Membrane cholesterol modulates the fluid shear stress response of polymorphonuclear leukocytes via its effects on membrane fluidity

Xiaoyan Zhang, Jonathan Hurng, Debra L. Rateri, Alan Daugherty, Geert W. Schmid-Schönbein, Hainsworth Y. Shin


Continuous exposure of polymorphonuclear leukocytes (PMNLs) to circulatory hemodynamics points to fluid flow as a biophysical regulator of their activity. Specifically, fluid flow-derived shear stresses deactivate leukocytes via actions on the conformational activities of proteins on the cell surface. Because membrane properties affect activities of membrane-bound proteins, we hypothesized that changes in the physical properties of cell membranes influence PMNL sensitivity to fluid shear stress. For this purpose, we modified PMNL membranes and showed that the cellular mechanosensitivity to shear was impaired whether we increased, reduced, or disrupted the organization of cholesterol within the lipid bilayer. Notably, PMNLs with enriched membrane cholesterol exhibited attenuated pseudopod retraction responses to shear that were recovered by select concentrations of benzyl alcohol (a membrane fluidizer). In fact, PMNL responses to shear positively correlated (R2 = 0.96; P < 0.0001) with cholesterol-related membrane fluidity. Moreover, in low-density lipoprotein receptor-deficient (LDLr−/−) mice fed a high-fat diet (a hypercholesterolemia model), PMNL shear-responses correlated (R2 = 0.5; P < 0.01) with blood concentrations of unesterified (i.e., free) cholesterol. In this regard, the shear-responses of PMNLs gradually diminished and eventually reversed as free cholesterol levels in blood increased during 8 wk of the high-fat diet. Collectively, our results provided evidence that cholesterol is an important component of the PMNL mechanotransducing capacity and elevated membrane cholesterol impairs PMNL shear-responses at least partially through its impact on membrane fluidity. This cholesterol-linked perturbation may contribute to dysregulated PMNL activity (e.g., chronic inflammation) related to hypercholesterolemia and causal for cardiovascular pathologies (e.g., atherosclerosis).

  • flow
  • cell deactivation
  • pseudopod retraction
  • mechanotransduction
  • hypercholesterolemia

whether adhered to a surface or freely suspended in plasma, polymorphonuclear leukocytes (PMNLs; particularly neutrophils) sense and respond to fluid shear stress (force per unit area acting tangential to the cell surface; dyn/cm2) imposed by blood flow. Specifically, fluid shear stress maintains PMNLs in a deactivated state by reducing or preventing formation of pseudopod(s), a morphological hallmark of an activated PMNL, while at the same time, promoting cleavage of cell surface-associated β2 (CD18) integrins involved in cell adhesion and migration (30, 48). The cell-inactivating action of shear stress is depressed upon treatment of cells with threshold concentrations of biochemical agonists, such as formyl peptide (N-formyl-Met-Leu-Phe or FMLP; > 0.01 μM) and platelet-activating factor (>0.1 μM) (15), pointing to circulatory fluid flow as a negative mechano-regulator of PMNL activity in the absence of inflammatory agents, i.e., cell stimulation overrides this negative fluid shear stress control. Moreover, impaired leukocyte responses to shear stress, due to agonist stimulation or a pathologically inflamed state, have been linked to increased PMNL entrapment in the microcirculation of normotensive rats (15, 34) and elevations in peripheral vascular resistances of spontaneously hypertensive (14) or glucocorticoid-treated (13) rats. These observations are consistent with reported correlations between chronic cell activation and reductions in leukocyte deformability that influence the rheology of microvascular blood flow (33, 54). Deficiency in leukocyte mechanotransduction may, therefore, be a contributing factor in the onset or progression of cardiovascular pathologies.

Notably, leukocyte responses to shear stress take place under shear magnitudes incapable of eliciting significant passive viscoelastic deformation of the cell (51). This characteristic points to a transducing element(s) present on the outer surface of the plasma membrane that converts shear stresses or gradients into biological cues. In this regard, recent evidence points to two families of membrane proteins that serve as putative mechanosensors for PMNLs: G protein-coupled receptors (GPCRs) and CD18 integrins. GPCRs (e.g., formyl peptide receptor or FPR) exposed to shear exhibit conformational shifts leading to pseudopod retraction likely via the influence on rac1-mediated actin polymerization (28, 29). CD18 also undergoes structural changes under the actions of shear and, in doing so, promotes its cleavage from the surfaces of both nonadherent and migrating leukocytes (48). Since these proteins are anchored in the cell membrane, shear stress-induced shifts in the conformation of these mechanosensors may be influenced by the composition and/or physical properties of the lipid bilayer. On the basis of this premise, the present study assessed the effects of changes in membrane properties on the sensitivity of PMNLs to shear stress.

Cholesterol is an essential structural component of cell membranes that plays a critical role in regulating cell signaling and functions. Depleting cholesterol from plasma membranes with methyl-β-cyclodextrin (MβCD) suppresses FMLP-induced ruffling, polarization, and F-actin polymerization in differentiated neutrophilic HL-60 cells and PMNLs (37, 39). Similar to cholesterol depletion, sequestration of cholesterol (and thus disruption of its organization) within the membrane, using filipin, influences activation of CD18 integrins leading to impaired T-cell adhesion (32). In this fashion, changes in membrane cholesterol have been shown to interfere with signaling related to lipid rafts, which provide spatial organization for select transmembrane proteins and act as signaling platforms responsible for coordinating outside-in and inside-out signal transduction (21, 32, 49). Interestingly, these cholesterol-rich microdomains have also been implicated as mechanotransduction centers such as caveolae, a subtype of lipid rafts that play a role in the mechanotransduction of shear stress and pressure by endothelial cells (41, 43, 44).

