Secondary Logo

Journal Logo

Basic Science Aspects

Abrupt Reflow Enhances Cytokine-Induced Proinflammatory Activation of Endothelial Cells During Simulated Shock and Resuscitation

Li, Ranran*; Zijlstra, Jan G.; Kamps, Jan A.A.M.*; van Meurs, Matijs*†; Molema, Grietje*

Author Information
doi: 10.1097/SHK.0000000000000223



Vascular endothelial cells are constantly exposed to blood flow under normal physiological conditions. Shear stress produced by blood flow modulates the structure and function of endothelial cells (1). The occurrence of blood flow disturbances in the vascular system is a hallmark of circulatory shock, which is defined as reduced organ perfusion or (micro)circulatory arrest and contributes to patient morbidity and mortality (2). Systemic inflammatory response is conceived as the leading cause of the development of MODS (multiple organ dysfunction syndrome) after shock. In hemorrhagic shock mice, endothelial cells show early proinflammatory activation and vascular destabilization, a phenotype that is augmented in the early stage of resuscitation after hemorrhagic shock (3). Activated endothelial cells express adhesion molecules and chemokines that mediate interactions between leukocytes and endothelial cells as well as subsequent leukocyte migration into tissues that plays a critical role in organ injury (4).

Events proposed to be involved in endothelial activation and organ injury during hemorrhagic shock and resuscitation encompass leukocyte activation (5), the release of cytokine by innate immune cells (6), and hemodynamic alterations of microcirculatory blood flow (2). Elucidating endothelial responses to these different variables may shed light on their relative contributions to proinflammatory activation and vascular destabilization. They furthermore can support the rational design of therapeutic strategies for maintaining proper organ function during microcirculatory arrest in shock and resuscitation as well as the period thereafter.

In the present study, our interest was to investigate the role in endothelial activation of flow cessation and its later recovery as appearing in hemorrhagic shock and subsequent resuscitation. Because it is difficult, if not impossible, to study in vivo flow alterations separately from consequential cytokine production and leukocyte activation and tissue hypoxia, we chose an in vitro approach to dissect flow changes from the other coexisting factors. The aim of this study was to better understand the role of shear stress and shear stress changes per se in endothelial activation and to evaluate their effects on concomitant cytokine challenge. For this purpose, we used endothelial cells that were preadapted to 48-h laminar shear stress (LSS) to represent the condition of continuous blood flow in vivo until hemorrhagic shock and resuscitation happen. Endothelial proinflammatory activation and vascular integrity–related molecules were studied during these flow changes (Fig. 1A). They include Tie2, the expression of which was lost in hemorrhagic shock/resuscitation (7), and Angiopoietin-2 (Ang2), the vasculature-destabilizing ligand of Tie2 signaling, which is known to be released by Weibel-Palade bodies on stimulation and has a role in ischemia-induced proinflammatory activation (8). We furthermore examined the responses of VE-cadherin and CD31 (or PECAM1) that function among others in endothelial permeability control (9, 10). Lastly, we included cellular adhesion molecules E-selectin, vascular cell adhesion molecule 1 (VCAM-1), and intercellular adhesion molecule 1 (ICAM-1), as well as chemokine IL-8, that were previously found to be regulated in vivo by installment of shock and after resuscitation (3). Transcription factor KLF2 was analyzed as a control gene to assess cell responsiveness to flow changes.

Fig. 1
Fig. 1:
Schematic representation of selected molecules investigated in endothelial cells and experimental setup of in vitro flow studies performed. A, E-selectin, VCAM-1, and ICAM-1 are cellular adhesion molecules that mediate leukocyte-endothelial interactions in inflammation. Interleukin-8 is a chemokine that is released on activation to regulate leukocyte attraction. Angiopoietin-2, released by Weibel-Palade bodies on proinflammatory activation, binds to receptor Tie2, thereby destabilizing the vasculature. Angiopoietin-2/Tie2 as well as CD31 and VE-cadherin play important roles in maintaining vascular integrity. KLF2 is a transcription factor that shows high sensitivity to shear stress exposure and is analyzed in this study to molecularly validate flow alterations. B, Scheme of the experimental design: human umbilical vein endothelial cells were seeded in ibidi microchambers and cultured overnight in static condition before exposure to unidirectional LSS. After 48 h of 20 dyne/cm2 LSS adaptation, flow was stopped for indicated periods to mimic the marked loss of shear associated with hemorrhagic shock in vivo. To mimic the subsequent resuscitation after hemorrhagic shock, laminar flow at 20 dyne/cm2 was restarted for indicated periods. At time points indicated in certain experiments, cells were further exposed to TNF-α stimulation, which is prominently produced in vivo during shock events.


Cell culture

Human umbilical vein endothelial cells (HUVECs) were obtained from the Endothelial Cell Facility of the University Medical Center Groningen. Human umbilical vein endothelial cells were isolated from at least two umbilical cords and cultured on 1% gelatin–coated plates or microchambers with medium (RPMI 1640; Lonza, BE) supplemented with endothelial growth factors, 20% heat inactivated fetal calf serum (FCS), 2 mM l-glutamine, 5 U/mL heparin, 50 μg/mL endothelial cell growth factor, and antibiotics (100 IE/mL penicillin and 50 μg/mL streptomycin). Human umbilical vein endothelial cells of passages 2 to 4 were used in this study.

