatment in mice may not always be predictive of their response in humans.

1. Introduction

Inflammatory dysfunction is the underlying cause of a number of pathologies. For example, sepsis which is a clinical syndrome defined as life-threatening organ dysfunction caused by dysregulated host response to infection, has incidence of more than one million per year in the US leading to more than 200,000 deaths/year (1–5). Of particular concern is that the reduction in mortality from sepsis has been primarily due to supportive care rather than effective medicines. Furthermore, all of the ~150 drugs recently developed in animal models to treat sepsis have failed in clinical trials (6). The lack of clinical applicability of anti-inflammatory therapies that work in rodents is the result of multiple factors including the use of injury models which do not mimic the clinical situation, the differential response of in bred vs. out bred rodents, age and sex differences, and the vastly different compositions of mouse and human systemic immune cells (e.g. circulating number of neutrophils) (6, 7).

Activation of endothelium is an important player in the host response to inflammation that results in increased neutrophil trafficking, capillary leak, loss of endothelial barrier function, and edema in critical organs (8, 9). In the case of sepsis, patients often die of organ failure and therapies directed against endothelial cell dysfunction and tissue damage may be of great interest in treating the disease (2, 8, 10). Key to inflammation-induced tissue damage is excessive neutrophil migration across the vascular endothelium, and neutrophil dysregulation has a critical role in the early course of organ damage through release of proteases and oxygen radicals (9, 11–15). Potential therapeutic target sites include local control of the vascular endothelial response to systemic inflammation as well as direct modulation of neutrophil recruitment and activation (11, 16–20). Neutrophil rolling on the endothelium is mediated by selectins, whereas firm adhesion and migration into the tissue are mediated by integrins/immunoglobulins and chemoattractants in the tissue. During inflammation, mediators damage the vascular endothelium resulting in increased permeability and excessive neutrophil migration into critical organs, with the lung being an early target (21). Modulating neutrophil recruitment reduces symptoms of inflammation and prevents lung injury (22), and identification of key molecules involved in the neutrophil adhesion cascade will allow for the development of novel therapeutics (23). However, the underlying pathophysiology of neutrophil-mediated tissue damage is yet to be understood and there are no specific pharmacologic therapies available that protect from neutrophil-mediated tissue damage (24, 25).

Most in vivo studies of mechanisms of inflammatory disease primarily employ murine models. However, concerns regarding the level of correspondence between mice and cell culture models, as well as phenotypic heterogeneity of different types of endothelial cells, and their relevance to human disease, have been expressed in the literature (26, 27). Therefore, a significant limitation of mouse models may be that a given therapeutic may impact mice differently as compared to humans (26–30). In support of this concept, one recent study employing bulk and single cell transcriptomics to map the innate immune response demonstrated significant species differences in cytokines, chemokines and their respective receptors (31). Recently, an international panel of experts emphasized the continuing need for mouse models in sepsis research but outlined significant limitations and the need for models that better represent human disease (32). Thus, there is a significant need for an “in vitro reconstitution of disease-related cell types or tissues” to study human inflammatory diseases (27).

In part to address the limitations of animal models, microfluidic systems have been developed for studies of human cells in a more controlled environment (33). Unfortunately, most of these devices do not realistically reproduces the complex geometry of in vivo microvascular network to model the complete inflammatory response (e.g. rolling, adhesion, migration) in the vasculature nor have they been validated against in vivo data. To address these important limitations, we developed a novel biomimetic microfluidic assay (bMFA) for studying the entire neutrophil adhesion cascade in a single assay (34–38). This assay reproduces the topography and flow conditions of the in vivo microvascular networks in a physiologically realistic, 3D environment that resolves and facilitates direct assessment of individual steps including rolling, adhesion, and extravasation of the leukocytes into the extra-vascular tissue space in a single system. This assay has been validated against in vivo models (36) and allows us to compare the differential response to inflammation and the impact of therapeutics on neutrophil-endothelial interactions in murine and human based mimetic systems.

