Lipopolysaccharide reduces intercellular coupling in vitro and arteriolar conducted response in vivo

¹ A. C. Burton Laboratory and ² Child Health Research Institute, Lawson Health Research Institute, and ³ Department of Medical Biophysics, University of Western Ontario, London, Ontario, Canada N6A 5C1

	ABSTRACT

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION APPENDIX REFERENCES

Our recent in vitro study (Lidington et al. J Cell Physiol 185: 117-125, 2000) suggested that lipopolysaccharide (LPS) reduces communication along blood vessels. The present investigation extended this study to determine whether any effect of LPS and/or inflammatory cytokines [tumor necrosis factor-, interleukin (IL)-1, and IL-6] on endothelial cell coupling in vitro could also be demonstrated for an arteriolar conducted response in vivo. Using an electrophysiological approach in monolayers of microvascular endothelial cells, we found that LPS (10 µg/ml) but not these cytokines reduced intercellular conductance (c_i) (an index of cell communication) and that LPS together with these cytokines did not further reduce c_i. Also, c_i was restored after LPS washout, and the LPS-induced reduction was prevented by protein tyrosine kinase (PTK) inhibitors (1.5 µM Tyr A9 and 10 nM PP-2). In our in vivo experiments in arterioles of the mouse cremaster muscle, local electrical stimulation evoked vasoconstriction that conducted along arterioles. LPS in the muscle superfusate did not alter local vasoconstriction but reduced the conducted response. Washout of LPS restored the conducted response, whereas PTK inhibitors prevented the effect of LPS. On the basis of a newly developed mathematical model, the LPS-induced reduction in conducted response was predicted to reduce the arteriolar ability to increase resistance to blood flow. We conclude that LPS can reduce communication in in vitro and in vivo systems comparably in a reversible and tyrosine kinase-dependent manner. Based on literature and present results, we suggest that LPS may compromise microvascular hemodynamics at both the arteriolar responsiveness and the conduction levels.

endothelial cell monolayer; mouse cremaster muscle; tyrosine kinase; mathematical model

	INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION APPENDIX REFERENCES

THE ABILITY OF THE ARTERIOLE to change its diameter is a key determinant in microvascular blood flow control. However, the response of arteriolar wall cells to stimuli impinging directly on these cells and the subsequent local diameter change may not necessarily alter the blood flow. Kurjiaka and Segal (19) showed that a stimulus causing only a local arteriolar dilation yielded no increase in blood flow in the microvascular network fed by the stimulated arteriole. However, a stimulus causing both local and conducted dilation did elevate blood flow. Thus both the local responsiveness of blood vessel wall cells and their ability to conduct responses along the blood vessel length may be required for microvascular blood flow control (19).

Sepsis, a systemic inflammatory response to a local infectious insult, impairs vascular responsiveness and may lead to maldistribution of blood flow in organs and eventually to organ dysfunction (22). Although it is well known that sepsis reduces the vasoconstrictive (17) and vasodilative (35) responsiveness of arterioles, it is not known whether sepsis also compromises arteriolar ability to conduct responses along the arteriolar length. Our recent study demonstrated that lipopolysaccharide (LPS), an initiating factor in sepsis, increased intercellular resistance (r_i) in cultured microvascular endothelial cell monolayers by about 60% (24). This finding suggested that LPS could reduce conduction and/or communication along blood vessels and implied that sepsis could compromise microvascular blood flow control at the arteriolar responsiveness as well as the conduction level.

