Role of matrix metalloproteases in the kinetics of leukocyte-endothelial adhesion in post-capillary venules

Abstract

Background:

The endothelial glycocalyx serves as a barrier to leukocyte (WBC)-endothelium (EC) adhesion. Shedding of glycans, by matrix metalloproteases (MMPs) exposes EC integrin receptors to facilitate firm adhesion. However, the effect of shedding on the strength of the adhesive bond remains to be determined.

Objectives:

Examine the effect of MMP inhibition on the kinetics of WBC-EC adhesion under normal and inflammatory conditions to delineate differences in the duration and number of adhesive bonds.

Methods:

WBC adhesion in post-capillary venules was observed in rat mesentery. Adhesion duration and off-rates ( $K_{OFF}$ ) were correlated with shear stress during adhesion in response to 1 µM fMLP or 0.5 µM doxycycline (doxy, to inhibit MMP activation).

Results:

Doxy increased mean adhesion time significantly from 2.5 (control) to 5.6 s, whereas fMLP increased it 8-fold to 20 s, which was not affected by pre-treatment with doxy. Estimates of the number of adhesive bonds (simplified Bell-model) revealed a significantly greater increase with fMLP compared to doxy alone, with no effect on fMLP by pretreatment with doxy. With doxy alone, $K_{OFF}$ was significantly 4-fold greater compared to fMLP, suggesting a much weaker bond.

Conclusions:

Although the increased number of bonds by MMP inhibition with doxy alone and fMLP were similar, the bonds due to doxy appeared weaker as evidenced by their shorter duration, and lesser reduction in $K_{OFF}$ relative to control. Thus doxy limits the availability of integrin binding sites during fMLP stimulated adhesion, but has a pro-adhesive effect due to increased ligands for WBC binding that arises from inhibition of normal sheddase activity on the EC.

Keywords

Leukocyte adhesion matrix metalloproteases MMPs adhesion duration bond force

1. Introduction

The adhesion of leukocytes (WBCs) to the endothelium (EC) in post-capillary venules has been firmly established as dependent upon the affinity and avidity of integrin molecules on the WBC [1–4]. Alterations in ligand-binding affinity and modulation of avidity by integrin cell surface diffusion and clustering have been postulated as the basis for increased WBC-EC adhesion in response to cytokines and chemoattractants (e.g. TNF-α and fMLP). Intravital observations of the dynamics of WBC-EC adhesion in post-capillary venules (in mesentery) have revealed a rapid response to stimulation by chemoattractants resulting in as many as 12 firmly adherent WBCs per 100 µm length of venule within 30 s of application of 1 µM fMLP [5]. This amount of adhesion was sufficient to raise the resistance to blood flow 2-fold and thus brings to focus the deleterious effect of WBC-EC adhesion on microvascular blood flow during the inflammatory process. Subsequent studies of the endothelial response to fMLP revealed that superfusion of mesenteric tissue induced a shedding of glycans from the EC surface that resulted in increased exposure of the principal ligand for WBC adhesion, ICAM-1 [6], thus suggesting that the endothelial surface layer may serve as a labile barrier to WBC adhesion.

The endothelial surface layer (ESL) has been shown to consist of a layer of adsorbed plasma borne proteins that covers the EC glycocalyx [7]. The structure of the endothelial glycocalyx has been reviewed extensively [7–11] as well as its role in leukocyte adhesion [12,13]. The most prominent components of the glycocalyx are the glycosaminoglycans (GAGs) heparan sulfate (HS), chondroitin sulfate (CS), and hyaluronan (HA). The GAGs, HS, and CS are covalently linked to membrane-bound proteoglycans (PGs). The composition of the glycocalyx has been shown to reflect a balance between shear dependent removal of components and the biosynthesis of new glycans [13]. The endothelial glycocalyx has been shown to be shed in response to inflammation, hyperglycemia, endotoxemia and septic shock, the presence of oxidized LDL, TNFα, atrial natriuretic peptide, abnormal blood shear stress, ischemia–reperfusion injury, light-induced production of free radicals, and during by-pass surgery [14]. While this broad range of events encompasses several mechanisms of shedding, it has been suggested that proteolytic cleavage of proteoglycans may result from the convergence of multiple pathways that activate a cell surface metalloproteinase [15]. Direct in situ observations of shedding in post-capillary venules suggest that matrix metalloproteases, may be responsible for shedding of glycocalyx components [6]. Matrix metalloproteases on the surface of the venular endothelium are rapidly activated by superfusion of the mesenteric tissue with fMLP and can be inhibited by superfusion with sub-antimicrobial doses (0.5 µM) of the antibiotic doxycycline. The inhibitory activity of doxycycline on shedding stems from its direct effect on MMP activation and not by its ability to chelate divalent cations [16] or scavenge reactive oxygen species [17].