Flow experiments

Human umbilical vein endothelial cells were detached with trypsin/EDTA and seeded in 1% gelatin–coated microchambers (μ-Slide I 0.4 Luer, sterile, ibidi, Martinsried, Germany). A total of 60,000 cells/cm2 were used to obtain confluent monolayers overnight. Cells were subjected to 20 dyne/cm2 LSS produced by flowing medium for 48 h, where appropriate, cells were subjected to flow cessation for indicated periods followed by reflow with 20 dyne/cm2 LSS for different periods as indicated. For gene expression analysis, cells were lysed with RLT buffer containing 10% β-mercaptoethanol (see below). For flow cytometric assays, cells were first detached with EDTA/trypsin from the slides before further processing (see below).

The experimental setup is illustrated schematically in Figure 1B.

RNA isolation and quantitative RT-PCR

Total RNA from cultured cells was isolated with the RNeasy Mini plus Kit (Qiagen, Leusden, The Netherlands) according to the manufacturer’s instructions. Integrity of RNA was determined by gel electrophoresis, whereas RNA concentration (OD260) and purity (OD260/OD280) were measured using an ND–1000 UV–vis spectrophotometer (NanoDrop Technologies, Rockland, Del). One microgram of total RNA was subsequently used for the synthesis of cDNA with SuperScript III RNase reverse transcriptase (Invitrogen, Breda, The Netherlands) in 20 μL final volume containing 250 ng of random hexamers (Promega, Leiden, The Netherlands) and 40 units of RNase OUT inhibitor (Invitrogen). One microliter cDNA was used for each PCR reaction. The Assay-on-Demand primers purchased from Applied Biosystems (Nieuwerkerk aan den IJssel, The Netherlands) for quantitative PCR included housekeeping gene GAPDH (glyceraldehyde-3-phosphate dehydrogenase, assay ID Hs99999905_m1), CD31 (platelet endothelial cell adhesion molecule, PECAM-1, assay ID Hs00169777_m1), VE-cad (VE-cadherin, assay ID Hs00174344_m1), KLF2 (Kruppel-like factor-2, assay ID Hs00360439_g1), Tie2 (assay ID Hs00176096_m1), Ang2 (Angiopoietin-2, assay ID Hs00169867_m1), E-selectin (assay ID Hs00174057_m1), VCAM-1 (assay ID Hs00365486_m1), ICAM-1 (assay ID Hs00164932_m1), and IL-8 (assay ID Hs00174103_m1). Quantitative PCR was performed in a ViiA 7 real-time PCR System (Applied Biosystems, Nieuwerkerk aan den IJssel, The Netherlands). Amplification was performed using the following cycling conditions: 15 min 95°C and 40 two-step cycles of 15 s at 95°C and 60 s at 60°C.

Duplicate real-time PCR analyses were executed for each sample, and the obtained threshold cycle values (CT) were averaged. According to the comparative CT method described in the ABI manual, gene expression was normalized to the expression of the housekeeping gene, yielding the ΔCT value. The average messenger RNA (mRNA) level relative to GAPDH was calculated by 2-ΔCT.

Data show mean ± SD of one experiment with three replicates (n = 3), which is representative of three independent experiments.

Flow cytometric analysis

After incubation at different conditions, HUVECs were detached with Trypsin/EDTA and subsequently resuspended in 5% FCS solution to neutralize Trypsin/EDTA. After washing with PBS/5% FCS, HUVECs were incubated with monoclonal antibodies against E-selectin, VCAM-1, and ICAM-1 (hybridoma supernatants kindly provided by Dr. M. Gimbrone, Boston, Mass) or with monoclonal antibody against CD31 (dilution 1:25; DakoCytomation, The Netherlands) for 45 min on ice followed by 45 min incubation with FITC-labeled rabbit anti-mouse antibodies (dilution 1:50; Jackson Immunoresearch, Suffolk, UK) on ice. Cells were fixed with 0.5% paraformaldehyde in PBS, and the fluorescence was detected by flow cytometry using the FACSCalibur (Becton Dickinson, The Netherlands). Nonspecific binding was assessed by staining with irrelevant isotype-matched monoclonal antibody, and fluorescence intensity was measured in arbitrary units corrected by isotype control.

Data show mean ± SD of one experiment with three replicates (n = 3), which is representative of three independent experiments.

Quantification of IL-8 protein levels by ELISA

To quantify the concentration of IL-8 protein in the medium under different conditions, medium supernatants were collected and the protein expression of IL-8 was quantified using human IL-8 enzyme-linked immunosorbent assay (ELISA) max standard sets (Biolegend, UK), according to the manufacturer’s instructions. In graphs, IL-8 levels were expressed as picograms per milliliter per square centimeter total cell surface to normalize for cell surface differences in different experimental conditions.

Statistical analysis

Statistical significance of differences was studied by means of the Student t test or a one-way analysis of variance followed by a Bonferroni correction for selected pairs of columns. All statistical analyses were performed using GraphPad Prism software (GraphPad Prism Software Inc., San Diego, Calif). Differences were considered to be significant when P < 0.05.