Moreover, the content of cholesterol within lipid bilayer determines the physical properties of membranes, such as membrane fluidity (7, 10, 18). While increases in cholesterol reduce membrane fluidity, decreases in membrane cholesterol levels lead to the opposite as shown in liposomes as well as in bloodborne cells (7, 10). Importantly, membrane fluidity, as it relates to the mobility (e.g., rotational and lateral diffusion) of membrane proteins within the lipid bilayer [e.g., concanavalin A receptors on lymphocytes (1)], plays a key role in regulating cell functions. In addition, membrane fluidity also influences changes in protein tertiary structure that govern the activity of membrane-bound receptors, e.g., conformation-dependent activation of GPCRs on endothelial cells (8). The involvement of conformational changes of putative mechanosensors [i.e., GPCRs and CD18 (29, 48)] in PMNL shear-induced pseudopod retraction leaves open the possibility that membrane cholesterol alters PMNL mechanotransduction physically by affecting membrane fluidity which in turn influences structural perturbations related to the activation of membrane-bound, mechanosensitive proteins.

Along this line, we hypothesized that the degree to which PMNLs transduce fluid shear stress into a biological response (i.e., mechanosensitivity) depends on the cholesterol-dependent physicochemical properties of cell membranes. To test this hypothesis, we modulated the membrane cholesterol and fluidity of human PMNLs using cholesterol-enhancing (cholesterol:MβCD conjugates) or cholesterol-disrupting (MβCD and filipin) agents as well as a membrane-fluidizer (benzyl alcohol or BnOH) (6, 7, 12, 32) and examined the PMNL pseudopod retraction in response to well-defined fluid shear stresses. Since hypercholesterolemia is associated with enriched membrane cholesterol in bloodborne cells and dysregulated leukocyte activity (11, 27), we assessed the shear-responses by PMNLs from mice during the development of hypercholesterolemia as an in vivo model. In this regard, we explored a putative link between extracellular cholesterol levels, the cell membrane, and the shear stress mechanosensitivity of PMNLs.


Preparations of human PMNLs.

Procedures used for collecting blood samples from asymptomatic human volunteers were approved by the Institutional Review Board at either the University of Kentucky (Lexington, KY) or the University of California, San Diego (La Jolla, CA).

For some experiments, fresh blood was harvested by subcubital vein finger prick as described previously (34). After erythrocyte sedimentation at room temperature and 1 g for 45 min, the leukocyte-enriched plasma fraction containing neutrophils, monocytes, platelets, and sporadic erythrocytes was diluted 1:20 (vol/vol) with Plasma-Lyte (Baxter Healthcare) buffer containing 2.5 mM CaCl2 (Sigma-Aldrich) (13). To mimic the in vivo interactions between different types of blood cells, particularly for analyzing the shear-responses of nonadherent PMNLs, whole blood drawn into vacutainers (with anticoagulant K2-EDTA; BD) using standard venipuncture was diluted 1:20 (vol/vol) with Hanks' buffered saline solution (HBSS) (48).

Modification of leukocyte membranes.

Migrating leukocytes were allowed to adhere to plasma proteins adsorbed on glass substrates (Fisher Scientific) in Plasma-Lyte buffer (containing 2.5 mM CaCl2) for 10 min and subsequently incubated with either 2 mM MβCD (Sigma-Aldrich), 5 μg/ml filipin (Sigma-Aldrich), or 2 μg/ml cholesterol:MβCD complexes (Sigma-Aldrich) for 15 min. After these treatments, migrating cells were washed with excess Plasma-Lyte containing 2.5 mM CaCl2 and exposed to shear stress in the presence of 0–15 mM BnOH (Acros Organics).

For experiments with nonadherent PMNLs, whole blood was treated with 0–10 μg/ml cholesterol:MβCD complexes for 15 min and subsequently allowed to incubate in a 1:20 (vol/vol) dilution with HBSS containing BnOH and FMLP (Sigma-Aldrich) for 10 min before shear exposure; the final concentration of BnOH ranged from 0 mM to 7 mM while that of FMLP was 10 nM, a concentration that elevated leukocyte activity without influencing the shear-response (15).

PMNL viability remained unchanged at >90% after all membrane-modifying treatments tested in the present study.

Membrane cholesterol quantification.

Membrane lipids were extracted from human leukocytes after membrane-modifying treatments with a 1:2:1.8 (vol/vol/vol) chloroform:methanol:water solution according to Bligh and Dyer (2). The amount of unesterified cholesterol in these membrane extracts was quantified spectrophotometrically (at 600 nm) using a Free Cholesterol E Kit (Wako) in conjunction with a spectrophotometer (BioTek; μQuant).

Membrane fluidity measurements.

The fluidity of leukocyte membranes was assessed using 1-pyrenedecanoic acid (PDA; Molecular Probes) as described previously (5). Briefly, human leukocytes were incubated with 5 μM PDA at 37°C for 1 h, and then with 0–10 μg/ml cholesterol:MβCD complexes for 15 min. Fluorescence emissions were evaluated using a fluorescence spectrophotometer (Hitachi; F-2500) at 375 nm for monomer (Im) and at 470 nm for excimer (Ie) under 344 nm excitation (5). Membrane fluidity was expressed as a ratio of Ie/Im; higher Ie/Im ratios indicate higher membrane fluidity.

Micropipette shear stress exposure for adherent leukocytes.