In the context of these observations, the current studies were undertaken to quantitate the kinetics of WBC-EC adhesion with the aim of determining the extent to which MMP activation affects the strength of their adhesive bond. To that end, WBC adhesion in post-capillary venules in intestinal mesentery of the rat was observed in response to superfusion of the tissue with fMLP, in the presence and absence of MMP inhibitors. The effect of MMP inhibition was examined in terms of the duration of adhesion for a given wall shear stress to quantitate the affinity of the adhesive bond, and the force generated by shear stresses on the WBC and the number of bonds were calculated using established models of cell adhesion.

2. Methods

2.1. Animal preparation

All animal studies conformed to the Guiding Principles in the Care and Use of Animals established by the American Physiological Society, and all protocols have been approved by the Institutional Animal Care and Use Committee of The Pennsylvania State University.

Male Wistar rats, weighing 250–400 g, were anesthetized with Inactin (120 mg/kg, i.p.), tracheostomized, and allowed to breathe under spontaneous respiration. The right jugular vein and its paired carotid artery were cannulated with polyethylene tubing (PE-50, Becton Dickinson, Franklin Lakes, NJ, USA). Supplemental anesthetic was administered via the jugular catheter, as needed, to maintain a surgical plane of anesthesia. The carotid catheter was connected to a strain-gage pressure transducer to monitor central arterial pressure, which averaged a nominal 125 mmHg. Core temperature was monitored by a rectal probe and was maintained between 36 and 37°C with the aid of a heating pad.

2.2. Intravital microscopy

The intestinal mesentery was exteriorized through a midline abdominal incision and placed on a glass pedestal to permit viewing under brightfield microscopy by trans-illumination using a 40x/0.75 NA water immersion objective (Zeiss, Thornwood, NY, USA) under tungsten illumination. The tissue was superfused with 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid buffered Ringer’s solution (pH = 7.4) at a temperature of 37.0°C. Visual recordings of post-capillary venules were digitized using a PCO 1600 digital CCD camera (PCO Imaging, Kelheim, Germany) at a spatial resolution of 1600 × 1200 pixels with a depth of 14 bits for subsequent analysis. The effective magnification yielded a pixel spacing of 17.4 pixels/µm and a width of the video field of 92 µm.

2.3. Hemodynamic measurements

Red cell velocity ( $V_{RBC}$ ) along the centerline of mesenteric venules was measured using the two-slit photometric technique [18]. The image of a microscopic field was switched to project onto a CCD camera (WV-51CCD; Panasonic, Secaucus, NJ, USA) for an effective width of the video field of 100 µm. $V_{RBC}$ along the centerline of post-capillary venules was measured with the two-slit photometric technique using a self-tracking correlator (IPM, San Diego, CA, USA). Mean flow velocity ( $V_{MEAN}$ ) was calculated from the empirical relationship $V_{MEAN} = V_{RBC} / 1.6$ [19]. Vessel diameter (D) was measured by the video image shearing technique using an image shearing monitor (IPM). Wall shear rates were estimated from the Newtonian-flow relationship for flow in a cylindrical tube, viz. $\dot{γ} = 8 V_{MEAN} / D$ . Wall shear stresses ( $τ_{WALL}$ ) were estimated from the Newtonian relationship, $τ_{WALL} = η \dot{γ}$ , where the apparent viscosity (η) was taken as 2.5 cP, which equaled the value corresponding to the average venular hematocrit and a linear relationship between microvessel hematocrit and η [20].

2.4. Experimental protocols

Following exteriorization of the mesentery, the tissue was allowed to stabilize for 20 min under superfusion with either Ringer’s solution (control conditions) or Ringers with the addition of either the broad spectrum MMP inhibitor doxycycline (0.5 µM), or the hydroxamic acid zinc chelator GM-6001 (2.6 µM, US Biological). To examine the effects of MMP inhibition on stimulated adhesion, following the stabilization period the tissue was superfused with 1 µM fMLP in Ringer’s solution. To establish a frame of reference indicative of non-specific adhesion, the tissue was superfused with 3 mM EDTA to inhibit selectin mediated rolling and adhesion.

Fig. 1.

Representative frequency distribution of WBC adhesion times for 300 adhesion events measured in a 30 µm diameter venule: (A) Raw adhesion times, (B) distribution normalized to area of 1.0, (C) cumulative probability (Pr) that an adhesion event will occur within time t, obtained by integrating the normalized frequency distribution, and (D) calculated survival time of adhesion taken as the probability that a cell will not adhere after time T, calculated at $1 - Pr$ .

For each of these treatments, frame by frame video analysis of the duration of WBC-EC adhesion events on the EC surface was made from digitized images. An adhesion event was defined as the arrest and firm adhesion of a WBC that lasted two or more successive video frames. From these measurements the frequency distribution of adhesion times within venules under various levels of shear stress were obtained to establish a correlation between mean adhesion time and shear stress. The off-rate of debonding of WBCs, and the average number of bonds formed between WBC and EC, were then calculated.

2.5. Statistical analysis

Statistical analyses of trends in the data were performed using SigmaStat (Systat, San Jose, CA) with either Student’s t-test for paired measurements, or the Holm–Sidak method for ANOVA of multiple comparisons.