In this study, we use bMFA to investigate whether mouse neutrophil-endothelial cell interactions and microvascular endothelium barrier characteristics, as well as response to a novel anti-inflammatory agent, are predictive of human cell responses. Previously, we demonstrated that PKCδ is an important regulator of neutrophil-endothelial cells interaction during inflammation and that a PKCδ-TAT peptide inhibitor may serve as a potential novel anti-inflammatory therapeutic (39–43). In this study, we tested the hypothesis that mouse and human cells do not respond similarly to activation and investigated whether the response of mouse cells to an anti-inflammatory agent is likely to provide data that would be predictive of its efficacy in human cells.

2. Materials and Methods

2.1. Materials, equipment, and reagents

A mouse monoclonal anti-human ICAM-1 (sc-107), a mouse monoclonal anti-human VCAM-1 (sc-18854), a mouse monoclonal anti-human E-selectin (sc-5262) were purchased from Santa Cruz Biotechnology (Dallas, TX, USA); a goat polyclonal anti-human JAM-C (AF1189) and a goat polyclonal anti-mouse JAM-C (AF1213) were purchased from R&D systems (Minneapolis, MN, USA); a mouse monoclonal anti-mouse ICAM-1 (ab171123), a rabbit monoclonal anti-mouse VCAM-1 (ab134047) and mouse fibronectin were purchased from Abcam PLC (Cambridge, MA, USA); a rat monoclonal anti-mouse E-selectin (553749) and human fibronectin were purchased from BD Biosciences (San Jose, CA, USA). Measurement of ICAM-1 expression under static conditions as described in section 2.12 was performed using two different anti-mouse ICAM-1 mAb purchased from Abcam (ab171123) and Invitrogen (14-0541-85). Protein A, fluorescent 9.9 μm microparticles (green: excitation 468 nm, emission 508 nm), carboxyfluorescein diacetate succinimidyl ester (CFDA/SE) probe, Hanks’ Balanced Salt Solution (HBSS), Trypsin/EDTA, Formalin, Triton X-100, Draq5, 40kDa Texas Red conjugated dextran, Alexa Fluor® 488 Phalloindin, and Hoechst 33342 were purchased from ThermoFisher Scientific (Rockford, IL, USA). Human lung microvascular endothelial cells (HLMVEC) were purchased from Lonza (Basel, Switzerland) and mouse lung microvascular endothelial cells (MLMVEC) were purchased from Cell Biologics (Chicago, IL, USA). Bovine Serum Albumin (BSA) was purchased from Sigma-Aldrich (St. Louis, MO, USA). Recombinant mouse TNF-α was purchased from EMD Millipore (Burlington, MA, USA) and recombinant human TNF-α was purchased from BioVision (Milpitas, CA, USA).

2.2. PKCδ-TAT peptide inhibitor synthesis

PKCδ activity was selectively inhibited by a peptide antagonist that consisted of a peptide derived from the first unique region (V1) of PKCδ (SFNSYELGSL: amino acids 8-17) coupled via an N-terminal Cys-Cys bond to a membrane permeant peptide sequence in the HIV TAT gene product (YGRKKRRQRRR: amino acids 47-57 of TAT) (44). The PKCδ TAT peptide (PKCδ-i) produces a unique dominant-negative phenotype that effectively inhibits activation of PKCδ but not other PKC isotypes (44–46). The peptide was synthesized by Mimotopes (Melbourne, Australia) and purified to >95% by HPLC.

2.3. Design and fabrication of the microfluidic assay

We have published the methods for design and fabrication of the novel microfluidic assay and in vivo validation previously (36, 39, 47). Briefly, a modified Geographic Information System (GIS) approach was used to digitize the in vivo microvascular networks (Figure 1A) that were lithographically patterned on polydimethylsiloxane (PDMS) (Figure 1B). Microfabricated pillars (10 μm diameter) were used to create the 3 μm ×100 μm pores resulting in a network of vascular channels connected to a tissue compartment via a 3 μm porous barrier (Figure 1C), which is the optimum size for neutrophil migration.

Figure 1.

The bMFA mimics a physiologically relevant microvascular environment. Intravital microscopy was used to map microvascular networks in animals (A). These maps are then used to fabricate the vascular network on PDMS and assemble the biomimetic microfluidic assay (B). The bMFA includes vascular channels that are connected to the tissue compartment through a 3μm barrier (C). Confocal microscopy shows that human microvascular endothelial cells form a complete lumen in the vascular channel of bMFA (F-actins labeled in green using phalloidin; cell nuclei labeled in red using Draq5) (D).