The expression of endothelial genes is affected by prolonged LSS

Because LSS exposure for 1 to 2 days is essential for endothelial cells in vitro to develop maximal responses to the changes of flow in a more physiological way (11), endothelial cells in this study were allowed to be adapted to 48 h of LSS before they were subsequently subjected to flow alterations (Fig. 1B).

We first investigated the response of endothelial cells to prolonged LSS exposure at 20 dyne/cm2 for 48 h. As shown in Figure 2, HUVECs cultured under LSS showed a significantly increased mRNA level of KLF2 compared with static control, which is in agreement with other studies (11, 12). At the same time, Tie2 was induced more than 3.5-fold by prolonged LSS, whereas Ang2 was downregulated by more than 90% of initial expression in static conditions. Flow exposure suppressed the expression of adhesion molecules E-selectin and VCAM-1 as well as chemokine IL-8. In contrast, ICAM-1 expression was induced. The expression of CD31 and VE-cadherin in endothelial cells was unchanged during long-term flow exposure. Similar results were found at 24 h, although the changes were less extensive (data not shown).

Fig. 2
Fig. 2:
The expression of endothelial genes is affected by prolonged LSS. Human umbilical vein endothelial cells were kept under static conditions or exposed to LSS at 20 dyne/cm2 for 48 h before gene expression analysis. Messenger RNA levels of genes were determined by quantitative RT-PCR using GAPDH as the housekeeping gene. Values represent fold changes compared with those of static control. Data are expressed as mean ± SD, n = 3. *P < 0.05, LSS exposure vs. static control.

Flow cessation after LSS preadaptation leads to a proinflammatory response of endothelial cells

To study the effect of loss of flow to represent the shock period in extensive hemorrhage, after initial 48-h exposure of HUVECs to 20 dyne/cm2 LSS flow was ceased completely for different periods (Fig. 3A). The gene expression results in Figure 3B showed that the expression of flow sensor KLF2 was rapidly abolished by flow cessation. In contrast, Tie2 expression was gradually downregulated after an initial slight increase, whereas the loss of flow significantly upregulated Ang2 expression starting at a later time point, that is, at 8 h, and continuing up to 24 h. E-selectin, VCAM-1, and IL-8 were induced by the loss of flow compared with LSS control, with an apparent peak expression of all three genes at 8 h. After 24-h loss of shear stress, VCAM-1 expression was still as high as its 8-h expression level, whereas the expression of E-selectin and IL-8 had decreased but were still significantly higher than their levels under LSS conditions. The expression of ICAM-1, which was the only cellular adhesion molecule that was upregulated under LSS condition, was decreased on loss of flow. Neither CD31 nor VE-cadherin expression was markedly affected by flow cessation in vitro.

Fig. 3
Fig. 3:
Flow cessation after LSS preadaptation leads to a proinflammatory response of endothelial cells. A, Human umbilical vein endothelial cells were kept under LSS (20 dyne/cm2) for 48 h and next subjected to different periods of flow cessation before mRNA analysis. B, Gene expression levels were determined by quantitative RT-PCR using GAPDH as the housekeeping gene. Values represent mRNA levels relative to LSS control. Data are shown as mean ± SD, n = 3. *P < 0.05, flow cessation vs. LSS control. C, The concentration of IL-8 in the medium after 8 h of flow cessation was determined by ELISA. Data are shown as mean of three independent experiments. *P < 0.05, LSS vs. static control; # P < 0.05, 8 h of flow cessation vs. LSS. D, The protein expression levels of surface adhesion molecules were analyzed by flow cytometry and shown as mean fluorescence intensity (MFI) arbitrary units corrected by isotype control. Data are expressed as mean ± SD, n = 3. *P < 0.05, LSS vs. static control; # P < 0.05, 8 h or 24 h of flow cessation vs. LSS.

At the protein level, the concentration of IL-8 in the medium of cells exposed to LSS was decreased compared with that of the static control, whereas subsequent exposure to 8-h flow cessation induced its upregulation (Fig. 3C). Protein expression of ICAM-1 followed its mRNA changes, that is, increased protein levels were measured on LSS exposure whereas they decreased on flow cessation (Fig. 3D). At the same time, the expression levels of E-selectin and VCAM-1 protein did not change compared with baseline, which might relate to the limited extent of mRNA expression induction on flow cessation. We also examined the expression of adhesion molecules at 24 h, showing similar results as at the 8-h time point (Fig. 3D). Angiopoietin 2 protein that is stored in Weibel-Palade bodies may have been released into the medium on the loss of flow; we, however, did not assess its concentration in the circulating medium in this study.

Although focusing on complete flow cessation, we found that a reduction of shear stress from 20 dyne/cm2 to 5 dyne/cm2 also leads to endothelial activation. Under this condition, the proinflammatory molecules and Ang2 were induced, whereas Tie2 showed a significant downregulation (Fig. 4), the latter being in agreement with a previous study (13).