Our recent finding of increased r_i after exposure to LPS could be questioned for its applicability to an in vivo model; for example, it is not clear what reduction in conducted arteriolar response may correspond to the measured 60% increase in r_i in the cell monolayer. Although a number of reports have characterized the arteriolar-conducted response in several animal models (8, 14, 18, 31, 32) and have addressed the mechanism of conduction (3, 6, 9, 11, 29, 39), little is known about the effect of any particular disease process on the conducted response. Thus in the context of these reports, the main objective of the present study was to determine whether the conducted response can be modulated by agent(s) of the inflammatory process. We tested whether any effect of LPS and/or inflammatory cytokines [tumor necrosis factor- (TNF-), interleukin (IL)-1, IL-6, or a combination thereof] on endothelial cell coupling in vitro could also be demonstrated for a conducted arteriolar response in vivo.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION APPENDIX REFERENCES

Isolation and Culture of Microvascular Endothelial Cells

Rat microvascular endothelial cells (RMEC) were harvested from the extensor digitorum longus muscle as we described previously (24, 37). Briefly, the muscle was enzymatically digested and RMECs were separated from other cells by using Griffonia simplicifolia lectin-coated magnetic beads and a magnetic concentrator. Harvested RMECs were plated and grown on 12-mm diameter glass coverslips in culture medium in a standard incubator. The culture medium consisted of medium 199 (M199, GIBCO, Mississauga, ON, Canada) supplemented with fetal bovine serum (10%, GIBCO), endothelial growth supplement (50 µg/ml, Collaborative Research; Bedford, MA), heparin (5 U/ml, Leo Laboratories; Ajax, ON, Canada), L-glutamine (0.1 mg/ml, GIBCO) and antimycotic-antibiotic solution (10 µl/ml, GIBCO). Cells were periodically tested for markers of endothelial phenotype as previously described (37) and were used between passages 5 and 15.

Measurement of r_i and Intercellular Conductance in Cell Monolayers

In RMEC monolayers, r_i () was determined by an electrophysiological approach and a Bessel function mathematical model as detailed in our recent study (24). Briefly, two cells of the monolayer were impaled by two microelectrodes, each connected to an electrometer (Intra 767, WPI). After reading a stable membrane potential (E_m) in each cell (Fig. 1), four to five hyperpolarizing pulses of current (50 nA, 100- ms duration) were injected into one cell and a change in E_m (E_m) was noted in the other cell. This process was repeated three times for other cell pairs of the monolayer at different microelectrode distances (d) (range 50-400 µm). The plot of E_m versus log (d) was fitted by a zero order Bessel function to determine the r_i of the monolayer (24). The intercellular conductance, c_i = 1/r_i (Siemens), was used in the present study as a measure of the spread of the injected current (an index of coupling/communication) between the cells of the monolayer.

View larger version (26K): [in this window] [in a new window]   Fig. 1.   Design of electrophysiological experiments in rat microvascular endothelial cell (RMEC) monolayers. Two cells of the monolayer were impaled by two microelectrodes separated by 50 µm, aiming to obtain a stable membrane potential (Em) in each cell (left, Em1 in one cell; right, Em2 in the other cell). Then, 4-5 hyperpolarizing pulses were injected into the first cell and the resulting hyperpolarization was noted in the other cell (change in Em2). Then, electrode Em1 was lifted from the cell and inserted into another cell at a distance of 100 µm from electrode Em2 (this electrode remained inserted in its cell), pulses were injected into the new cell, and the resulting hyperpolarization in Em2 was noted. The same procedure was repeated for interelectrode distances of 200 and 400 µm. Changes in hyperpolarization in Em2 vs. the interelectrode distance were used to compute the intercellular resistance, ri, as detailed by Lidington et al. (24). (Adapted from Ref. 24).