Migrating PMNLs with extended pseudopod(s) were exposed to shear stress (≈2 dyn/cm2) for 2 min using a micropipette shear device as described previously (15, 34). Controls were migrating PMNLs maintained under static (no-flow), but otherwise similar experimental conditions. During experiments, video recordings of PMNLs were acquired using a videocassette recorder interfaced to a color charge-coupled device camera (Sony). Images from these recordings were digitized and processed with ImageJ (National Institutes of Health, Bethesda, MD). To quantify the pseudopod activity (i.e., degree of extension or retraction), the instantaneous length of the major axis of each cell tested was measured and normalized to either its average major axis length over the 2-min time period before shear exposure (for 6-min shear experiments) or its major axis length at 5 s before the start of the experiment (for 2-min experiments).

Cone-plate shear exposure for suspended leukocytes.

For nonadherent cells, aliquots (1 ml) of cell suspensions prestimulated with 10 nM FMLP were exposed to a constant shear stress field at 5 dyn/cm2 in a cone-plate device for 10 min. Controls were parallel suspensions of cells maintained under no-flow, but otherwise similar experimental conditions. At the end of experiments, aliquots (500 μl) of samples were immediately fixed with an equal volume of 2% paraformaldehyde and incubated for 1 h. The erythrocytes were then lysed using red blood cell lysing buffer (Sigma-Aldrich). Finally, leukocytes were labeled with DAPI (final concentration: 2 μg/ml) and observed with a microscope at a ×100 magnification and ×10 eyepiece. The fraction of activated PMNLs with pseudopod(s) out of at least 20 cells was visually assessed. From this, we calculated the PMNL activity level for blood samples according to the following equation: PMNL Activity Level=Number of Activated PMNLsTotal Number of PMNLs Analyzed×100% (1)Similarly, the shear-induced pseudopod retraction response index (PRR) was estimated with the following formula: PRR=[PMNL Activity Level]control[PMNL Activity Level]sheared[PMNL Activity Level]control×100% (2)

Animal studies.

All animal handling procedures used for the present study were approved by the University of Kentucky Institutional Animal Care and Use Committee. LDLr−/− mice (male; B6.129S7-Ldlrtm1Her; catalog no. 2207) were purchased from the Jackson Laboratory and subsequently acclimated for 1 wk before use. To induce hypercholesterolemia, mice (starting at 8 wk of age) were fed a diet enriched in saturated fat (HFD) (21% wt/wt fat and 0.15% wt/wt cholesterol; catalog no. TD88137, Harlan Teklad) ad libitum for up to 8 wk; age-matched LDLr−/− mice fed a normal mouse diet (ND) served as control animals. At selected intervals (i.e., 2, 4, and 8 wk of diet), mice were anesthetized with a lethal dose of xylazine-ketamine mixture (90 and 10 mg/kg body wt ip) followed by thoracotomy. Blood (500 μl) was drawn by cardiac puncture into a syringe containing EDTA (20 μl of 0.2 M). Within 3 h of blood collection, aliquots of blood were diluted 1:20 (vol/vol) with HBSS and subjected to cone-plate shear (5 dyn/cm2; 10 min) to assess the shear-induced PRR index of PMNL populations. The remaining (unused) blood was centrifuged at 1,000 g and 4°C for 5 min, after which the supernatant plasma was harvested. Finally, concentrations of unesterified and total cholesterol in plasma samples were determined spectrophotometrically using Free Cholesterol E (Wako) and Infinity Cholesterol Reagent (Sigma-Aldrich) kits, respectively. The mass concentration of esterified cholesterol was calculated as that of the sterol moiety plus fatty acid using the established relationship (16), Esterified Cholesterol=(Total CholesterolUnesterified Cholesterol)×1.67 (3)


Parametric data are expressed as means ± SE. Comparisons between the means of experimental treatments were conducted using Student's t-test with Bonferroni's corrected P values. Regression analyses were used to assess correlations between PMNL shear-responses and key parameters of interest, namely, PMNL membrane fluidity and plasma cholesterol concentrations in LDLr−/− mice.


Membrane cholesterol-modifying agents alter shear-responses of migrating PMNLs.

Membrane-modifying chemicals selectively altered unesterified cholesterol content in leukocyte membranes (Fig. 1A). Whereas treatment of leukocytes with 2 μg/ml cholesterol:MβCD complexes significantly (P < 0.008) increased membrane cholesterol content, incubation of cells with 2 mM MβCD significantly (P < 0.008) reduced cholesterol levels. Incubating leukocytes with filipin had no effect on membrane cholesterol levels. Interestingly, all of these membrane treatments influenced responses of adherent PMNLs to fluid shear stress.

Fig. 1.

Membrane cholesterol-modifying agents alter shear-responses of migrating polymorphonuclear leukocytes (PMNLs). A: normalized cholesterol concentrations in membrane extracts harvested from untreated PMNLs (UT) and cells that had been incubated with either 2 mM methyl-β-cyclodextrin (MβCD), 5 μg/ml filipin (FIL), or 2 μg/ml cholesterol:MβCD (CH). n ≥ 7. *P < 0.008 compared with UT. B–E: pseudopod activity for UT cells (B) and PMNLs pretreated with either cholesterol:MβCD (CH; C), MβCD (D), or filipin (FIL; E) is shown at 15-s intervals for 2 min before [0 ≤ time (t) < 2 min], during (2 ≤ t ≤ 4 min, solid horizontal bar on time axis), and after (4 < t ≤ 6 min) exposure to 2 dyn/cm2 shear stress. Images in B–E show PMNL morphology at 2 min before (t = 0 min) and at flow onset (t = 2 min) as well as at t = 4 min and 2 min after (t = 6 min) flow cessation. Cells were outlined (indicated by dotted line) to determine the instantaneous major axis lengths for cells at each time point (Lt). F: pseudopod activity of untreated PMNLs maintained under no-flow conditions for 6 min; these data served as controls. n = 3 (≥9 PMNLs analyzed for each sample). #P < 0.05 compared with corresponding time points for control (no-flow) cells.