Fig. 4
Fig. 4:
Endothelial responses to acute reduction of LSS to 5 dyne/cm2. Human umbilical vein endothelial cells were exposed to LSS (20 dyne/cm2) for 48 h, and then shear stress was reduced to 5 dyne/cm2 for 8 h. Cells cultured in static condition were taken as control. Genes were determined by quantitative RT-PCR, and GAPDH was taken as the housekeeping gene. Values represent fold change of genes in comparison with the levels in static control. Bars show mean ± SD, n = 3. *P < 0.05 vs. 48 h 20 dyne/cm2.

Thus, endothelial cells are quickly activated by flow reduction/cessation after long-term adaptation to laminar flow, although the kinetics of changes in the expression of the genes involved in vascular integrity as well as proinflammatory activation of endothelium are variable.

Abrupt reflow–associated endothelial responses after flow cessation

Fluid resuscitation to improve microvascular blood flow is an essential therapy for the treatment of any form of shock (2). Our previous in vivo study showed that endothelial proinflammatory status was significantly induced by fluid resuscitation after hemorrhagic shock (3). To investigate the role of shear stress changes by itself in the process of endothelial proinflammatory activation during resuscitation, we therefore examined the effect of acute reflow on the endothelium (Fig. 5A). Considering that prolonged no-reflow does exist because of reperfusion defects after hemorrhagic shock in the clinic (14), we continued our study with the 8-h flow cessation time point.

Fig. 5
Fig. 5:
Abrupt reflow associated endothelial responses after initial flow cessation. A, To mimic the resuscitation procedure after hemorrhagic shock in vivo, LSS-preadapted HUVECs were subjected to LSS (20 dyne/cm2) for 0.5, 1, and 4 h after an 8-h flow cessation period. B, Messenger RNA was harvested, and genes were determined by quantitative RT-PCR using GAPDH as the housekeeping gene. Values represent fold changes of genes in comparison with the levels after 8-h flow cessation conditions. Data are expressed as mean ± SD, n = 3. *P < 0.05, 0.5-, 1-, 4-h reflow vs. flow cessation, which was set as 100%.

On reflow, KLF2 expression was rapidly regained (Fig. 5B). Interestingly, Tie2, which was downregulated during flow cessation, was decreased even further by reflow within the 4-h period studied. On the other hand, the induction of adhesion molecules by 0.5-h reflow after 8-h flow cessation suggests a quick proinflammatory effect of rapid onset of reflow. Four hours after the start of reflow, VCAM-1, IL-8, and Ang2 were reduced significantly, whereas ICAM-1 remained elevated. In addition, reflow did not affect the expression of CD31 and VE-cadherin.

The proinflammatory response of endothelial cells to TNF-α stimulation was enhanced by 0.5 h of reflow

Cytokine production is a prominent systemic inflammatory response during hemorrhagic shock and resuscitation. After hemorrhage, tumor necrosis factor-α (TNF-α) production shows an early increase at the 1-h time point and remains high until 5 h (15). Furthermore, systemic TNF-α levels induced during hemorrhagic shock are further increased by subsequent resuscitation (6). It is not known whether the changes of flow during circulatory shock and resuscitation will influence the proinflammatory consequences of endothelial cells induced by cytokine exposure. To study this, LSS adapted endothelial cells were stimulated with TNF-α for the final 4 h of flow cessation period and next subjected to 0.5-h reflow exposure or maintained without reflow (Fig. 6A). As shown in Figure 6B, the additional drop in KLF2 mRNA level by 4-h TNF-α exposure during flow cessation was rapidly increased on 0.5-h reflow to the similar level as it was under flow cessation conditions. Tie2 and Ang2, both not responsive to TNF-α per se, showed a further reduction, respectively, induction in their expression by the installation of abrupt reflow. At the same time, reflow increased the expression levels of proinflammatory molecules E-selectin, VCAM-1, and ICAM-1, as well as IL-8 beyond the levels induced by the TNF-α challenge during flow cessation. The protein expression data corroborated that subsequent acute reflow further upregulated the protein levels of VCAM-1 and ICAM-1 induced by TNF-α exposure during flow cessation (Fig. 6C). In addition, there is a tendency to increased IL-8 production on reflow (Fig. 6D). These results indicate that TNF-α exposure of endothelial cells during flow cessation strongly activates the cells, and that the application of subsequent reflow aggravates this cytokine-induced endothelial response.

Fig. 6
Fig. 6:
TNF-α–induced proinflammatory activation of endothelial cells during flow cessation is aggravated by subsequent abrupt reflow. A, Forty-eight–hour LSS (20 dyne/cm2)–preconditioned human umbilical vein endothelial cells were subjected to 8-h flow cessation with or without subsequent 0.5-h reflow (20 dyne/cm2), and TNF-α (10 ng/mL) was added for the last 4 h under both conditions. B, Cells were lysed for mRNA isolation, and genes were analyzed by quantitative RT-PCR using GAPDH as the housekeeping gene. Values represent fold change of genes in comparison with the levels under 8-h flow cessation set as 1. Results are expressed as mean ± SD, n = 3. *P < 0.05, 4-h TNF-α under flow cessation vs. flow cessation control; # P < 0.05, 4-h TNF-α with reflow vs. 4-h TNF-α without reflow. C, The expression levels of adhesion molecules on the endothelial surface were analyzed by flow cytometry and shown as mean fluorescence intensity (MFI) arbitrary units corrected by isotype control. Data are expressed as mean ± SD, n = 3. *P < 0.05, vs. flow cessation condition alone; # P < 0.05, significant difference between with and without abrupt reflow exposure. D, The concentration of IL-8 in the medium was determined by ELISA. Data are shown as mean of three independent experiments.