3. Results

Representative measurements taken under control conditions (Ringer’s) are shown in Fig. 1(A) where the frequency distribution of adhesion times is given for 300 adhesion events in a single post-capillary venule with diameter of 30 µm and calculated wall shear stress of 7.6 dyn/cm². Shown are the first 2 s of a 40 s range, within which mean adhesion time averaged 1.6 ± 4.8 SD s. The smallest time resolved was 0.033 s (dictated by the image capture frame rate of 30 frames/s). To permit interpretation of these data as a probability density distribution (pdf) of adhesion times, the data were normalized to an area of 1.0 (Fig. 1(B)). Integration of this curve from 0 to time T thus yielded the cumulative probability of an adhesion event occurring, Pr in Fig. 1(C), i.e. $Pr (t ⩽ T) = \int_{0}^{T} pdf d t$ . Taking $1 - Pr$ thus yields the probability of a cell not adhering after time T that is the survival time of an adhesive bond shown in Fig. 1(D). This curve was found to be best fit by a double exponential function, $1 - P_{r} = a e^{- t / τ_{1}} + b e^{- t / τ_{2}}$ . The off-rate ( $K_{OFF}$ ) representative of the debonding event was taken as the inverse of the smaller of the two time constants.

To illustrate the effects of prevailing shear stresses, presented in Fig. 2 are the mean adhesion times and off-rates of 300 adhesion events in each of 17 venules, ranging in diameter from 18–59 µm for a range of $τ_{WALL}$ from 1.0 to 13.3 dyn/cm². Mean adhesion time declined and $K_{OFF}$ rose significantly with increasing shear stress. In light of the skewed distribution of adhesion times (Fig. 1) it should be noted that the mean value may not adequately reflect the central tendency of the data, and the median adhesion time may be more statistically relevant. However, the correlation of median adhesion time with wall shear stress was not significant. Red cell velocity ( $V_{RBC}$ ) averaged 2.0 ± 1.2 SD mm/s, ranged from 0.5 to 5 mm/s and was not significantly correlated with diameter ( $r^{2} = 0.003$ , $p = 0.849$ ) for this sampling of venules. Similarly, wall shear rate ( $\dot{γ}$ ) averaged 289 ± 162 SD s⁻¹, and was not significantly correlated with diameter ( $r^{2} = 0.119$ , $p = 0.173$ ). Nonetheless, mean adhesion time rose significantly with increasing diameter ( $r^{2} = 0.351$ , $p < 0.011$ ). In contrast, the off-rate was not correlated with diameter ( $r^{2} = 0.031$ , $p = 0.499$ ).

Fig. 2.

Mean adhesion times and off-rate for 300 adhesion events in each of 17 venules as a function of wall shear stress. A linear regression is shown for each, bracketed by the 95% confidence interval of the slope (dashed curves). Adhesion time fell significantly with shear stress ( $r^{2} = 0.280$ , $p < 0.027$ ) and off-rate rose significantly ( $r^{2} = 0.254$ , $p < 0.037$ ). The mean adhesion time averaged 1.73 ± 1.74 SD s, and off-rate averaged 4.92 ± 2.51 SD s⁻¹ with average shear stress in all venules of 7.2 ± 4.0 SD dyn/cm².

Fig. 3.

Representative frequency distributions of adhesion times for superfusion of mesentery with: (A) Control (Ringer’s solution), (B) $10^{- 7}$ M fMLP, (C) 0.5 µM doxycycline, and (D) $10^{- 7}$ M fMLP following 20 min exposure to 0.5 µM doxycycline. Statistics of hemodynamic parameters are given in Table 1. These distributions have been normalized to area 1.0 as in Fig. 1(B).

To contrast the effects of fMLP stimulated adhesion and the effects of MMP inhibition, the distribution of WBC adhesion time was obtained in separate experiments during superfusion of the mesentery with either Ringer’s solution alone, or with the addition of 1 µM fMLP, 0.5 µM doxycycline, or fMLP following a 20 min exposure to 0.5 µM doxycycline. Representative frequency distributions of the normalized frequency distributions of adhesion times are illustrated in Fig. 3. Statistics of the number of venules, adhesion events and hemodynamic parameters are given in Table 1. To facilitate comparisons of the treatments in Fig. 3, a lesser number of venules and adhesion events per venule were observed compared to those of Fig. 1. The range of adhesion times observed varied with the treatment, and in the case of fMLP applied after 20 min superfusion with doxy (fMLP + doxy) only 26 WBCs were adherent in the 10 venules studied. Further, due to the smaller sampling of adhesion events, adhesion time vs. $V_{RBC}$ or $τ_{WALL}$ was not significantly correlated. It should be noted that the rate of occurrence of adhesion events for each treatment was not significantly different, thus suggesting no effect of the margination process.