Typically, under no-flow conditions, naïve PMNLs migrating on glass substrates extended and retracted pseudopods in a cyclical fashion [0 ≤ time (t) < 2 min; Fig. 1B]. Upon exposure to fluid shear flow, these cells retracted existing pseudopods, exhibited significantly (P < 0.05) reduced cell lengths at t = 3.5 and 3.75 min during the flow period (2 ≤ t ≤ 4 min; Fig. 1B), and adopted a compact/rounded morphology at flow cessation (Fig. 1B) compared with those observed at the corresponding time points during control experiments, i.e., cells maintained under no-flow conditions for 6 min (Fig. 1F). After removal of shear stress (t > 4 min; Fig. 1B), PMNLs resumed cyclical extension and retraction of pseudopods; i.e., the shear-response is reversible.

Migrating PMNLs pretreated with 2 μg/ml membrane cholesterol-enhancing cholesterol:MβCD conjugates exhibited a different pattern of pseudopod activity when subjected to a 6-min no flow/flow/no flow experimental regimen (Fig. 1C). In this case, PMNLs under an initial 2-min time period of no flow, on average, projected and retracted pseudopods cyclically as indicated by fluctuations in cell lengths (0 ≤ t < 2 min; Fig. 1C). During shear exposure, these cells appeared to exhibit an initial period of retraction within ∼1 min of flow onset followed by pseudopod extension (2 ≤ t ≤ 4 min; Fig. 1C). Once the shear application was removed, these PMNLs tended to retract pseudopods and adopt a compact morphology (t > 4 min; Fig. 1C). Observed changes in the pattern of pseudopod activity by membrane cholesterol-enhanced PMNLs under fluid flow, however, were not associated with any significant differences in length indices (Fig. 1C) when compared with untreated cells maintained under no-flow conditions for the entire 6-min duration of experiments (Fig. 1F).

Finally, incubation with either MβCD (Fig. 1D) or filipin (Fig. 1E) completely abolished pseudopod retraction by PMNLs in response to fluid shear exposure. Specifically, these cells exhibited cyclical extension and retraction of pseudopods characterized by fluctuating cell length indices that were independent of shear stress exposure.

Benzyl alcohol counteracts the blocking effects of membrane cholesterol enrichment on PMNL shear-responses.

Leukocytes incubated with cholesterol:MβCD conjugates followed by BnOH exhibited significantly (P < 0.02) elevated cholesterol levels in their membranes compared with untreated cells. Observed increases in membrane cholesterol content were solely attributed to treatment of cells with cholesterol:MβCD since incubation of PMNLs with BnOH alone had no effect on membrane cholesterol level (data not shown).

In terms of the effects of these chemical agents on the PMNL responses to flow, whereas naïve PMNLs exhibited pseudopod retraction with significant (P < 0.05) reductions in length indices after 1 and 2 min of shear exposure (relative to their length indices at flow onset), cells in the presence of 2 mM BnOH exhibited similar cell lengths throughout the 2-min shear exposure (Fig. 2A). Similar to trends from our prior experiments (Fig. 1C), preincubation with 2 μg/ml cholesterol:MβCD conjugates altered the responses of PMNLs to shear stress. Again, pseudopod activity of cholesterol-treated PMNLs under flow stimulation was characterized by an initial retraction for 1 min followed by extension of new pseudopods. Moreover, the initial period of pseudopod retraction was associated with significant (P < 0.05) reductions in cell length indices relative to those at the onset of flow (Fig. 2B).

Fig. 2.

Membrane fluidizers selectively recover impaired shear-responses of migrating PMNLs after membrane cholesterol enrichment. A and B: pseudopod activity of adherent PMNLs exposed to 2 dyn/cm2 shear stress in the absence (UT; ●) or presence of 2 mM BnOH (▵), 2 μg/ml cholesterol:MβCD (CH; □), 2 μg/ml CH + 2 mM BnOH (X), 2 μg/ml CH + 7 mM BnOH (◊) or 2 μg/ml CH + 15 mM BnOH (○). n = 3 (≥9 PMNLs analyzed for each sample). *,§,+,#P < 0.05 compared with the major axis lengths at the flow onset under each condition.

Notably, PMNLs pretreated with 2 μg/ml cholesterol conjugates followed by shear exposure in the presence of 2 mM BnOH retracted existing pseudopods without new pseudopod formation (Fig. 2B), a response similar to that of untreated cells. Interestingly, although cholesterol-treated PMNLs exposed to flow in the presence of 7 mM BnOH exhibited significantly (P < 0.05) reduced cell lengths after 1 min, their cell lengths increased thereafter (Fig. 2B). Moreover, when the BnOH concentration in cell suspensions was increased to 15 mM, pseudopod retraction by cholesterol-treated PMNLs was completely blocked throughout the 2-min duration of shear exposure (Fig. 2B).

Involvement of membrane cholesterol and fluidity in the PMNL shear-response is independent of cell adhesion.