Fluctuations in blood flow during shock and resuscitation lead to flow disturbances in vessels. In parallel, endothelial proinflammatory activation and vascular leakage occur (3, 16). These activations are more extensive at 1 h after resuscitation compared with the endothelial priming observed at 90 min of hemorrhagic shock (3). In this in vitro study, we addressed whether, and to what extent, the changes in shear stress resulting from the loss as well as the subsequent regaining of blood flow during hemorrhagic shock and resuscitation contribute to endothelial activation. Using an in vitro model to mimic the flow changes as occurring during hemorrhagic shock and resuscitation after long-term LSS exposure, we here demonstrated that flow changes per se affected the expression of proinflammatory adhesion molecules by the endothelial cells. Each gene studied showed its own kinetic of expression on the challenge of flow cessation and reflow. For these proinflammatory molecules E-selectin, VCAM-1, ICAM-1, as well as IL-8, however, the effects of flow alteration per se were minor compared with their expression induced by TNF-α challenge. Yet, the induction of proinflammatory adhesion molecules by TNF-α stimulation during flow cessation was significantly higher because of the combined exposure of reflow, implying that both blood flow changes and proinflammatory cytokines may contribute to endothelial activation during clinical shock and resuscitation. In addition, we showed that at 4 h after reflow Tie2 had not regained its normal expression under LSS control, and that the combined exposure of reflow and TNF-α extended the downregulation of Tie2 and upregulation of Ang2 expression (Fig. 7).

Fig. 7
Fig. 7:
Summary of the effects of shear stress changes in the absence/presence of TNF-α on endothelial gene expression in vitro. Genes associated with endothelial integrity (Tie2, Ang2) and inflammatory responses (E-selectin, VCAM-1, ICAM-1, and IL-8) that were affected by the studied experimental conditions are shown. VE-cadherin and CD31 mRNA levels did not change under the conditions studied (data not shown in this table). Effects of TNF-α exposure during flow cessation, respectively, flow cessation followed by reflow are given as fold changes of mRNA expression (*fold change versus flow cessation control). ↑↑, >10-fold induction; ↑, <10-fold induction; ↔, no change; ↓, <50% reduction; ↓↓, >50% reduction.

Hemorrhagic shock/resuscitation can be conceived as a systemic ischemia/reperfusion injury insult, and the extent of ischemia in tissues during hemorrhagic shock is considered a major determinant of the systemic inflammatory response (17). We demonstrated in the present study that flow changes per se, that is, flow cessation and reflow implementation, induce proinflammatory activation of endothelial cells. The flow characteristics in different microvascular beds from different vital organs in health and disease are not exactly known. It has been reported that microcirculatory responses to hemorrhagic shock are dependent on the vessel type. After a hemorrhage period, slow and constant blood flow existed in some preferential capillary channels whereas other capillaries were eliminated from the circulation, and the blood flow in arterioles ceased several times (18). This indicates that endothelial cells from different microvascular beds likely experience variable flow changes during hemorrhagic shock. Similarly, hypotension is a hallmark of septic shock, and microcirculatory alterations characterized by an increased number of intermittent- or stopped-flow capillaries may play an important role in sepsis-associated organ dysfunction (19). In addition, heterogeneity of microvascular endothelial cells likely underlies high cell-to-cell variability in their adaptation to pathology-related microenvironmental changes, which can be organ specific and microvascular bed specific (20). The molecular nature of heterogeneous responses of endothelial cells to microcirculatory changes per se are largely unknown and difficult to simulate in vitro. An important follow-up study will therefore be to investigate the effects of different shock forms like hemorrhagic shock and septic shock in animal models and to study in detail the different microvascular segments within organs using, for example, quantification of gene expression in laser microdissected microvascular segments and intravital microscopy to study blood flow specificities in different segments (21, 22).

Several other in vitro studies have reported on the relation between shear stress and endothelial activation, with the majority focusing on effects of shear stress loss. Krizanac-Bengez et al. (23) found that, under normoxic and normoglycemic conditions, loss of shear stress was able to independently induce leukocyte-mediated proinflammatory activation in brain microvascular endothelial cells, leading to failure of the blood-brain barrier function. Wei and colleagues (24) reported that flow-adapted bovine pulmonary artery endothelial cells (at 1 dyne/cm2 for 48 h) generated reactive oxygen species (ROS) on 1-h flow cessation (mimicking ischemia), which resulted in the activation of NF-κB and AP-1. They furthermore showed that, in flow-adapted bovine pulmonary artery endothelial cells (5 dyne/cm2, 24 h), the sudden removal of shear stress led to ROS-dependent phosphorylation of ERK1/2 as well as Ca2+-dependent NO generation within 60 s (25). Nuclear factor-κB, AP-1, and ERK1/2 are also involved in the induction of proinflammatory activation of endothelial cells (26, 27), and NF-κB shows time-dependent activation during hemorrhagic shock and resuscitation in mice (28). Therefore, it can be speculated that, in our study, the activation of these kinases/transcription factors is associated with the endothelial activation observed. Because endothelial cells responded to flow changes with a similar gene expression profile as that observed in cytokine stimulation, we measured TNF-α production in this in vitro system and found that its mRNA level was undetectable (data not shown). The exact molecular control of endothelial activation during flow cessation as well as reflow exposure are as of yet unknown and will be investigated in future studies using pharmacological inhibition of these different pathways.