Mouse Cremaster Muscle Preparation for Intravital Microscopy

To study the conducted response, we used a mouse preparation rather than a rat preparation to permit future mechanistic studies in genetically altered mice. The procedure for preparation of mouse cremaster and all experimental protocols were approved by the Council on Animal Care at the University of Western Ontario. Male mice (C57BL/6, 20-30 g) were anesthetized with a mixture of ketamine (80 mg/kg) and xylazine (4 mg/kg) injected intraperitoneally. Supplementary injections of this mixture (30% of initial dose) were given as needed. In some mice, the left carotid artery was cannulated to permit measurement of the arterial pressure. The mouse was placed in a supine position on a styrofoam board and kept warm by radiant heat to keep the rectal temperature at 37°C. The cremaster muscle preparation was similar to that described for the rat (34). The skin of the scrotum and its underlying connective tissue were cut longitudinally along the anterior midline, exposing the cremaster muscle sac. The sac was isolated from the scrotum, placed on a histological glass slide, and irrigated with physiological saline solution (PSS), pH 7.4, composed of (in mM) 131.9 NaCl, 4.7 KCl, 2.2 CaCl₂, 1.2 MgSO₄, and 20.0 NaHCO₃ bubbled continuously with 95% N₂-5% CO₂ gas. A longitudinal cut was made in the ventral surface of the muscle to open the sac. Sutures were stitched at the muscle edge, pulled, and pinned to the styrofoam board to gently spread the muscle over the glass slide. The testicular mesentery and the vessels attached to it were cauterized and cut to separate the contents of the sac from the muscular sac itself. The contents were pushed into the abdominal cavity; the cremaster muscle was then continuously superfused with PSS (33-34°C) at a rate of 3-4 ml/min. The muscle was epi-illuminated by means of a fiber-optic light guide and visualized with an intravital microscope (ELR, Leitz) with a long-working distance objective (×20/0.32 numerical aperture) and an eyepiece (×6.3). The resultant field of view (0.56 × 0.38 mm) was video recorded by a closed circuit system (MTI camera, Panasonic WV5410 monitor, and Mitsubishi U82 sVHS tape recorder). Inner luminal arteriolar diameters were measured off-line from the video screen with resolution of about ±1 µm.

Experimental Protocols in Vitro

We have previously shown (24) that exposure of RMECs to LPS for 1-24 h at a concentration of 10 µg/ml increases r_i from 3.3 to 5.3 M (a maximal r_i already occurs at 1 h). In terms of LPS concentration dependency, increased r_i was already seen at 10 ng/ml (24). In the present study, we aimed to determine whether 1) exposure of RMECs to a combination of LPS and inflammatory cytokines further elevates r_i using TNF-, IL-1, and IL-6, agents known to be elevated in septic plasma (1, 10); 2) the effect of LPS + cytokines is reversible; and 3) the effect can be prevented by pretreatment of RMECs with protein tyrosine kinase (PTK) inhibitors. The outcome of these in vitro experiments dictated the design of subsequent in vivo experiments. The concentrations of LPS (10 µg/ml, isolated from Escherichia coli serotype 055:B5), TNF- (100 ng/ml), IL-1 (50 ng/ml) (all from Sigma), and IL-6 (50 ng/ml, R&D Systems; Minneapolis, MN) were chosen at the upper end of the concentration range reported in the literature (2, 30) and were expected to produce a maximum effect (30). We used PTK inhibitors tyrphostin (Tyr) A9 (1.5 µM, ICN Radiochemicals; Costa Mesa, CA) and PP-2 (10 nM, Calbiochem; La Jolla, CA). The reason for choosing these two structurally different agents was to minimize the possibility of their nonspecific effect. To this end, we also used the inactive PTK inhibitor Tyr A1 (0.1 mM, ICN Radiochemicals) as an "isoform control." The concentrations of PP-2 and Tyr A1 and A9 were based on previous reports (4, 7, 16) and the IC₅₀ values reported by the manufacturer. Stock solutions of inhibitors were dissolved in 100% DMSO; the final concentration of DMSO in the culture medium was 0.01% unless otherwise stated.