2.4. Seeding of endothelial cells in bMFA

Human (HLMVEC) or mouse (MLMVEC) lung microvascular endothelial cells were cultured in the corresponding microvascular endothelial growth media (MV-EGM) and used between passages 1-3. The bMFA was coated with human or mouse fibronectin and corresponding endothelial cells were cultured under shear flow (inlet flow rate of 0.1 μl/min) for 48 hrs (36). The bMFA was then treated for 4 hours under flow (inlet flow rate of 0.1 μl/min) with buffer, TNF-α (10 U/mL), or TNF-α (10 U/mL) + PKCδ-TAT peptide inhibitor (5 μM) before introduction of neutrophils or antibody coated microparticles into bMFA. The bMFA was similarly treated for permeability and TEER studies. Neutrophils were treated for 15 minutes with buffer, TNF-α (10 U/mL), or TNF-α (10 U/mL) + PKCδ-TAT peptide inhibitor (5 μM) before introduction into bMFA.

Images were acquired using an ORCA Flash 4 camera (Hamamatsu Corp., USA) on a Nikon TE200 fluorescence microscope equipped with an automated stage. An Olympus FluoView FV1000 confocal microscope equipped with a fully automated stage was used for capturing confocal image stacks. PhD Ultra Syringe pump (Harvard Apparatus, USA) was used for injecting growth medium, permeability dye, or neutrophil/microparticles suspension into the bMFA with high precision. A stage warmer was used to keep the bMFA at 37 °C. NIS Elements software (Nikon Instruments Inc., Melville, NY) was used to control the microscope stage and the camera.

Consistent with our published data (39, 48, 49), endothelial cells in bMFA form a confluent and complete 3D lumen in vascular channels under physiological conditions; the formation of a complete lumen was confirmed using confocal microscopy (Figure 1D) (39, 48). Assays in which neutrophils freely entered the tissue compartment without attachment were discarded.

2.5. Immunofluorescence staining and quantification

The formation of F-actin filaments and cell coverage in the vascular channels of the bMFA was characterized using immunostaining with Draq5 against cell nuclei and phalloidin against F-actin filaments. Images were taken using the microscope and camera system described before. Orientation of F-actin fibers in the flow direction was evaluated using OrientationJ plugin for ImageJ which takes orientation and coherency of fibers into account (50).

2.6. Permeability measurements

The vascular compartment was connected to a Hamilton gas tight syringe filled with Texas Red 40 kDa dextran (25 μM in MV-EGM) mounted on a syringe pump. Permeability was measured by imaging the bMFA every minute for 2 hrs while the dextran solution flowed through the vascular channels (1 μl/min inlet flow). Using our previously published method (48), permeability (P) of dextran across the endothelium was calculated using:

$$P=\frac{1}{I_{v_0}}\frac{V}{S}\frac{d{I}_t}{dt}$$ (1)

where I_t is the average intensity in the tissue compartment, I_v0 is the maximum fluorescence intensity of the vascular channel, and V/S is the ratio of vascular channel volume to its surface area.

2.7. Transendothelial electrical resistance measurements

Transendothelial electrical resistance (TEER) was measured following our established method (48) using an electrode compartment outside the vascular channels. Ag/AgCl electrodes were placed on either side of the HLMVEC or MLMVEC in the vascular and tissue compartments and connected to SynVivo Cell Resistance Analyzer (SynVivo Inc., Huntsville, AL). Impedance measurements were acquired at 10 kHz with a voltage of 10 mV.

2.8. Human neutrophil isolation and labeling

Heparinized human blood was obtained from healthy, male and female adult donors, following informed consent as approved by the Institutional Review Board of Temple University (Philadelphia, PA, USA). Human neutrophils were isolated using ficoll-hypaque separation, dextran sedimentation, and hypotonic lysis to remove erythrocytes (40, 51). Isolated neutrophils were suspended in HBSS (5 × 10⁶ cells/ml) and labeled using CFDA/SE probe for 10 min at room temperature. Fluorescently labeled neutrophils (5 X 10⁶ cells/ml) were introduced into the vascular channels of the bMFA at a flow rate of 1 μl/min.