Table 1

Venule and hemodynamic parameters

Treatment	# Venules	# Adhesion events	Venule diameter (µm)	$V_{RBC}$ (mm/s)	Wall shear rate, $\dot{γ}$ (s⁻¹)	Wall shear stress, $τ_{WALL}$ (dyn/cm²)
Control	13	322	26.2 ± 5.4	2.5 ± 0.69	484 ± 151	12.1 ± 3.8
fMLP	16	104	25.6 ± 6.1	2.5 ± 0.65	502 ± 153	12.6 ± 3.8
Doxy	11	273	30.2 ± 5.9	2.9 ± 1.2	476 ± 164	11.9 ± 4.1
fMLP + doxy	10	26	29.4 ± 6.6	3.6 ± 1.6	679 ± 348	17.0 ± 8.7
EDTA	9	135	29.2 ± 6.1	2.4 ± 0.45	423 ± 86	10.6 ± 2.1
GM6001	15	165	31.0 ± 5.7	3.0 ± 1.0	466 ± 103	11.6 ± 2.6

Notes: Given above are mean ± SD.

Presented in Fig. 4 are survival curves ( $1 - Pr$ ) of adhesion times for the four treatments of Fig. 3 and during superfusion of the hydroxamic acid zinc chelator GM6001 that inhibits MMP activation. Both GM6001 and doxycycline extended the duration of adhesion, although doxycycline appeared slightly more effective. Topical application of fMLP dramatically increased the duration of adhesion, which was not affected by subsequent MMP inhibition with doxycycline. As a frame of reference, superfusion with 3 mM EDTA is also shown, which exhibited the anticipated reduction in WBC rolling and adhesion due to Ca²⁺ chelation.

Fig. 4.

Survival curves representing the probability $(1 - Pr)$ that a WBC will not adhere after a specific adhesion time, for the data of Fig. 3. For each treatment, all adhesion events have been pooled, regardless of shear rate. Also shown are survival curves for the zine chelator GM6001 that inhibits MMP activation, curves for superfusion with 3 mM EDTA, which limits Ca²⁺ dependent rolling and adhesion of WBCs.

The mean adhesion times and off rates ( $K_{OFF}$ ) for the survival curves of Fig. 4 are shown in Fig. 5. The mean adhesion time under control conditions averaged 2.5 s. For all treatments, mean adhesion times and off-rates were significantly different from control. As expected, EDTA significantly reduced adhesion time whereas fMLP significantly increased it. No significant effect of pretreating the tissue with doxy prior to fMLP (fMLP + doxy) was observed (ANOVA with multiple comparisons). Inhibition of MMP activity with either doxycycline or the zinc chelator GM6001 significantly increased mean adhesion time. This behavior is consistent with the hypothetical inhibition of constitutive sheddases that limit the availability of ligands for WBC rolling and adhesion on the EC. The off-rates were not significantly different for doxycycline compared to GM6001, nor were they different for fMLP with and without pretreatment with doxycycline. These results suggest that the strength of the adhesive bonds between WBC and EC in response to fMLP were not affected by MMP activity.

Fig. 5.

(A) Mean adhesion time and (B) $K_{OFF}$ for the survival curves of the pooled data shown in Fig. 4. Mean adhesion time and $K_{OFF}$ were significantly different from control for all treatments, $p < 0.05$ , ANOVA. Multiple pairwise comparison by ANOVA using the Holm–Sidak method showed no difference in adhesion time and $K_{OFF}$ for doxy vs. GM6001 and fMLP vs. doxy + fMLP; and doxy, EDTA and GM6001 were significantly different from fMLP.

To estimate the avidity (strength) of the adhesive bonds formed during each treatment, the number of bonds formed was calculated using the adhesion model of Bell [21]. As schematized in the inset of Fig. 6, the hydrodynamic drag force (F) acting on an adherent WBC was calculated for the estimated wall shear stress, $τ_{WALL}$ , from in vivo measurements [22]. In this study, F was found by conservation of momentum in venules of diameter (D) for a specific red cell velocity, and varied as $F / τ_{WALL} = 2.5 - 0.045 D$ , where $F [=] dyn$ , $τ_{WALL} [=] {dyn/cm}^{2}$ , and $D [=] µ m$ . The aggregate force acting on the cluster of bonds at the WBC trailing edge was calculated as $F / cos θ$ , where θ is the contact angle of the WBC membrane in the adhesion zone. The angle θ was measured from video recordings under control conditions and during fMLP and found to average 46.8 ± 9.5 SD degrees under control conditions (for 18 WBCs) and 60.5 ± 19.4 SD degrees, with fMLP (for 13 WBCs). Shear rates and stresses were similar to those found in the present study. The greater contact angle with fMLP is consistent with the decreased deformability of activated WBCs that become more spherical with polymerization of their internal actin filaments. Under control conditions, the average drag force on an adherent WBC was $16.4 \pm 7.2 SD \times 10^{- 5} dynes$ , which did not vary significantly with treatments.

Fig. 6.