Experimental Protocols in Cremaster Muscle in Vivo

Arterioles in the mouse cremaster muscle have been shown to exhibit conducted responses after application of local dilatory and constriction stimuli (18, 20). We used the approach of Gustafsson and Holstein-Rathlou (14) to stimulate arterioles locally with a train of unipolar depolarization pulses. The rationale for using this approach rather than that of local agonist application (18, 20) was our wish to exclude possible effect(s) of LPS/cytokines on the initiation of conducted response (agonist-induced local diameter change). Briefly, glass micropipettes of outer tip diameter 6-10 µm were backfilled with 2 M NaCl solution and connected to a Grass stimulator (model S88) via an Ag-AgCl wire. A reference electrode was inserted into the mouse tail. Each pipette was mounted on a micromanipulator and the tip was positioned as close to an arteriole as possible to deliver a 40-s train of pulses at 10 Hz frequency, 2 ms duration, and 40-80 V amplitude. In each experiment, after the surgical exposure of the cremaster muscle and a 30-min stabilization period, a site in the middle portion of a 1.5- to 3-mm-long unbranched arteriole (~50 µm diameter, 1A or 2A branching order) was chosen for local electrical stimulation. A key criterion for the arteriole selection was clear visibility of the arteriolar wall at the local stimulation site and at a site 500 µm upstream from the pipette tip. At the beginning of the experiment (one arteriole per mouse), the amplitude was adjusted to produce ~50% reduction in the local arteriolar diameter near the pipette tip. For any subsequent treatments of the arteriole/cremaster muscle, we have stimulated exactly the same local site with the same amplitude determined at the beginning of the experiment. In the majority of experiments, we simultaneously video recorded diameter changes occurring locally and at the 500-µm upstream site. On the basis of prestimulation diameter measurements (D_local,pre and D_500,pre) and the minimal diameter measurements during the 40-s stimulation period (D_local,min and D_500,min), we defined the relative local and upstream diameter changes as D_local(%) = 100% × (D_local,min  D_local,pre)/D_local,pre and D₅₀₀(%) = 100% × (D_500,min  D_500,pre)/D_500,pre, respectively. We used the communication ratio at the 500-µm site, computed as CR₅₀₀(%) = 100% × D₅₀₀(%)/D_local(%), as an index of the conducted response.

Effect of LPS on conducted response. The design of the present in vivo experiments was driven by our preceding cell culture work. Because cytokines had no effect on c_i (see RESULTS for details), the present in vivo work dealt with LPS only. On the basis of LPS dose, time dependencies determined by us in vitro (24), and limited stability of the cremaster muscle preparation (3-4 h), the in vivo protocol included 10 µg/ml LPS application for 1 h to ensure maximal effect. The protocol consisted of the 30-min stabilization period and cremaster muscle superfusion with PSS + 0.01% DMSO, followed by two to three repeated stimulations (~3 min between stimulations) of the same local arteriolar site with the predetermined amplitude. The outcome of these repeated stimulations served to verify the reproducibility of the response. The muscle was then superfused for 1 h with PSS containing 0.01% DMSO and 10 µg/ml LPS; the same local arteriolar site was again stimulated 2-3 times. Finally, the muscle was superfused for 1 h with PSS + 0.01% DMSO only (washout period), and the same arteriole was again stimulated 2-3 times. For each set of the 2-3 stimulations, D_local,pre, D_500,pre, D_local,min, and D_500,min were measured and then averaged. On the basis of these averages, D_local(%), D₅₀₀(%), and CR₅₀₀(%) were computed.

Effect of PTK inhibitors and LPS on conducted response. Following the stabilization period, an arteriole was initially stimulated 2-3 times and then subjected to one of the following three protocols. The first protocol included 1-h superfusion with PSS + 0.01% DMSO, followed by 1-h superfusion with 10 µg/ml LPS in PSS + 0.01% DMSO followed by 2-3 repeated stimulations. The second protocol consisted of 1-h superfusion with a PTK inhibitor (1.5 µM Tyr A9 or 10 nM PP-2) in PSS + 0.01% DMSO or inactive Tyr A1 (0.1 mM) in PSS + 0.1% DMSO, followed by 1-h superfusion with PTK inhibitor plus 10 µg/ml LPS in PSS + 0.01/0.1% DMSO, and then followed by 2-3 repeated stimulations. The third protocol consisted of 2-h superfusion with PSS + 0.01% DMSO followed by 2-3 repeated stimulations (yielding the time-matched control response).