In line with the observations on individual adherent cells, populations of nonadherent PMNLs prestimulated with 10 nM FMLP were deactivated by cone-plate shear stress (5 dyn/cm2, 10 min); this shear-response was significantly (P < 0.001) attenuated by 2 μg/ml cholesterol:MβCD (Fig. 3A). Interestingly, whereas 1 mM BnOH alone significantly (P < 0.001) reduced the PMNL shear-response, addition of 2 mM BnOH to suspensions of cells pretreated with 2 μg/ml cholesterol complexes restored their shear-responses to levels exhibited by untreated cells. BnOH at concentrations either lower or higher than 2 mM (i.e., 1, 3, 5, and 7 mM), however, failed to counteract the attenuating effects of 2 μg/ml cholesterol complexes on PMNL responses to shear. Moreover, when PMNLs were pretreated with 1 μg/ml cholesterol:MβCD, the concentration of BnOH needed to reinstate their shear-responses was shifted to 0.5 mM (Fig. 3B). Again, BnOH at concentrations less or greater than 0.5 mM (i.e., 0.2 and 1 mM) failed to completely recover the PMNL shear-response depressed by pretreatment with 1 μg/ml cholesterol:MβCD.

Fig. 3.

Membrane fluidizers selectively counteract the attenuating effects of membrane cholesterol enrichment on shear-induced deactivation of nonadherent PMNLs. A and B: pseudopod retraction responses (PRR; %) of PMNLs were determined after sequential incubations with cholesterol:MβCD (CH) [2 μg/ml (A) or 1 μg/ml (B)] followed by 0–7 mM BnOH and 10 nM FMLP before exposure to 5 dyn/cm2 shear stress for 10 min. n ≥ 3 (≥30 PMNLs analyzed for each sample). #P < 0.001, §P < 0.005 compared with samples without CH and BnOH treatments; *P < 0.001, ‡P < 0.005 compared with samples pretreated with CH alone.

PMNL responses to shear stress depend on the cholesterol-related fluidity of cell membranes.

The attenuating effects of membrane cholesterol enrichment on PMNL shear-responses (i.e., shear-induced deactivation as indicated by reduced pseudopod activity) depended on the concentration of cholesterol:MβCD used to preincubate PMNLs (Fig. 4A). Increasing the concentration of cholesterol complexes reduced PMNL responses to shear dose-dependently. When incubated with 10 μg/ml cholesterol:MβCD, the shear-responses of PMNLs were reversed. Concomitantly, incubation with increasing cholesterol:MβCD decreased the fluidity of plasma membranes in a similar dose-dependent fashion (Fig. 4B). Particularly, when cholesterol complexes were greater than 1 μg/ml, membrane fluidity was significantly (P < 0.001) lower than that of naïve cells.

Fig. 4.

Dose-dependent impairment of PMNL shear-responses by membrane-modifying cholesterol:MβCD conjugates correlates with their effects on cell membrane fluidity. A: PMNL shear-responses (PRR; %) were determined for cells pretreated with 0–10 μg/ml cholesterol:MβCD (CH) and 10 nM FMLP followed by exposure to 5 dyn/cm2 shear stress for 10 min. n = 4 (≥30 PMNLs analyzed for each sample). *P < 0.002 compared with samples without CH treatment. B: membrane fluidity for cells incubated with 0–10 μg/ml cholesterol:MβCD was measured and normalized to that of untreated cells. n ≥ 3. #P < 0.001 compared with samples without CH treatment. C: linear regression analysis for PMNL shear-responses and membrane fluidity. PRR indices for cells pretreated with 1 and 2 μg/ml CH (○) were included from prior experiments (see Fig. 3).

To confirm the existence of a link between PMNL shear-responses and membrane fluidity, a regression analysis was conducted between the membrane fluidity measurements and PRR values obtained from all experiments with cholesterol-treated cells. For PMNL populations pretreated with 0–10 μg/ml cholesterol:MβCD, shear-response index PRR highly correlated (R2 = 0.96; P < 0.0001) with membrane fluidity in a linear fashion (Fig. 4C).

Hypercholesterolemia alters shear-induced PMNL deactivation.

LDLr−/− mice placed on HFD, compared with their ND counterparts, exhibited significantly (P < 0.05) elevated blood plasma concentrations of total cholesterol (Fig. 5A) that were due to increases in both unesterified (Fig. 5B) and esterified (Fig. 5C) forms after 2, 4, and 8 wk of the HFD diet. Notably, after only 2 wk, the HFD group exhibited approximately fourfold higher serum unesterified and esterified cholesterol concentrations than those in the ND group. Moreover, as HFD progressed for up to 8 wk, plasma concentrations of total, esterified, and unesterified cholesterol rose in a time-dependent fashion. In blood from ND animals, concentrations of unesterified cholesterol remained constant throughout the 8-wk duration of the study despite time-dependent increases in esterified and total cholesterol levels.

Fig. 5.

High-fat diet (HFD) differentially affects levels of cholesterol components in plasma of low-density lipoprotein receptor-deficient (LDLr−/−) mice. A–C: concentrations of total (A), unesterified (B), or esterified (C) cholesterol in serum of LDLr−/− mice fed either a normal (ND; □) or a fat-supplemented diet (HFD; ○) for 2, 4, and 8 wk. n ≥ 4. £,‡,€P < 0.05 compared with the total, unesterified, or esterified cholesterol concentrations of animals on ND at each diet duration, respectively; #,¶P < 0.02 compared with total or esterified cholesterol concentrations of 2-wk ND animals, respectively; *,§,+P < 0.02 compared with the total, unesterified, or esterified cholesterol levels of 2-wk HFD animals, respectively.