Comparison of the number of bonds formed between WBC and EC during adhesion for each of the five treatments shown corresponding to the pooled data of Fig. 4. The number of bonds formed was calculated using the adhesion model suggested by Bell [21], for the drag force (F) acting on each WBC and the measured adhesion time, assuming that the WBC was tethered to the endothelium at its trailing edge, as schematized in the inset. The number of WBCs observed in each case is given in Table 1. ^∗ Significantly different from control, # significantly different, NS not significantly different.

The number of bonds formed was calculated from Bell’s equation derived from the kinetic theory of fracture [21], viz.: $\begin{matrix} τ = τ_{0} exp [(E_{0} - r_{0} F_{TOTAL} / N_{BONDS}) / k T], \end{matrix}$ where: τ = lifetime of the adhesive bond, τ ₀ = reciprocal of natural frequency bond oscillation, 10⁻¹³ s, E ₀ = bond energy, r ₀ = maximum stretch of a bond, 0.5 nm, F _TOTAL = total adhesive force, k = Boltzmann’s constant, $1.38 \times 10^{- 23} J/ ° K$ , T = absolute temperature, 310°K.

The bond energy, $E_{0}$ , represents the energy barrier that must be overcome to disrupt the bond, and was estimated from linear regressions of the natural log of adhesion time vs. $τ_{WALL}$ for control and fMLP experiments. Under control conditions, $E_{0}$ equaled $8.8 \times 10^{- 13} dyn-cm$ and with fMLP was $9.4 \times 10^{- 13} dyn-cm$ . The increase in bond energy with fMLP was significant, $p < 0.00002$ , t-test on intercepts. With this formulation, the number of bonds formed between WBC and EC were calculated for each adhesive event. The average number of bonds is presented for each treatment in Fig. 6, where it was assumed that the bond energy for EDTA and doxy were equal to that under control conditions, and that for fMLP + doxy was the same as fMLP.

With EDTA, the number of bonds fell significantly from 244 to 89, $p < 0.05$ (ANOVA with post-hoc comparison with control, Holm–Sidak method). The cases of doxy, fMLP and fMLP + doxy all increased significantly about 3-fold above control, $p < 0.05$ . All pairwise multiple comparisons revealed that doxy was significantly less than fMLP, but fMLP was not significantly different from fMLP + doxy.

4. Discussion

Saltation and rolling of leukocytes along the venular endothelium and their subsequent transient adhesion is an essential sequence in the inflammatory process that leads to transmigration into the parenchymal tissue [23]. It has been firmly established that rolling of leukocytes on the EC is facilitated by their adhesive interactions with selectins on the EC membrane, primarily E- and P-selectin [24], that result in their firm arrest by ligands for WBC integrin receptors, primarily ICAM-1 and, to a lesser extent, VCAM-1 [25]. Components of the EC glycocalyx have also been observed to support WBC adhesion. The most prominent component, heparan sulfate, HS, [7] has been shown to serve as a ligand for the WBC $β_{2}$ integrin CD11b/CD18 that promotes firm adhesion of neutrophils to HS coated surfaces [26].

Under control conditions, the trends in adhesion time (Tadh) were consistent with prior in vivo observations in omentum (rabbit) [27] where an average value of 2.5 ± 0.2 SE s was observed for fluorescently labeled unfractionated WBCs infused into the circulation (cf. $⟨ Tadh ⟩ = 1.73 \pm 1.74 SD s$ , Fig. 2(A)). In those studies, a double exponential fit of the survival curve was also noted, with an off-rate (corresponding to smallest time constant) equal to 10.0 ± 2.78 SE s⁻¹, which was comparable to the average $⟨ K_{OFF} ⟩ = 4.92 \pm 2.51 {SD s}^{- 1}$ found herein. Differences between the two sets of measurements may be attributed to species and experimental protocols. Prior studies of transient adhesion of lymphocytes to high endothelium venules in Peyer’s patches [28] revealed a best fit of the survival curve with a double exponential plus an asymptotic constant offset, and an off-rate equal to 3.0 s⁻¹ corresponding to the shorter of the two time constants. As shown therein, the mean adhesion time was determined to be about 0.3 s for about two-thirds of adhesion events, and about 3.4 s for 1/7th of the cell population. It was noted therein that the double exponential decay may reflect differences between binding properties of the B and T cells that comprised the lymphocyte population, as well as geometry of the interaction site.

In prior studies from this laboratory with Wistar rats, differential WBC counts averaged about 22% neutrophils and 77% lymphocytes in peripheral blood samples [29]. In spite of the predominance of lymphocytes in the circulation, analysis of the subtype of WBC engaged in rolling along the venular EC of mesentery revealed that greater than 94% were granulocytes [30]. However, analysis of WBC rolling in response to activation of the endothelium with cytokines (e.g. IL-1) suggested that the number of monocytic cells rolling may increase in neutrophil depleted animals [31]. Hence, mean adhesion times may be weighted toward receptor–ligand interactions typical of the neutrophil population, although the possibility of adhesive interactions representative of monocytic cells remains and may be reflected by the double exponential decay of adhesion time.