Control Experiments

To assess the direct effect of electrical stimulation on the diameter response at the 500-µm upstream site, we positioned the microelectrode tip at a site 100 µm away from the original local site (the tip to upstream site distance was maintained at 500 µm) and observed responses at the local and upstream sites. To assess the role of arteriolar innervation in the presently studied conducted responses, we evaluated CR₅₀₀ and CR₁₀₀₀ (computed similarly to CR₅₀₀) after tetrodotoxin (TTX, 10 µM final concentration) was added to the PSS superfusate and allowed to equilibrate for 15-20 min. The positive test for TTX included surgical exposure of the right peroneal nerve and the right extensor digitorum longus muscle of the mouse, supramaximal stimulation of the nerve with a pair of Ag electrodes (5-10 V, 6 Hz), and the subsequent observation of muscle twitching. The nerve was then superfused with the 10 µM TTX solution for 15-20 min, and the response to supramaximal stimulation (no twitching) was noted. Finally, in separate mice, we assessed the vasomotor tone (or vasodilative potential) of the approximately 50-µm arterioles used in the present study. Similar to Hungerford et al. (18), a maximal dilation was obtained by adding ACh (0.1 mM final concentration) to the PSS superfusate and by allowing 10- to 15-min equilibration.

Statistics

All data were expressed as means ± SE. Parameters were analyzed using an analysis of variance, followed by t-test with Bonferroni correction for multiple comparisons when applicable. A level of P < 0.05 was considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION APPENDIX REFERENCES

Effect of LPS and cytokines on c_i in in vitro experiments. Figure 1 exemplifies the electrophysiological approach we used to determine the r_i (or c_i) of the RMEC monolayer. Figure 1A demonstrates that the size of the current-induced hyperpolarization (E_m2) decreased with increasing interelectrode distance. Using the Bessel function mathematical model, the rate of decrease with distance was used to estimate r_i (details in Ref. 24). The average resting E_m was 29 ± 1 mV (n = 316 impaled cells); any of the subsequent treatments of the monolayer had no significant effect on resting E_m. For untreated control monolayers, the values were r_i = 3.5 ± 0.1 M, c_i = 0.28 ± 0.01 µS (Fig. 2). LPS significantly reduced c_i but LPS applied together with all of the three cytokines did not result in a further reduction in c_i (Fig. 2). None of these cytokines applied alone had a significant effect on c_i. Based on these findings, all of our subsequent in vivo experiments used LPS alone to mimic the effect of sepsis on conducted arteriolar response.

View larger version (13K): [in this window] [in a new window]   Fig. 2.   Effect of lipopolysaccharide (LPS) and cytokines on intercellular conductance (ci). LPS (10 µg/ml) and the cytokines tumor necrosis factor (TNF)- (100 ng/ml), interleukin (IL)-1 (50 ng/ml), and IL-6 (50 ng/ml) were tested alone and in combination for an effect on intercellular conductance (ci = 1/ri). After a 2-h exposure, LPS alone and the combination of three cytokines plus LPS caused a 33% reduction in ci. Cytokines alone had no effect. Similar results were obtained for a 24-h exposure, instead of 2-h exposure (data not shown). The control ci value was based on 36 measurements of ci from 12 monolayers; each of the remaining values was based on 12 measurements of ci from 3 monolayers. *Significant difference from control.