Under static conditions, similar levels of PMNL activity were observed in the whole blood harvested from ND mice throughout the 8-wk course of our study (Fig. 6A). Although there was a gradual rise in the numbers of activated PMNLs in blood of ND mice after 4 and 8 wk, these were not significantly different compared with numbers of activated PMNLs in blood of similar animals after 2 wk of ND. On the other hand, whereas HFD mice had PMNL activity levels similar to those of their ND counterparts after 2 wk of high saturated fat feeding, their activated PMNL counts were significantly (P < 0.02) elevated (by over 2-fold) after 4 wk (Fig. 6A). Interestingly, after 8 wk, PMNL activity of HFD mice decreased to the levels that were similar to the levels of 2-wk HFD mice but was significantly (P < 0.05) lower than those of 8-wk ND mice.

Fig. 6.

PMNLs from LDLr−/− mice on HFD exhibit altered shear-responses. A: PMNL activity levels (%) were assessed for blood drawn from LDLr−/− mice fed either a normal (ND; □) or a saturated-fat diet (HFD; ○) for 2, 4, and 8 wk. #P < 0.05 compared with PMNL activity levels of ND mice at each diet duration; *P < 0.02 compared with PMNL activity levels of 2-wk HFD mice. B: shear-induced pseudopod retraction responses (PRR; %) were determined for PMNLs of LDLr−/− mice fed either a normal (ND; ■) or a high-fat diet (HFD; □) for up to 8 wk after exposure to 5 dyn/cm2 for 10 min. ‡P < 0.05 compared with the shear-responses of PMNLs from animals on ND at each diet duration; +P < 0.02 compared with the shear-responses of PMNLs from 2-wk HFD mice. n ≥ 4 (≥20 PMNLs analyzed for each sample).

In terms of mechanosensitivity to fluid flow, PMNLs from LDLr−/− mice under ND conditions for all diet durations tested (i.e., 2, 4, and 8 wk) were deactivated by shear stress exposure (Fig. 6B). In the case of HFD animals, the PMNL shear-response was altered (Fig. 6B). Specifically, PMNLs from blood of LDLr−/− mice on HFD for 2 wk exhibited significantly (P < 0.05) reduced shear-response indices compared with cells from animals on ND at this interval. After 4 wk, the PMNL shear-response by LDLr−/− HFD mice was completely abolished; in fact, a large number of PMNLs from three out of eight HFD animals were activated by shear exposure. Finally, after another 4 wk of HFD (i.e., 8-wk HFD animals), PMNL shear-responses were completely reversed as reflected by negative PRR indices. In this case, fluid shear stress appeared to stimulate PMNLs to project pseudopods relative to unsheared cells from the same animal.

We also examined our measurements for any significant correlation between the PMNL shear-response and serum cholesterol concentrations without regard for diet type or duration (Fig. 7). Overall, serum concentrations of cholesterol (either the total, unesterified, or esterified forms) significantly (P < 0.01) correlated with the PMNL shear-response index for first (linear)-, second (quadratic)-, and third (cubic)-order regression analyses when combining data from both ND and HFD animals (Table 1). Specifically, as serum cholesterol concentrations increased beyond some threshold level, shear-induced PMNL deactivation decreased and eventually reversed, i.e., cells projected pseudopods (Fig. 7). We repeated these regression analyses within each diet group (either ND or HFD) (Table 1) as well. Interestingly, within the ND group, the shear-response index was independent of serum concentrations of any forms of cholesterol regardless of whether regression fitting was conducted with linear, quadratic, or cubic relationships. In contrast, shear-responses of PMNLs from the HFD animals exhibited significant (R2 = 0.5; P < 0.01) correlations only with blood concentrations of unesterified cholesterol.

Fig. 7.

PMNL shear-responses change as plasma cholesterol levels increase in LDLr−/− mice. A–C: PMNL shear-response index (PRR; %) was plotted versus the concentrations of total (A), unesterified (B), or esterified (C) cholesterol in the plasma of LDLr−/− mice fed either a normal diet (ND; ×) or a high-fat diet (HFD; ●) for up to 8 wk.

View this table:
Table 1.

Regression analysis of PMNL shear-response and serum cholesterol concentrations


The present study brought to light a novel relationship between membrane cholesterol, membrane fluidity, and shear-induced PMNL deactivation as reflected by reductions in pseudopod activity. Our results provide the first evidence that the cell membrane plays a central role in PMNL mechanotransduction particularly as it relates to cellular mechanosensitivity. Specifically, changes in membrane cholesterol content, such as those due to elevated cholesterol abundance in local microenvironments, modified the ability of a PMNL to sense and functionally respond to local shear stresses generated by the surrounding flow field.

Cholesterol exists in either an unesterified or an esterified form. Of interest to this study was the unesterified cholesterol abundance within PMNLs since 1) the majority (>90%) of cellular unesterified cholesterol is contained within plasma membranes of eukarytotic cells (24, 25) and 2) changes in membrane cholesterol levels influence a number of cell processes via their effects on either lipid bilayer organization or fluidity. For instance, enhanced abundance of unesterified cholesterol in cell membranes influences protein translocation and calcium influx during GPCR-mediated cell activation (21). In this case, the effects of unesterified cholesterol uptake occurred due to changes in the composition, structure, and function of lipid rafts. In PMNLs, lipid rafts regulate a number of functions such as chemokine-induced calcium signaling, extracellular regulated kinase activity, cell polarization, shape change, adhesion, migration, integrin expression, and actin polymerization (32, 37, 39, 47, 53).