2.9. Mouse neutrophil isolation and labeling

Animal procedures and handling were conducted in accordance to National Institutes of Health standards and were approved by the Institutional Animal Care and Use Committee at Temple University (Philadelphia, PA, USA). Following euthanasia, the femur and tibias from both hind legs were harvested from C57BL/6 mice (Jackson Labs, Bar Harbor, ME, USA). The distal tip of each bone was cut off, bones were rinsed using HBSS, and cell clumps were dispersed. Neutrophils were isolated using a Percoll gradient sedimentation, followed by hypotonic lysis to remove erythrocytes. Isolated neutrophils were suspended in HBSS and labeled using the CFDA/SE probe for 10 min at room temperature. Fluorescently labeled neutrophils (5 X 10⁶ cells/ml) were introduced into the vascular channels of the bMFA at a flow rate of 1 μl/min.

2.10. Preparation of antibody coated microparticles

Using our established methodology (52), the level of binding of antibody (e.g. anti-ICAM-1) coated microparticles to microvascular endothelial cells was used as an index of the level of upregulation of adhesion molecules. Briefly, 9.9 μm fluorescent polystyrene microparticles were washed with a sodium bicarbonate buffer and coated with protein A (300 μg/ml) via passive adsorption and incubated overnight at room temperature. Microparticles were then washed and incubated in a blocking buffer (1% BSA in HBSS) at room temperature. Microparticles (5 X 10⁶ particles/ml) were counted, diluted in blocking buffer, and incubated with antibodies to the appropriate adhesion molecule for 30 min. Antibody coated microparticles were then suspended in the corresponding media and introduced into the bMFA (39) and their adhesion to endothelial cells was measure as described below.

2.11. Quantification of neutrophil-endothelial and microparticle-endothelial interaction

Fluorescently labeled human or mouse neutrophils or antibody coated microparticles were introduced in the vascular compartment with an inlet flow rate of 1 μl/min. Cells or microparticles that did not move for 30 seconds were considered adherent. Adhesion level of neutrophils or microparticles to the endothelium was quantified by scanning the entire network. The number of migrated neutrophils was quantified using time-lapse imaging every 3 min for 60 min. Nikon Elements and Fiji software were used to collect and analyze the data (53).

2.12. Measurement of VCAM-1 and ICAM-1 expression under static conditions in MLMVEC

To rule out the potential role of shear flow in regulating expression of adhesion molecules (54), a cell surface ELISA method (40) was used to measure VCAM-1 and ICAM-1 expression on TNF-α activated mouse cells under static conditions. Briefly, MLMVEC were grown to confluence on 96-well plates and then were pretreated with buffer or TNF-α (10 U/mL) for 24 hr. The cell monolayers were washed, tetramethylbenzidine liquid substrate (Sigma-Aldrich) was added, and optical density at 650 nm(OD650) was determined. Constitutive expression of VCAM-1 and ICAM-1 was normalized to 1 and expression in response to TNF-α was compared with constitutive ICAM-1 and VCAM-1 expression. All measurements were done in triplicates.

2.13. Statistical analysis

Data are presented as Mean ± SEM. All numerical data passed the Shapiro-Wilk normality test. Statistical significance was determined by one-way or two-way analysis of variance (ANOVA) with Holm-Šidák method post-hoc using the SigmaPlot software. Differences were considered statistically significant if p<0.05.

2.14. Data Availability Statement

Datasets are available on request: the raw data supporting the conclusions of this manuscript will be made available by the authors, without undue reservation, to any qualified researcher.

3. Results

3.1. Human endothelial cell microfilaments align more strongly in the direction of flow as compared to mouse endothelial cells

Classically, HUVECs have been shown to strongly align under shear flow but the degree of alignment of endothelial cells from other organs and different species has not been widely studied. We and others have observed that some endothelial cells do not align as well as HUVEC in the direction of flow (48, 50, 55). As a gross measure of differential response of human and mouse cells to flow, the expression of F-actin filaments in lung microvascular endothelial cells from human (HLMVEC) and mouse (MLMVEC) under physiologically relevant flow condition was investigated using immunofluorescence staining and fluorescence microscopy. Both HLMVEC and MLMVEC completely covered the vascular channels and formed a 3D lumen under flow conditions (Figure 2A) and both HLMVEC and MLMVEC aligned in the direction of flow in all vessels of the network under shear flow ranging from 10-150 s^–1. However, when accounting for both orientation and coherency of fibers, F-actin filaments of HLMVEC aligned significantly more uniformly in the direction of flow as compared to F-actin filaments of MLMVEC (Figure 2B).