Baseline measurements and control experiments in in vivo model. The mean arterial pressure was typically 85-100 mmHg during experiments (average 93 ± 3 mmHg, subset of 17 mice). On the basis of visual assessment of the microvascular flow in the cremaster muscle and on blood pressure measurement, preparations were stable during the 3- to 4-h experimental protocol. In the subgroup of 1A and 2A arterioles used to determine the vasodilative potential (average diameter (D) 44 ± 7 µm, n = 10), ACh dilated arterioles by D = 4.1 ± 5 µm (13 ± 4% diameter increase). Figure 3A exemplifies the time course of arteriolar diameter changes measured at the tip of the electrode (local) and 500 µm upstream during a 40-s electrical stimulation period in PSS superfused muscle. At the local site, the diameter reached a minimum quickly (within the first 10-s period after the onset of the stimulus), and then it tended to recover before the end of the stimulus. At the 500-µm site, the arteriole also constricted quickly, but the minimum diameter was reached somewhat later than the minimum at the local site. Figure 3B underscores these features, based on the average local and 500-µm site responses normalized with respect to the prestimulation diameter. The size of the standard error bars in Fig. 3B reflects an appreciable variability observed in the time course of local and upstream diameter responses among different arterioles. Figure 4A summarizes observed diameter changes at the local, 500-µm, and 1,000-µm upstream sites during control PSS superfusion and after TTX application. Figure 4B shows the communication ratios (CR₅₀₀ and CR₁₀₀₀) for these two protocols. Because our control experiments with peroneal nerve stimulation showed that our TTX solution was effective, we show in Fig. 4 that the nerves did not participate in the conducted response measured under the present experimental conditions. Finally, repositioning of the electrode tip 100 µm away from the arteriolar wall but delivering a comparable stimulus (63 ± 4 V, n = 6) as in Fig. 4 abolished diameter responses at the local and upstream sites (data not shown). Thus responses at the upstream sites required the presence of local diameter responses.

View larger version (16K): [in this window] [in a new window]   Fig. 3.   A: Example of time course of diameter (D) changes in a mouse cremaster muscle arteriole. Changes were induced by a 40-s local electrical stimulation of the arteriole and were measured at the microelectrode tip (local) and 500 µm upstream from the tip (500 µm). B: time course of relative diameter [D(%)] changes caused by electrical stimulation. To compute D(%), the stimulation-induced reduction in diameter (in µm) was divided by the prestimulation diameter. Data are based on measurements from 6 arterioles in 6 mice.

View larger version (30K): [in this window] [in a new window]   Fig. 4.   A: Effect of tetrodotoxin (TTX, 10 µM) on the relative arteriolar diameter change measured at the tip of the stimulation pipette (local) 500-µm and 1,000-µm upstream sites. In 5 arterioles, diameter changes were measured at local and 500-µm sites. In 3 of these arterioles, diameter changes were also measured at the 1,000-µm site. Local prestimulation diameter before application of TTX (58.1 ± 8.2 µm, n = 5 arterioles in 5 mice) was not altered by the TTX treatment. Average amplitude of the local stimulus was 68 ± 6 V. B: communication ratio, CR, determined from the data in A for the 500-µm and 1,000-µm sites [CR500(%) = 100% × D500(%)/ Dlocal(%), CR1000(%) = 100% × D1000(%)/Dlocal(%)]. CR was used as an index of the conducted response. There was no effect of TTX on the conducted response.

Effects of LPS, wash, and PTK inhibitors on c_i in vitro and conducted response in vivo. Although our in vitro experiments were completed before in vivo work began, data on LPS, wash, and PTK inhibitors are presented together to permit comparison between our in vitro and in vivo models. Figure 5A demonstrates that 10 µg/ml LPS was significantly reduced, but a subsequent 1-h wash restored, the c_i in monolayers in vitro (raw electrophysiological data for 5A were also used in Ref. 24). Figure 5C shows that the same concentration of LPS in the superfusate reduced the conducted response in the cremaster muscle in vivo, and that 1-h wash restored it [changes in D₅₀₀(%), Fig. 5B, showed comparable reduction and restoration]. LPS did not affect the local diameter response [D_local(%), Fig. 5B] nor the time delays of the measured diameter minima (time after the stimulus onset 11 ± 1 and 28 ± 3 s for pre-LPS stimulation, n = 6; and 14 ± 2 and 28 ± 3 s for post-LPS stimulation for local and upstream minima, respectively). Thus based on comparison of Fig. 5, A and C, LPS caused qualitatively similar effects on communication in our in vitro and in vivo models.