KLF2 plays an important role in maintaining a quiescent endothelial phenotype. The current study showed that the upregulation of KLF2 by long-term shear stress adaptation was rapidly abolished by flow cessation. Interestingly, the ability of 0.5-h reflow to restore KLF2 expression was inhibited by TNF-α challenge (see Figure, Supplemental Digital Content 1, which demonstrates the expression of KLF2 on 0.5-h TNF-α exposure under flow cessation or reflow conditions), implying a diminished adaptive capacity of endothelial cells to cytokine stimulation under abrupt reflow conditions. Gracia-Sancho et al. (29) have shown that endothelial cells with reduced KLF2 expression on flow cessation under cold storage condition (4 degree) show higher responsiveness to cytokine IL-1β. Possibly, the TNF-α–related inhibition of KLF2 reexpression on reflow in our study contributes to the relatively higher responsiveness of endothelial activation under reflow conditions. Clinically, this may be relevant to sepsis where flow changes and cytokines are both abundantly present. It is worthwhile to follow up on this observation and to investigate the in vivo kinetics of KLF2 expression during hemorrhagic shock and resuscitation and to verify whether a diminished KLF2 reexpression by TNF-α is involved in the underlying molecular mechanisms of endothelial activation during resuscitation.

There is accumulating evidence emphasizing the critical role of the Ang/Tie2 system in vascular dysfunction during critical illness. We previously showed in vivo that Tie2 expression was reduced at 4 h after hemorrhagic shock and resuscitation and normalized at 24 h (7). In the present study, we showed that the removal of shear stress in vitro caused loss of Tie2, and that the subsequent acute readministration of flow did not recover its expression. At the same time, Ang2 was induced whereas the reexpression of KLF2 was abolished by the TNF-α challenge, showing an inverse relationship between KLF2 and Ang2 that was previously described in a study with KLF2-overexpressing endothelial cells (30). The decrease of Tie2 as well as the induction of Ang2 caused by flow alterations may thus contribute to microvascular dysfunction, including increased vascular leakage and extended proinflammatory activation as observed in vivo.

Our current study has several limitations that should be kept in mind. First of all, it is difficult to directly translate the in vitro flow-induced endothelial changes to the bed of the shock patient. Beyond the studied flow effects, other processes will likely occur simultaneously to induce endothelial activation during shock and resuscitation. Among others, impaired oxygen delivery induced tissue hypoxia (31), leukocyte activation (5), and cytokine release (6) are all known to be involved in endothelial activation. At the same time, exclusion of other cell types coexisting in vivo in the microenvironment of the endothelium may also make that the in vitro flow model dose not recapitulate the in vivo situation during shock and resuscitation. Second, unidirectional laminar flow has been used throughout our present study, whereas in the shock patient, the different features of shear stress in various vessel types, for example, arterial bifurcations and stenotic vessels, can influence downstream endothelial behavior (32). Third, we have used 20- to 5-dyne/cm2 and 20- to 0-dyne/cm2 flow changes in our study. Although the shear stress values in various vessels have been estimated (33), the exact flow magnitudes and patterns in distinct macrovascular and microvascular beds under healthy condition as well as in critically ill patients are not known. Fourth, we chose specifically predefined time points of flow cessation and reflow to compare this in vitro study with our in vivo hemorrhagic shock (3) and lipopolysaccharide exposure mouse models (13). In patients, this time frame may be shorter or much longer because microvascular flow derangements can be limited to a few seconds (preoperative hypotension after anesthesia induction) to several minutes (massive surgical bleeding) or be present for days (critically ill sepsis patients). Therefore, only carefully combining our in vitro data presented here with data from animal shock model and human organ biopsies of deceased patients (34) will help to bridge this bench-to-bedside gap.

Collectively, our study revealed that flow alterations per se, as experienced by endothelial cells during microcirculatory shock and resuscitation, act as a proinflammatory stimulus, and that the flow changes in combination with other proinflammatory factors such as TNF-α that coexist in critically ill patients likely interact with each other. The abrupt reflow–related enhancement of cytokine-induced endothelial inflammatory activation suggests that inflammation-related signaling cascades at the time of resuscitation may be potential targets of pharmacological interventions to attenuate endothelial activation and the development of consequent multiple organ failure. Moreover, the changes in expression of Ang2/Tie2 by flow alterations observed in this study indicate that improving organ perfusion, a hallmark of modern critical care, is likely to positively affect microvascular function. In addition, it has been shown that phosphorylation-dependent activation of Tie2 can be induced by the onset of shear stress (35). This Tie2 phosphorylation on flow cessation and reflow will be further investigated in our in vitro flow study as well as in vivo models (hemorrhagic shock and septic shock mice, patients) in the future. Because the flow alteration–associated endothelial dysfunction may be a critical determinant for the development of pathological changes during shock and resuscitation as well as for postresuscitation organ function, future studies will be executed to investigate which molecular pathways can be therapeutically interfered with in the microvasculature.