Figure 2.

Alignment of F-actin fibers with flow is different between species. Human lung microvascular endothelial cells orient in the direction of flow more strongly as compared to mouse lung microvascular endothelial cells (F-actin labeled in green using phalloidin; cell nuclei labeled in red using Draq5) (A) – the scale bar indicates 50μm. Quantitative measurement using ImageJ shows significantly more alignment of F-actin fibers in the flow direction in human cells (N=3, Mean ± SEM, one-way ANOVA, ** p<0.01) (B).

3.2. Endothelial cell barrier responds differently to activation and treatment between species

Loss of endothelial barrier function is a hallmark of many inflammatory conditions. We therefore directly assessed the integrity of the endothelial cell barrier in the bMFA by measuring both the permeability of 40 kDa dextran from the vascular channels to the tissue compartment and transendothelial electrical resistance (TEER). The permeability coefficient of non-electrolyte tracers (fluorescent dextran) is an index of the rate at which dye molecules pass across endothelial cell monolayer, whereas TEER reflects the ionic conductance of the paracellular as well as transcellular transport in the endothelial cell monolayer (56, 57).

Permeability across the endothelial barrier was measured in 4 hrs TNF-α activated cells with or without PKCδ-TAT peptide inhibitor (PKCδ-i) treatment as described in Materials and Methods. While the overall permeability in no treatment cells was not significantly different between human and mouse cells, TNF-α activation had a significantly higher impact on permeability of human as compared to mouse cells (Figure 3). Furthermore, PKCδ inhibition significantly reduced (by 61%) the permeability of HLMVEC to background levels (Figure 3). Permeability of TNF-α treated MLMVEC was also significantly reduced by PKCδ inhibition (by 48%) but the magnitude of this reduction was smaller compared to human cells.

Figure 3.

Primary human and mouse lung microvascular endothelial cell permeability to 40 kDa dextran under No Treatment conditions are not significantly different from each other. TNF-α activation had a significantly higher impact on permeability of human as compared to mouse cells. Furthermore, PKCδ inhibition significantly reduced the permeability of HLMVEC to background levels. Permeability of TNF-α treated MLMVEC was also significantly reduced by PKCδ inhibition but the magnitude of this reduction was smaller compared to human cells.

(N=4, Mean ± SEM, two-way ANOVA, * p<0.05).

TEER across human cells was significantly higher compared to mouse cells under all experimental conditions tested (Figure 4). Furthermore, TNF-α activation significantly reduced TEER in both HLMVEC (by 51%) and MLMVEC (by 75%). However, while PKCδ-TAT peptide inhibitor treatment significantly increased TEER in both human and mouse cells and TEER returned to background levels in human cells, TEER in mouse cells remained at 65% of no treatment levels even after PKCδ inhibition.

Figure 4.

Transendothelial electrical resistance (TEER) across human cells is significantly higher compared to mouse cells under No Treatment, TNF-α activated, and PKCδ-TAT peptide inhibitor (PKCδ-i) treated conditions. While TNF-α activation significantly reduces TEER in both human and mouse lung endothelial cells and PKCδ inhibition significantly increases TEER in both human and mouse cells, TEER returns to background levels only in PKCδ-i treated human cells. (N=3, Mean ± SEM, two-way ANOVA, * p<0.05, ** p<0.01, *** p<0.001).

3.3. Kinetics of neutrophil migration across endothelium is different between species

Neutrophil migration from vasculature into tissue is an important clinical manifestation of inflammatory disease. We therefore determined how neutrophil migration kinetics may be different between mouse and human cells and how TNF-α activation may impact this kinetics.