At low values of wall shear stress ( $τ_{WALL} < 1.0 {dyn/cm}^{2}$ , Fig. 2(B)), the present off-rates approached 3.0 s⁻¹ when extrapolating the regression of Fig. 2(B) to zero shear stress. This value is consistent with in vitro measurements of neutrophils rolling on lipid bilayers with embedded P-selectin [32]. For a P-selectin site density of 5 sites per µm² neutrophil off-rates rose nearly linearly with $τ_{WALL} 0.17 < τ_{WALL} < 1.1 {dyn/cm}^{2}$ , with a slope of about 3.5 s⁻¹ per dyn/cm². Similar studies of neutrophils rolling on polystyrene slides coated with adsorbed P-selectin at 9 sites/µm² revealed an off-rate of about 7.9 s⁻¹ for $τ_{WALL} = 1 {dyn/cm}^{2}$ , which increased only 30% as shear stress was doubled. These in vitro studies revealed a complete loss of adhesion events within a 0.5 s exposure to shear stress ranging from 1–2 dyn/cm², whereas the present in vivo data fell on average 65% for all 17 venules within the first 0.5 s, for $1.0 ⩽ τ_{WALL} ⩽ 13.2 {dyn/cm}^{2}$ , as obtained by extrapolation of a regression of % adhesion after 0.5 s vs. $τ_{WALL}$ .

To explore the effect of MMP inhibition on fMLP stimulated adhesion a separate set of experiments was conducted, albeit with fewer observations per treatment, compared to those of Fig. 2 (where 300 adhesion events in each of 17 venules yielded a total of 5100 observations). In contrast, the observations of the various treatments entailed on the order of 10–30 observations per venule necessary to expedite the comparisons, except for the fMLP + doxy experiments where only 26 events were monitored in 10 venules. The number of events in each case was also determined by the physiological response, which in the case of fMLP + doxy was muted as noted previously [6]. Pooling of observations for each treatment did not adversely affect the mean adhesion time as evidenced by the absence of a significant difference in mean adhesion time for control conditions between the large data set of Fig. 2, and the limited snapshot of Fig. 4. On average, $τ_{WALL}$ was about 70% greater for the pooled data ( $12.1 \pm 3.8 SD vs . 7.2 \pm 4.0 {SD dyn/cm}^{2}$ , $p < 0.05$ ). However, mean adhesion time averaged 1.73 ± 1.74 SD s for the data of Fig. 2 and 2.51 ± 3.36 s for the pooled data (Fig. 5), which were not significantly different $p = 0.208$ . In contrast, under control conditions $K_{OFF}$ for the pooled data (Fig. 5(B)) was significantly less than that of the large data set (Fig. 2(B)), where averages were 0.60 ± 0.21 SD and 4.9 ± 2.5 SD s⁻¹, respectively, $p < 0.05$ . This disparity may be due, in part, to the larger range of shear stresses in the large data set (Fig. 2) compared to the pooled data (Fig. 4), as evidenced by a coefficient of variation of 0.56 vs. 0.31, respectively, as well as the effect of a smaller sampling of adhesion events for the pooled data.

The trends shown in Fig. 5 are consistent with prior observations on WBC rolling and adhesive interactions with the endothelium. The diminished adhesion time and increased off rate in response to chelation of divalent cations with EDTA has been well established [16,33–35]. The effectiveness of doxycycline as an MMP inhibitor is also supported by the equal effectiveness of the hydroxamic acid zinc chelator, GM6001, that also increases the mean adhesion time (Fig. 5(A)) and diminishes $K_{OFF}$ (Fig. 5(B)), most likely by inhibition of basal sheddase activity on the endothelial surface. The finding that MMP inhibition did not affect the duration and off-rate of adhesion in response to fMLP, while it does diminish the total number of adherent WBCs [6], suggests that MMPs do not affect the avidity of the integrin bonds, but only the availability of ligands on the endothelium. This conclusion is further supported by the simplified calculations of the number of adhesive bonds formed during each treatment (Fig. 6). The number of bonds estimated during spontaneous and stimulated adhesion appear to be a small fraction of the total number available on the neutrophil, which may be on the order of about 50,000 LFA-1 and 15,000 Mac-1 receptors for ICAM-1 that are constitutively expressed, of which about 10,000 are in an activated state [36] and about 2500 VLA-4 receptors for VCAM-1 [37]. Hence the number of bonds estimated for the unstimulated (control) state appears reasonable.

Pretreatment of the tissue with doxycycline had no effect on the number of bonds formed during fMLP stimulated adhesion. The increased number of bonds formed in response to doxycycline was significantly less than that with fMLP (Fig. 6). However, the number of adherent WBCs was not significantly different [16]. Further, the lifespan of these bonds was significantly shorter (Fig. 5(A)). While the precise mechanism of this apparent down-regulation of doxycycline induced adhesion remains to be determined, the adhesion may arise from WBC interactions with glycoproteins or glycosaminoglycans (GAGs) that are limited by normally active MMP activity. Heparin (a more highly sulfated form of heparan sulfate) was found to inhibit adhesion of fMLP treated PMNs to endothelium in a concentration dependent manner [38]. Heparan sulfate, infused into the mesenteric circulation, was found to inhibit WBC rolling [39]. Thus, the competitive binding of exogenous glycans to PMNs may lessen their binding to GAGs on the EC surface. This suggests that increased GAG concentration due to MMP inhibition of basal GAG shedding may enhance WBC-EC adhesion.