The authors thank Henk E. Moorlag, Rianne M. Jongman, and Peter J. Zwiers (University Medical Center Groningen) for excellent technical assistance. The authors also thank Dr. Guido Krenning, Dr. Jan-Renier Moonen, and Ms. Ee Soo Lee (University Medical Center Groningen) for their support and help with the use of the ibidi flow system.


1. Wasserman SM, Topper JN: Adaptation of the endothelium to fluid flow: in vitro analyses of gene expression and in vivo implications. Vasc Med 9 (1): 35–45, 2004.
2. Vincent JL, De Backer D: Circulatory shock. N Engl J Med 369 (18): 1726–1734, 2013.
3. van Meurs M, Wulfert FM, Knol AJ, De Haes A, Houwertjes M, Aarts LP, Molema G: Early organ-specific endothelial activation during hemorrhagic shock and resuscitation. Shock 29 (2): 291–299, 2008.
4. Muller WA: Mechanisms of transendothelial migration of leukocytes. Circ Res 105 (3): 223–230, 2009.
5. Douzinas EE, Livaditi O, Tasoulis MK, Pelekanou A, Giamarellos-Bourboulis EJ: Stimulation of monocytes is a pathway involved in systemic inflammatory response following haemorrhagic shock resuscitation: the effect of hypoxaemic resuscitation. Clin Exp Immunol 150 (3): 502–508, 2007.
6. Lee CC, Chang IJ, Yen ZS, Hsu CY, Chen SY, Su CP, Chiang WC, Chen SC, Chen WJ: Delayed fluid resuscitation in hemorrhagic shock induces proinflammatory cytokine response. Ann Emerg Med 49 (1): 37–44, 2007.
7. van Meurs M, Kurniati NF, Wulfert FM, Asgeirsdottir SA, De Graaf IA, Satchell SC, Mathieson PW, Jongman RM, Kumpers P, Zijlstra JG, et al.: Shock-induced stress induces loss of microvascular endothelial Tie2 in the kidney which is not associated with reduced glomerular barrier function. Am J Physiol Renal Physiol 297 (2): F272–F281, 2009.
8. Tressel SL, Kim H, Ni CW, Chang K, Velasquez-Castano JC, Taylor WR, Yoon YS, Jo H: Angiopoietin-2 stimulates blood flow recovery after femoral artery occlusion by inducing inflammation and arteriogenesis. Arterioscler Thromb Vasc Biol 28 (11): 1989–1995, 2008.
9. Dejana E, Orsenigo F, Lampugnani MG: The role of adherens junctions and VE-cadherin in the control of vascular permeability. J Cell Sci 121 (Pt 13): 2115–2122, 2008.
10. Carrithers M, Tandon S, Canosa S, Michaud M, Graesser D, Madri JA: Enhanced susceptibility to endotoxic shock and impaired STAT3 signaling in CD31-deficient mice. Am J Pathol 166 (1): 185–196, 2005.
11. Fledderus JO, Van Thienen JV, Boon RA, Dekker RJ, Rohlena J, Volger OL, Bijnens AP, Daemen MJ, Kuiper J, Van Berkel TJ, et al.: Prolonged shear stress and KLF2 suppress constitutive proinflammatory transcription through inhibition of ATF2. Blood 109 (10): 4249–4257, 2007.
12. Dekker RJ, Van Thienen JV, Rohlena J, De Jager SC, Elderkamp YW, Seppen J, De Vries CJ, Biessen EA, Van Berkel TJ, Pannekoek H, et al.: Endothelial KLF2 links local arterial shear stress levels to the expression of vascular tone-regulating genes. Am J Pathol 167 (2): 609–618, 2005.
13. Kurniati NF, Jongman RM, Hagen F, Spokes KC, Moser J, Regan ER, Krenning G, Moonen J-RaJ, Harmsen MC, Struys MMRF, et al.: The flow dependency of Tie2 expression in endotoxemia. Intensive Care Med 39 (7): 1262–1271, 2013.
14. Kloner RA: No-reflow phenomenon: maintaining vascular integrity. J Cardiovasc Pharmacol Ther 16 (3–4): 244–250, 2011.
15. Sato H, Tanaka T, Tanaka N: The effect of p38 mitogen-activated protein kinase activation on inflammatory liver damage following hemorrhagic shock in rats. PLoS One 7 (1): e30124, 2012.
16. Childs EW, Tharakan B, Byrge N, Tinsley JH, Hunter FA, Smythe WR: Angiopoietin-1 inhibits intrinsic apoptotic signaling and vascular hyperpermeability following hemorrhagic shock. Am J Physiol Heart Circ Physiol 294 (5): H2285–H2295, 2008.
17. Douzinas EE, Andrianakis I, Livaditi O, Paneris P, Tasoulis M, Pelekanou A, Betrosian A, Giamarellos-Bourboulis EJ: The level of hypotension during hemorrhagic shock is a major determinant of the post-resuscitation systemic inflammatory response: an experimental study. BMC Physiol 8: 15, 2008.
18. Colantuoni A, Bertuglia S, Intaglietta M: Microvessel diameter changes during hemorrhagic shock in unanesthetized hamsters. Microvasc Res 30 (2): 133–142, 1985.