Human neutrophil migration across human endothelial cells into the tissue compartment after TNF-α activation in response to fMLP increased gradually and significantly over the 60-min observation period (Figure 5A). The kinetics of mouse neutrophil migration across mouse endothelial cells was rapid and reached a maximal value that did not change significantly after the first 10 min (Figure 5B). While mouse neutrophil migration was more rapid, the total number of neutrophils that migrated was greater in human as compared to mouse. Under no treatment conditions, migration of human neutrophil across human endothelial cells was not significantly different from migration of mouse neutrophils across mouse endothelial cells. Surprisingly, PKCδ inhibition was significantly more effective in reducing neutrophil migration across human endothelium as compared to mouse endothelium (92±7% reduction in human vs. 49±6% reduction in mouse at 60 min) (Figure 5C). Furthermore, PKCδ inhibition reduced human neutrophil migration to background level but not mouse neutrophils. As a whole, the response of mouse cells to PKCδ inhibition significantly underestimates its response in human cells.

Figure 5.

The kinetics of human and mouse neutrophil migration across corresponding endothelial cells are different, with human neutrophils migrating gradually while mouse neutrophils migrate rapidly. Human neutrophil migration across human endothelial cells into the tissue compartment after TNF-α activation in response to fMLP increases gradually and significantly over the 60-min observation period (A). Mouse neutrophils migrate across mouse endothelial cells rapidly and reach a maximal value after the first 10 min (B). PKCδ inhibition (PKCδ-i) is significantly more effective in reducing neutrophil migration across human endothelium as compared to mouse, and PKCδ inhibition reduces human neutrophil migration to background levels but not mouse neutrophils (C). TNF-α activation has a significantly higher impact on human neutrophil migration as compared to mouse neutrophils and PKCδ inhibition is significantly more effective in reducing neutrophil migration in humans. (N=4, Mean ± SEM, two-way ANOVA, * p<0.05, ** p<0.01, *** p<0.001).

3.4. Neutrophil adhesion patterns to microvascular endothelium are similar between human and mouse

Neutrophil adhesion to endothelium under shear flow precedes its migration and is a key regulator of the inflammatory response. Overall, adhesion of human or mouse neutrophils to their corresponding endothelial cells increased significantly after TNF-α activation and returned to no treatment levels after treatment with the PKCδ-TAT peptide inhibitor (Figure 6A and 6B). An advantage of bMFA is that spatial and flow dependent adhesion and migration of neutrophils in a microvascular network can be readily studied in real time. The increase in neutrophil adhesion to TNF-α activated endothelial cells and its return to no treatment levels after PKCδ inhibition was most pronounced in vessels with low flow (shear values <30 s⁻¹) and, in the case of mouse cells, at bifurcations where the flow patterns are altered. Overall, in contrast to neutrophil migration, the adhesion patterns of human neutrophils to HLMVEC were not significantly different from adhesion patterns of mouse neutrophils to MLMVEC (Figure 6C).

Figure 6.

Adhesion of neutrophils to the corresponding endothelium is similar between human and mouse. TNF-α activation and treatment with the PKCδ-TAT peptide inhibitor (PKCδ-i) similarly impacts adhesion of human (A) or mouse (B) neutrophils to their corresponding endothelial cells. The impact of TNF-α activation of endothelial cells and treatment with the PKCδ-TAT peptide inhibitor was most pronounced in vessels with low flow. Mouse neutrophils adhered significantly near bifurcations where the flow patterns are altered. Overall, the adhesion patterns of neutrophils to endothelial cells in human is not significantly different from that of mouse (C) (N=4, Mean ± SEM, two-way ANOVA, * p<0.05, ** p<0.01, *** p<0.001).

3.5. Adhesion molecules are differentially expressed in mouse vs. human microvascular endothelial cells

On endothelial cells, selectins (e.g. P & E-selectin) are responsible for neutrophil capture and rolling, while adhesion molecules such as ICAM-1, VCAM-1, and JAM-C (junctional adhesion molecule-C) are critical regulators of neutrophil firm attachment and migration (58–61). Antibody coated microparticles were used to determine if various adhesion molecules are differentially upregulated on human and mouse microvascular endothelial cells in response to 4 hr TNF-α activation and treatment with the PKCδ-TAT peptide inhibitor under the experimental conditions described earlier. E-selectin was significantly upregulated on TNF-α activated human cells and returned to no treatment levels after treatment with the PKCδ-TAT peptide inhibitor. Similarly, levels of ICAM-1 and VCAM-1 were also significantly upregulated on TNF-α activated human cells and returned to no treatment levels after treatment with the PKCδ-TAT peptide inhibitor (Figure 7A). JAM-C was not upregulated in human cells following 4 hr TNF-α treatment.