An interesting finding of the present study relates to our observations that, compared with the consistent retraction of pseudopods by naïve PMNLs under shear flow (Fig. 1B) that is in agreement with previous findings (15, 28, 31, 34), either reducing or sequestering cholesterol in PMNL plasma membranes abolished fluid shear-induced pseudopod retraction (Fig. 1, D and E). In fact, membrane cholesterol depletion/sequestration blocks a number of cell functions (e.g., migration and adhesion) based on their ability to dissociate membrane proteins from lipid rafts with impact on downstream signaling cascades (19, 38, 49). It is important to point out that removal of such small amounts of cholesterol from PMNL membranes, i.e., only 10% by 2 mM MβCD, which is in line with reported 21% reductions in membrane cholesterol of neutrophils treated with 10 mM MβCD for 15 min (39), completely blocked shear-induced pseudopod retraction response, suggesting the critical importance of membrane cholesterol content in PMNL mechanotransduction. Similarly, enhancing membrane cholesterol (Fig. 1C), which also influences lipid raft organization (36), impaired pseudopod retraction of PMNLs under fluid shear stress. These findings point to the existence of one membrane cholesterol level permissive for an unimpaired fluid flow-induced PMNL deactivation and, in general, implicate membrane cholesterol as an important modulator of cellular mechanotransduction.

After membrane cholesterol enrichment, however, PMNLs tended to retract pseudopods during an initial 1-min period of shear exposure followed by rapid pseudopod extension for the remaining flow period (Fig. 1C), a response that was confirmed in later 2-min shear studies (Fig. 2B). Notably, this unique biphasic response was in contrast to the fluid flow-insensitive and cyclical extension and retraction of pseudopods by sheared PMNLs that had been pretreated with cholesterol-reducing (MβCD) or -chelating (filipin) agents (Fig. 1, D and E). These data substantiate a potential difference in the dependence of PMNL shear-responses on membrane cholesterol perturbations resulting from enriching compared with reducing or sequestering membrane cholesterol.

To address this discrepancy, we investigated an alternative explanation as to why modulating membrane cholesterol influences the PMNL sensitivity to shear stress. Along this line, we focused on the possibility that shear-induced PMNL deactivation depends, at least in part, on the physical properties, i.e., fluidity or the inverse of microviscosity, of cell membranes. In essence, membrane fluidity alters the dynamics of membrane protein activity, either by influencing macromolecular interactions or by modulating conformational changes, thereby affecting downstream cell signaling and functions (26). For instance, membrane fluidity influences the number, affinity, lateral mobility, as well as conformational changes of membrane receptors (e.g., FPR and concanavalin A receptor), thus regulating their sensitivity to corresponding ligands and associated downstream functions such as cell migration, phagocytosis, cell growth, and differentiation (1, 8, 52, 55).

Interestingly, BnOH, an established membrane fluidizer for a variety of cell types, including erythrocytes, lymphocytes, endothelial cells, and epithelial cells (4, 6, 12, 46), impaired shear-induced pseudopod retraction in adherent and nonadherent PMNLs on its own (Figs. 2A and 3A). These data combined with our observations that membrane cholesterol enrichment (which rigidifies cell membranes) also blocked shear-induced pseudopod retraction suggest the existence of an optimal membrane fluidity level for PMNL mechanotransduction. What was most dramatic, however, was our observation that BnOH, in a concentration-selective manner, recovered the shear-responses of adherent and nonadherent PMNLs that had been blocked by membrane cholesterol enrichment (Figs. 2B, 3A, and 3B). Only one concentration of BnOH completely recovered the shear-responses of cholesterol-pretreated PMNLs, and the concentration of BnOH required to counterbalance the blocking effects of cholesterol enrichment on PMNL shear-responses depended on the concentration of cholesterol conjugates used to initially treat the cells. In fact, incubation of PMNLs with cholesterol-enhancing agents reduced the fluidity of cell membranes as reported (7) in a dose-dependent manner (Fig. 4B), which was accompanied by diminished and even reversed cell responses to shear (Fig. 4A). Notably, there was a strong correlation (R2 = 0.96; P < 0.0001) between PMNL shear-responses and membrane fluidity related to cholesterol enrichment (Fig. 4C). Together, these data provide the first evidence that lipid bilayer fluidity is not only critical for PMNL mechanotransduction, but that there exists an optimal level of membrane fluidity permissive for an unimpaired PMNL shear-response.

The in vitro results of the present study, therefore, strongly point to the putative role of the cell membrane as a mechanotransducing compartment for PMNLs. This is reasonable due to its strategic location between the intra- and extracellular milieu as well as its enriched content of putative mechanosensors, e.g., FPR and CD18 integrins (29, 48). There is, however, a possibility that the cell membrane itself acts as a mechanotransducer. In this respect, literature reports demonstrate increased membrane fluidity of human (20) and bovine endothelial cells within 5 s of exposure to 10–26 dyn/cm2 shear stress, a response associated with increased mitogen-activated protein kinase activity and altered activity of G proteins and/or GPCRs (3, 4, 8). However, the concept that the membrane, on its own, serves as a mechanotransducer lacks the specificity that would explain the diversity of responses to shear stress exhibited by different cell types. It is more likely that membrane fluidity serves to modulate the sensitivity of cell-specific mechanoreceptors to local shear stress distributions rather than to initiate cell mechanotransduction.