In summary then, the present studies aimed to determine if MMP inhibition affected the strength of the adhesive bond between WBC and EC. To that end a simplified model of the mechanics of WBC adhesion was employed that utilized the fundamental model of receptor–ligand mechanics described by Bell [21]. This model has seen much use in the literature and its utility tested by elegant studies of cell–cell interactions by Goldsmith and colleagues [40–46]. Although the current estimates may lack the precision of in vitro determinations, the comparisons performed clearly suggest that the role of MMPs in WBC-EC adhesion warrants further exploration in light of their ability to increase availability of ligands on the endothelium.

Footnotes

Acknowledgements

This work was supported by NIH R01 HL-39286.

References

[1] Kinashi

Katagiri

. Regulation of lymphocyte adhesion and migration by the small gtpase rap1 and its effector molecule, rapl. Immunol Lett. 2004;93:1–5.

[2] Laudanna

Kim

Constantin

Butcher

. Rapid leukocyte integrin activation by chemokines. Immunol Rev. 2002;186:37–46.

[3] Luo

Carman

Springer

. Structural basis of integrin regulation and signaling. Annu Rev Immunol. 2007;25:619–647.

[4] Zarbock

Ley

. Neutrophil adhesion and activation under flow. Microcirculation. 2009;16:31–42.

[5] House

Lipowsky

. Leukocyte-endothelium adhesion: Microhemodynamics in mesentery of the cat. Microvasc Res. 1987;34:363–379.

[6] Mulivor

Lipowsky

. Inhibition of glycan shedding and leukocyte-endothelial adhesion in postcapillary venules by suppression of matrixmetalloprotease activity with doxycycline. Microcirculation. 2009;16:657–666.

[7] Pries

Secomb

Gaehtgens

. The endothelial surface layer. Pflugers Arch. 2000;440:653–666.

[8] Chappell

Jacob

Paul

Rehm

Welsch

Stoeckelhuber

, et al. The glycocalyx of the human umbilical vein endothelial cell: An impressive structure ex vivo but not in culture. Circ Res. 2009;104:1313–1317.

[9] Gotte

. Syndecans in inflammation. FASEB J. 2003;17:575–591.

10.

[10] Smith

Long

Damiano

Ley

. Near-wall micro-piv reveals a hydrodynamically relevant endothelial surface layer in venules in vivo. Biophys J. 2003;85:637–645.

11.

[11] Weinbaum

Tarbell

Damiano

. The structure and function of the endothelial glycocalyx layer. Annu Rev Biomed Eng. 2007;9:121–167.

12.

[12] Marki

Esko

Pries

Ley

. Role of the endothelial surface layer in neutrophil recruitment. J Leukoc Biol. 2015;98(4):503–515.

13.

[13] Lipowsky

. The endothelial glycocalyx as a barrier to leukocyte adhesion and its mediation by extracellular proteases. Ann Biomed Eng. 2012;40:840–848.

14.

[14] Becker

Chappell

Bruegger

Annecke

Jacob

. Therapeutic strategies targeting the endothelial glycocalyx: Acute deficits, but great potential. Cardiovasc Res. 2010;87:300–310.

15.

[15] Fitzgerald

Wang

Park

Murphy

Bernfield

. Shedding of syndecan-1 and -4 ectodomains is regulated by multiple signaling pathways and mediated by a timp-3-sensitive metalloproteinase. J Cell Biol. 2000;148:811–824.

16.

[16] Lipowsky

Sah

Lescanic

. Relative roles of doxycycline and cation chelation in endothelial glycan shedding and adhesion of leukocytes. Am J Physiol Heart Circ Physiol. 2011;300:H415–H422.

17.

[17] Lipowsky

Lescanic

. The effect of doxycycline on shedding of the glycocalyx due to reactive oxygen species. Microvasc Res. 2013;90:80–85.

18.

[18] Wayland

Johnson

. Erythrocyte velocity measurement in microvessels by a two-slit photometric method. J Appl Physiol. 1967;22:333–337.

19.

[19] Lipowsky

Zweifach

. Application of the “two-slit” photometric technique to the measurement of microvascular volumetric flow rates. Microvasc Res. 1978;15:93–101.

20.

[20] Lipowsky

Usami

Chien

. Invivo measurements of apparent viscosity and microvessel hematocrit in the mesentery of the cat. Microvasc Res. 1980;19:297–319.

21.

[21] Bell

. Models for the specific adhesion of cells to cells. vol. 200. New York, NY: Science; 1978. p. 618–627.