Figure 7.

Adhesion molecules are differentially expressed in human and mouse cells in response to TNF-α mediated activation. While the adhesion pattern of microparticles coated with anti-E-selectin, anti-VCAM-1 or anti-JAM-C was similar between human (A) and mouse cells (B), adhesion of microparticles coated with anti-ICAM-1 to human cells was significantly higher as compared to mouse cells; i.e. ICAM-1 is not upregulated on activated mouse endothelial cells (B). Treatment with the PKCδ-TAT peptide inhibitor (PKCδ-i) reduced the upregulated adhesion of E-selectin, ICAM-1 and VCAM-1 in human cells as well as E-selectin and VCAM-1 in mouse cells to their corresponding background levels. (N=3, Mean ± SEM, one-way ANOVA, ** p<0.01, and *** p<0.001)

In mouse cells, similar to human endothelial cells, E-selectin and VCAM-1 levels were also significantly upregulated in response to TNF-α activation and returned to no treatment levels after treatment with the PKCδ-TAT peptide inhibitor. Surprisingly however, ICAM-1 was not upregulated in TNF-activated mouse endothelial cells, and its expression level was not altered after treatment with the PKCδ-TAT peptide inhibitor (Figure 7B). To further verify these findings and to rule out the potential role of shear flow in regulating expression of adhesion molecules (54), we used a cell surface ELISA method in a static assay (40) to show that VCAM-1 was significantly upregulated on TNF-α activated mouse cells under static conditions clearly indicating that TNF-α properly activates mouse endothelial cells (see Supplemental Material). Furthermore, we show that ICAM-1 is not upregulated in C57BL/6 mouse cells under static conditions using two different antibodies to ICAM-1 (see Supplemental Material). Since ICAM-1 is known to be a key regulator of neutrophil migration (20, 22), this observation is consistent with other data presented in this study indicating that neutrophil-endothelial interaction in human and mouse are differentially regulated.

4. Discussion

The lack of clinical applicability of many anti-inflammatory therapies that work in rodents may be due to important physiological differences in the two species (6, 7). Furthermore, the underlying pathophysiology of neutrophil-mediated tissue damage is poorly understood and there are no specific pharmacologic therapies available that protect from neutrophil-mediated tissue damage (24, 25). By the same token, therapeutics that do not work well in animal models are often abandoned before their applicability to human disease is determined. To address these challenges, we have used a novel biomimetic microfluidic assay (bMFA) that reproduces the entire neutrophil adhesion cascade in a physiologically realistic three-dimensional environment and has already been validated against in vivo data (33, 36, 37, 47). In the present study we used the bMFA to investigate neutrophil-microvascular endothelial cell interaction in human and mouse, and to study a PKCδ-TAT peptide inhibitor efficacy and regulation of neutrophil-endothelial cell interaction during inflammation. Overall, our findings indicate that while many neutrophil-endothelial interactions are preserved across species, there are significant differences between human and mouse cells responses to activation and in response to therapeutics. For example, while mouse and human neutrophil adhesion patterns to their corresponding endothelial cells are similar, migration of neutrophils across the endothelial barrier has a faster time frame in mouse cells as compared to human cells. More significantly, application of a novel anti-inflammatory agent (PKCδ-TAT peptide inhibitor) in mouse cells significantly underestimates its efficacy in preventing neutrophil migration across human endothelial cells. Furthermore, microfilament alignment, permeability and TEER measurements indicate a consistent difference in microvascular endothelium barrier function between mouse and human. The degree and significance of alignment of endothelial cells under flow, as well as its relationship with permeability, has not been well-studied and while it may be tempting to conclude that better aligned cells have lower (or higher) permeability, more in-depth studies are required to determine how alignment could potentially affect cell function. Given the clinical significance of neutrophil migration and endothelial permeability, our findings indicate that mouse models may not always be good predictors of human response to anti-inflammatory therapeutics.