View larger version (29K): [in this window] [in a new window]   Fig. 5.   A: Washout of LPS. In endothelial cell monolayers, a 2-h exposure of LPS (10 µg/ml) was followed by replacement of the exposure medium by control medium. After 1 h of wash, ci returned to the control level. Control values are based on 12 measurements of ci from 4 monolayers; LPS and washout (1-2 h) values are each based on 20 measurements of ci from 5 monolayers. B: effect of LPS (10 µg/ml, 1 h) in the cremaster muscle superfusate and of wash (1 h) on the relative arteriolar diameter change measured at local and 500-µm sites. Local prestimulation diameter (55.6 ± 6.3 µm, n = 6 arterioles in 6 mice) was not altered by the LPS or wash treatments. Average amplitude of the local stimulus was 59 ± 6 V. C: communication ratio, CR500, was determined from the data in B. *Significant difference from control.

Figure 6, A and B, summarizes the effect of PTK inhibitors. In both models, pretreatment with the same concentration of Tyr A9 and PP-2 prevented the LPS-induced reduction in communication (raw data for LPS + PP-2 bar in Fig. 6A were also used in Ref. 24). As well, pretreatment with the inactive inhibitor Tyr A1 had no effect on the LPS-induced reduction. Thus comparison of Fig. 6, A and B, revealed that PTK inhibitors caused a qualitatively similar effect on LPS-induced reduction in communication in both of our models. In our in vivo model, the application of Tyr A1, A9, or PP-2 alone had no effect either on the local diameter response to the electrical stimulus or on D₅₀₀(%); in our in vitro model, these agents alone had no effect on c_i (data not shown).

View larger version (15K): [in this window] [in a new window]   Fig. 6.   Effect of tyrosine kinase inhibitors. A: in endothelial cell monolayers, a 1-h pretreatment with tyrphostin (Tyr) A9 (1.5 µM in 0.01% DMSO) but not the inactive control Tyr A1 (100 µM in 0.1% DMSO) prevented the reduction in intercellular conductance after 24-h exposure to 10 µg/ml LPS. The inhibitor PP-2 (10 nM in 0.01% DMSO) also prevented the LPS response. Values of ci in control and LPS groups are based on 44 measurements from 11 monolayers; all other values are based on 12-16 measurements of ci from 3-4 monolayers. B: CR500 for arterioles exposed to PSS + 0.01% DMSO (control group), LPS, Tyr A9 + LPS, PP-2 + LPS, or Tyr A1 + LPS at superfusate concentrations identical to those used in vitro experiments of A. Separate arterioles (one arteriole/mouse) were used in each of the 5 groups shown. Among the 5 groups, the number of arterioles, the average local prestimulation diameter, and the average stimulus amplitude ranged from 5 to 8, from 42.0 ± 2.5 to 64.4 ± 6.6 µm, and from 64 ± 5 to 75 ± 2 V, respectively. None of the treatments affected the prestimulation diameter, whereas there was no effect of DMSO (either 0.1 or 0.01%) on ci or CR500 (data not shown). *Significant difference from control.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION APPENDIX REFERENCES

The objectives of this study were to determine whether communication data obtained from our in vitro monolayer model could provide a suitable framework for our in vivo animal experiments, and whether LPS can modulate conducted arteriolar response in the mouse cremaster muscle. On the basis of qualitatively comparable data in both in vitro and in vivo models, our results indicate that LPS attenuates microvascular cell-to-cell communication in a reversible and tyrosine kinase-dependent manner.