22.

[22] House

. Lipowsky HH. In vivo determination of the force of leukocyte-endothelium adhesion in the mesenteric microvasculature of the cat. Circ Res. 1988;63:658–668.

23.

[23] Springer

. Traffic signals for lymphocyte recirculation and leukocyte emigration – the multistep paradigm. Cell. 1994;76:301–314.

24.

[24] Lawrence

Springer

. Neutrophils roll on e-selectin. J Immunol. 1993;151:6338–6346.

25.

[25] Springer

. Traffic signals on endothelium for lymphocyte recirculation and leukocyte emigration. Annu Rev Physiol. 1995;57:827–872.

26.

[26] Diamond

Alon

Parkos

Quinn

Springer

. Heparin is an adhesive ligand for the leukocyte integrin mac-1 (cd11b/cd1). J Cell Biol. 1995;130:1473–1482.

27.

[27] House

Lipowsky

. Dynamics of leukocyte-endothelium interactions in the splanchnic microcirculation. Microvasc Res. 1991;42:288–304.

28.

[28] Bjerknes

Cheng

Ottaway

. Dynamics of lymphocyte-endothelial interactions in vivo. Science. 1986;231:402–405.

29.

[29] Pearson

Lipowsky

. Influence of erythrocyte aggregation on leukocyte margination in postcapillary venules of rat mesentery. Am J Physiol Heart Circ Physiol. 2000;279:H1460–H1471.

30.

[30] Tangelder

Janssens

Slaaf

oude Egbrink

Reneman

. In vivo differentiation of leukocytes rolling in mesenteric postcapillary venules. Am J Physiol. 1995;268:H909–H915.

31.

[31] Xie

Raud

Hedqvist

Lindbom

. In vivo rolling and endothelial selectin binding of mononuclear leukocytes is distinct from that of polymorphonuclear cells. Eur J Immunol. 1997;27:2935–2941.

32.

[32] Alon

Hammer

Springer

. Lifetime of the p-selectin-carbohydrate bond and its response to tensile force in hydrodynamic flow. Nature. 1995;374:539–542.

33.

[33] Hoover

Folger

Haering

Ware

Karnovsky

. Adhesion of leukocytes to endothelium: Roles of divalent cations, surface charge, chemotactic agents and substrate. J Cell Sci. 1980;45:73–86.

34.

[34] Hoover

Briggs

Karnovsky

. The adhesive interaction between polymorphonuclear leukocytes and endothelial cells in vitro. Cell. 1978;14:423–428.

35.

[35] Atherton

Born

. Quantitative investigations of the adhesiveness of circulating polymorphonuclear leucocytes to blood vessel walls. J Physiol. 1972;222:447–474.

36.

[36] Simon

Green

. Molecular mechanics and dynamics of leukocyte recruitment during inflammation. Annu Rev Biomed Eng. 2005;7:151–185.

37.

[37] Lomakina

Waugh

. Adhesion between human neutrophils and immobilized endothelial ligand vascular cell adhesion molecule 1: Divalent ion effects. Biophys J. 2009;96:276–284.

38.

[38] Bazzoni

Beltran Nunez

Mascellani

Bianchini

Dejana

Del Maschio

. Effect of heparin, dermatan sulfate, and related oligo-derivatives on human polymorphonuclear leukocyte functions. J Lab Clin Med. 1993;121:268–275.

39.

[39] Ley

Cerrito

Arfors

. Sulfated polysaccharides inhibit leukocyte rolling in rabbit mesentery venules. Am J Physiol. 1991;260:H1667–H1673.

40.

[40] Kwong

Tees

Goldsmith

. Kinetics and locus of failure of receptor–ligand-mediated adhesion between latex spheres. II. Protein-protein bond. Biophys J. 1996;71:1115–1122.

41.

[41] Goldsmith

McIntosh

Shahin

Frojmovic

. Time and force dependence of the rupture of glycoprotein iib-iiia-fibrinogen bonds between latex spheres. Biophys J. 2000;78:1195–1206.

42.

[42] Tees

Coenen

Goldsmith

. Interaction forces between red cells agglutinated by antibody. IV. Time and force dependence of break-up. Biophys J. 1993;65:1318–1334.

43.

[43] Tees

Goldsmith

. Kinetics and locus of failure of receptor–ligand-mediated adhesion between latex spheres. I. Protein-carbohydrate bond. Biophys J. 1996;71:1102–1114.

44.

[44] Long

Goldsmith

Tees

Zhu

. Probabilistic modeling of shear-induced formation and breakage of doublets cross-linked by receptor–ligand bonds. Biophys J. 1999;76:1112–1128.

45.

[45] Xia

Goldsmith

van de Ven

. Flow-induced detachment of red blood cells adhering to surfaces by specific antigen-antibody bonds. Biophys J. 1994;66:1222–1230.

46.

[46] Simon

Goldsmith

. Leukocyte adhesion dynamics in shear flow. Ann Biomed Eng. 2002;30:315–332.