A notable feature of bMFA is the ability to determine the molecular mechanisms that may be responsible for the differential response of different cells to activation and therapeutics. Our findings indicate that E-selectin, which regulates leukocyte rolling, is upregulated in both 4-hour activated human and mouse cells and its levels return to normal after treatment with the PKCδ-TAT peptide inhibitor. Similarly, several adhesion molecules involved in leukocyte adhesion and migration, most notably ICAM-1, were significantly upregulated in activated human lung endothelial cells at the same time point and their upregulation was significantly modulated by inhibition of PKCδ. Surprisingly however, ICAM-1 was not significantly upregulated on activated mouse lung cells. ICAM-1 has been shown to be upregulated in mouse brain, heart, and intestine endothelial cells 5 hours after TNF or LPS treatment (62, 63) but no change in ICAM-1 expression on mouse pulmonary endothelial cells during pneumonia has been observed (64). Given the increasing efforts to develop drugs that specifically target upregulation of adhesion molecules during the inflammatory response (65) , additional studies are needed to determine the extent to which upregulation of various adhesion molecules in mouse models/cells correspond to those observed in humans and/or human cells.

The species differences observed in neutrophil-endothelial interactions in this study (Figures 5 and 6) may be due to the differential response of either neutrophils or endothelial cells to activation and treatment with anti-inflammatory agents. We hypothesize that these differential responses observed in this study are primarily due to the differences in endothelial cells and not neutrophils as the findings reported in Figures 5 and 6 are consistent with differences observed in other assays (Figures 2–4, and 7) that only involve endothelial cells. Furthermore, we used both permeability to a 40 kDa dextran particle and transendothelial electrical resistance to show that the endothelial barrier is differentially impacted by TNF-α activation and PKCδ treatment. Transport of macromolecules larger than 3 nm, such as the 40 kDa dextran, occurs through the transcellular pathway (transcytosis) (66–69) that is regulated in part by cytoskeleton proteins such as actin and is often altered by TNF-α activation (70). TEER on the other hand, is an index of current flow via the paracellular route (through the junctions between cells and regulated by junctional proteins) or via the transcellular route (through the cell body regulated by cell membrane lipid bilayer) (71). These different regulatory mechanisms may become more prominent depending on the phenomenon being studied and additional studies may be required to further determine the mechanisms responsible for differential response of endothelial barrier to various stimuli in different species.

Given these compelling findings, cataloging the response of cells in other species and/or mouse strains would be of interest in screening potential therapeutics. For the studies presented here, we selected the most commonly used mouse strain (C57BL) and one that is used for most therapeutic preclinical trials.

In summary, we have used a novel biomimetic microfluidic assay (bMFA) to investigate differences in neutrophil-endothelial cell interaction and on the microvascular endothelium barrier function between mouse and humans. Our findings indicate that in many cases the response of murine cells to inflammatory signals may be a poor predictor of response in human cells. Moreover, our results suggest that PKCδ is an important regulator of neutrophil-endothelial cells inflammatory cross-talk but differentially regulates the expression of endothelial adhesion molecules as well as microvascular endothelial barrier function in human and mouse cells. The novel biomimetic microfluidic assay (bMFA) provides a tool for rapid screening of novel therapeutics for predicting their response in humans.

Abbreviations

bMFA

biomimetic Microfluidic Assay

BSA

Bovine Serum Albumin

CFD

Computational Fluid Dynamics

CFDA/SE

Carboxyfluorescein diacetate succinimidyl ester

fMLP

N-Formylmethionyl-leucyl-phenylalanine

GIS

Geographic Information System

HBSS

Hank’s Balanced Salt Solution

HUVEC

Human Umbilical Vein Endothelial Cell

ICAM-1

Intercellular Adhesion Molecule 1

ICAM-2

Intercellular Adhesion Molecule 2

JAM-C

Junctional Adhesion Molecule C

PKCδ

Protein Kinase C-delta

PKCδ-i

Protein Kinase C-delta TAT peptide inhibitor

TEER

Trans Endothelial Electrical Resistance

TNF-α

Tumor Necrosis Factor α

VCAM-1

Vascular cell adhesion protein