19. Waisman D, Abramovich A, Brod V, Lavon O, Nurkin S, Popovski F, Rotschild A, Bitterman H: Subpleural microvascular flow velocities and shear rates in normal and septic mechanically ventilated rats. Shock 26 (1): 87–94, 2006.
20. Langenkamp E, Molema G: Microvascular endothelial cell heterogeneity: general concepts and pharmacological consequences for anti-angiogenic therapy of cancer. Cell Tissue Res 335 (1): 205–222, 2009.
21. Langenkamp E, Kamps JA, Mrug M, Verpoorte E, Niyaz Y, Horvatovich P, Bischoff R, Struijker-Boudier H, Molema G: Innovations in studying in vivo cell behavior and pharmacology in complex tissues—microvascular endothelial cells in the spotlight. Cell Tissue Res 354 (3): 647–669, 2013.
22. Huang TC, Lin WC, Wu CC, Zhang G, Lin KP: Experimental estimation of blood flow velocity through simulation of intravital microscopic imaging in micro-vessels by different image processing methods. Microvasc Res 80 (3): 477–483, 2010.
23. Krizanac-Bengez L, Mayberg MR, Cunningham E, Hossain M, Ponnampalam S, Parkinson FE, Janigro D: Loss of shear stress induces leukocyte-mediated cytokine release and blood-brain barrier failure in dynamic in vitro blood-brain barrier model. J Cell Physiol 206 (1): 68–77, 2006.
24. Wei Z, Costa K, Al-Mehdi AB, Dodia C, Muzykantov V, Fisher AB: Simulated ischemia in flow-adapted endothelial cells leads to generation of reactive oxygen species and cell signaling. Circ Res 85 (8): 682–689, 1999.
25. Wei Z, Al-Mehdi AB, Fisher AB: Signaling pathway for nitric oxide generation with simulated ischemia in flow-adapted endothelial cells. Am J Physiol Heart Circ Physiol 281 (5): H2226–H2232, 2001.
26. Kempe S, Kestler H, Lasar A, Wirth T: NF-kappaB controls the global pro-inflammatory response in endothelial cells: evidence for the regulation of a pro-atherogenic program. Nucleic Acids Res 33 (16): 5308–5319, 2005.
27. Furst R, Blumenthal SB, Kiemer AK, Zahler S, Vollmar AM: Nuclear factor-kappa B-independent anti-inflammatory action of salicylate in human endothelial cells: induction of heme oxygenase-1 by the c-jun N-terminal kinase/activator protein-1 pathway. J Pharmacol Exp Ther 318 (1): 389–394, 2006.
28. Korff S, Falsafi R, Czerny C, Jobin C, Nau C, Jakob H, Marzi I, Lehnert M: Time dependency and topography of hepatic nuclear factor kappaB activation after hemorrhagic shock and resuscitation in mice. Shock 38 (5): 486–492, 2012.
29. Gracia-Sancho J, Villarreal G Jr, Zhang Y, Yu JX, Liu Y, Tullius SG, Garcia-Cardena G: Flow cessation triggers endothelial dysfunction during organ cold storage conditions: strategies for pharmacologic intervention. Transplantation 90 (2): 142–149, 2010.
30. Van Agtmaal EL, Bierings R, Dragt BS, Leyen TA, Fernandez-Borja M, Horrevoets AJ, Voorberg J: The shear stress–induced transcription factor KLF2 affects dynamics and angiopoietin-2 content of Weibel-Palade bodies. PLoS One 7 (6): e38399, 2012.
31. Mccloskey CA, Kameneva MV, Uryash A, Gallo DJ, Billiar TR: Tissue hypoxia activates JNK in the liver during hemorrhagic shock. Shock 22 (4): 380–386, 2004.
32. Cicha I, Beronov K, Ramirez EL, Osterode K, Goppelt-Struebe M, Raaz D, Yilmaz A, Daniel WG, Garlichs CD: Shear stress preconditioning modulates endothelial susceptibility to circulating TNF-alpha and monocytic cell recruitment in a simplified model of arterial bifurcations. Atherosclerosis 207 (1): 93–102, 2009.
33. Papaioannou TG, Stefanadis C: Vascular wall shear stress: basic principles and methods. Hellenic J Cardiol 46 (1): 9–15, 2005.
34. Aslan A, Jongman RM, Moser J, Stegeman CA, Van Goor H, Diepstra A, Van Den Heuvel MC, Heeringa P, Molema G, Zijlstra JG: The renal angiopoietin/Tie2 system in lethal human sepsis. Crit Care 18 (2): 423, 2014.
35. Jong Lee H, Young Koh G: Shear stress activates Tie2 receptor tyrosine kinase in human endothelial cells. Biochem Biophys Res Commun 304 (2): 399–404, 2003.

Microcirculatory hypoperfusion; hemorrhagic shock; endothelial cells; laminar shear stress; flow cessation; reflow; proinflammatory activation

Supplemental Digital Content

© 2014 by the Shock Society