The physiological implications of the present study center around substantial evidence (30) pointing to the leukocyte shear-response as a fluid mechanics-based control mechanism to ensure that these cells remain in a round, passive, nonadhesive state permissive for their unhindered passage through the small vessels of the microcirculation. Under nonphysiological or pathological conditions associated with elevated cell activity levels above a putative threshold, leukocytes lacking intact shear-responses project pseudopods and become less deformable with detrimental effects on their transit through the microcirculation, contributing to increases in overall microvascular hemodynamic resistance (14). In this regard, the lack of a shear stress response (i.e., pseudopod retraction) by leukocytes is a putative pathogenic mechanism underlying the development of chronic inflammation associated with sustained leukocyte activation in the microcirculation during hypertension and contributing to downstream organ and tissue injury (14).

In general, chronic inflammation is a common denominator for a wide range of cardiovascular disease conditions including diabetes and hypercholesterolemia, both of which involve development of a hyperlipidemic blood environment (45). As such, the possibility that elevated blood cholesterol alters leukocyte sensitivity to fluid flow through its effects on the cell membrane implicates aberrant or dysfunctional leukocyte mechanotransduction in this pathological condition. Along this line, the present study examined the pseudopod retraction responses by PMNLs from LDLr−/− mice. When placed on a diet supplemented with fat, these mice develop gradually increasing plasma cholesterol levels over a period of weeks compared with those on regular chow (17). Moreover, LDLr−/− mice as well as wild-type mice subjected to hypercholesterolemia are associated with chronic leukocyte activation in the microcirculation (35, 50).

Interestingly, LDLr−/− mice fed a HFD for as short as 2 wk exhibited greatly elevated (>4-fold) serum concentrations of total, unesterified, and esterified cholesterol (Fig. 5) that were accompanied by a reduced PMNL shear-response (Fig. 6B). It is not surprising that the aberrant shear-response occurred as early as after 2 wk of HFD. This result is consistent with the reported increases in leukocyte adhesion and emigration in the microcirculation of wild-type mice after 2 wk of a high-cholesterol diet (50). Moreover, attenuation of the shear-response was not due to a deficiency in LDLr since PMNLs from LDLr−/− mice on ND exhibited an intact response to shear (Fig. 6B). In fact, without regard for diet type or duration, we detected significant (P < 0.01) correlations between serum concentrations of total, unesterified, or esterified cholesterol and PMNL shear-responses (Table 1) pointing to a general dependence of PMNL shear-responses on serum cholesterol concentrations. However, when examining for similar correlations for ND and HFD animals analyzed separately, we identified concentration of unesterified cholesterol as the most important predictor of an impaired shear-response.

Taken together, the results of the present study provide evidence that the PMNL mechanosensitivity to fluid shear stress depends on plasma concentrations of unesterified cholesterol which can be transported into the cells through receptor dependent/independent pathways (40) and alter the membrane cholesterol and fluidity (11, 27). Beyond a threshold level, plasma rich in unesterifed cholesterol is associated with a reversed shear-response, probably due to enhanced accumulation of membrane cholesterol, which is consistent with the observed changes in human PMNL responsiveness to shear brought about by incubation with cholesterol-enhancing agents. Finally, in conjunction with our in vitro experiments using human blood, our results pointed to membrane fluidity as the critical factor that relates changes in the regulation of PMNL activity by fluid shear stress with 1) blood cholesterol levels and 2) membrane cholesterol levels.

It is important to point out that half of the variances associated with the shear-responses of PMNLs from LDLr−/− mice on HFD were attributed to changes in the blood levels of unesterified cholesterol (R2 = 0.5; P < 0.01), leaving open the possible contributions of other factors. For example, elevated plasma concentrations of circulating proinflammatory mediators [e.g., oxidized LDL (42)] associated with LDLr−/− mice on HFD may have raised PMNL activity to levels that suppress shear-induced deactivation (15). Contrary to this was the independence of shear-responses on the PMNL activity levels in LDLr−/− mice on ND or HFD for up to 8 wk (Fig. 6). It is also possible that altered regulation of superoxide and nitric oxide levels related to hypercholesterolemia (22) impaired PMNL responses to shear since both of these molecules are involved in the leukocyte shear-response and exhibit membrane solubility (15, 23). In addition, the reduced PMNL shear-responses may be attributable to an enhanced cleavage of membrane-bound receptors such as shear-sensitive FPRs (9). However, the fact that the effects of these factors (e.g., proinflammatory agonists, superoxide/nitric oxide, receptor cleavage) together with additional contributors, such as other HFD components and age, account for the remaining ≈50% of the variances associated with hypercholesterolemia-related attenuation of PMNL shear-responses further supports the important and dominant role of cholesterol.

In conclusion, the results of the present study provide evidence that membrane cholesterol influences mechanotransduction of shear stress by PMNLs, at least in part, through its impact on membrane fluidity. This has implications in pathologies associated with elevated cholesterol levels in the blood milieu of PMNLs. Along this line, the present study points to the possibility that defective fluid flow regulation of PMNLs contributes to the etiology of cardiovascular disease particularly as it relates to hyperlipidemia and chronic inflammation, e.g., atherosclerosis and amplified microvascular injury due to ischemia-reperfusion.


This work was supported by the National Institutes of Health (HL10881 and HL083740), the University of Kentucky Research Support Fund, the American Heart Association (Beginning Grant-in-Aid), and the National Science Foundation (Kentucky Experimental Program to Stimulate Competitive Research-Bioengineering Initiative).


No conflicts of interest, financial or otherwise, are declared by the author(s).


We acknowledge Mr. Darby Chan for technical assistance with micropipette experiments and Dr. Abhijit Patwardhan for use of statistical software.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
  47. 47.
  48. 48.
  49. 49.
  50. 50.
  51. 51.
  52. 52.
  53. 53.
  54. 54.
  55. 55.
View Abstract