Experimental models. A considerable amount of work has been done in the area of vascular cell communication, based on in vitro (see review in Ref. 5) and in vivo approaches (see review in Ref. 15). However, to our knowledge, there are no studies exploiting both approaches simultaneously. In the present study, we took advantage of our recent in vitro work where LPS was shown to increase r_i in RMEC monolayers (24). In our preliminary experiments (unpublished observation) using the same electrophysiological approach as that of our RMEC study, LPS also increased r_i in monolayers derived from mouse skeletal muscle endothelial cells. Thus based on this background work and on the mechanistic potential of genetically engineered mice, we chose the mouse cremaster muscle as an in vivo model to complement our RMEC monolayer model. In general, the presently measured E_m and the intercellular communication ability in the RMEC monolayer were consistent with previous reports (27, 28). Similarly, the baseline hemodynamic parameters of the cremaster muscle preparation (e.g., mean arterial blood pressure, stability of perfusion) and the ability to conduct local diameter changes along the arteriole length agreed with published reports of conducted response in mice (6, 18, 20).

In the present study, we used the electrical depolarization protocol of Gustafsson and Holstein-Rathlou (14) to elicit conducted vasoconstriction. There were two reasons for choosing this protocol. First, because the vasodilative potential of our arterioles was modest (13%), the detectability of conducted vasodilation and its modulation by LPS could have been compromised by the limited resolution of our diameter measurement technique (±1 µm). Thus conducted vasoconstriction would provide a better "signal-to-noise" ratio. Second, we wanted to exclude the possible dependence of the size of the conducted response on the size of the initial local response (38). Thus we used the protocol to set the size of the local response to a desired level by adjusting the amplitude of the stimulus. In general, the maximal local diameter reduction elicited by a particular voltage was highly variable among arterioles (reflecting variability in the muscle surface-to-arteriole distance, pipette tip diameter, or responsiveness of the vessel itself) (20). The range of amplitudes (40-80 V) used to achieve the desired local constriction (~50%) was comparable to the voltage range reported for the rat mesentery (14).

Local and conducted responses under control experimental conditions. The size of local and conducted responses at 500-µm and 1,000-µm sites and the tendency of the conducted response to lag behind the local response (Figs. 3 and 4) were comparable to those reported for the rat mesentery (14). The time course of local vasoconstriction during the 40-s stimulation period (Fig. 3B) agreed with the time course of vasoconstriction in mouse cremaster muscle arterioles (20). The rapid onset of constriction followed by a gradual reduction in constriction during the stimulation (Fig. 3B) may be accounted for by coordinated smooth muscle and endothelial cell function (calcium signaling-induced nitric oxide release) (20, 40). Figure 4 shows that the present local and conducted responses were independent of perivascular nerves as TTX (a fast sodium channel antagonist) had no effect on these responses. This finding is consistent with that of Kumer et al. (20), but it appears to disagree with the results of electrical depolarization protocol reported by Hungerford et al. (18). The disagreement could be due to differences in stimulation parameters (1 ms pulse at 32 Hz, 60-100 V) (18) and the size of the local constriction employed (80%).

Effect of LPS on vascular cell communication. Although the effect of LPS on conducted response can be studied in mice injected with LPS, the experimental outcome could be difficult to interpret because it may depend on the direct effect of LPS on the microvasculature as well as on the systemic response to LPS. For this reason, the present study involved adding LPS to the cremaster muscle superfusate to keep LPS as local as possible to minimize systemic effects.

Data from cell culture experiments (Figs. 2, 5A, and 6A) extended the findings of our recent study (24) where LPS increased r_i in RMEC monolayers. The data (Fig. 2) show for the first time that cytokines TNF-, IL-1, and IL-6 applied alone had no effect on conductance and that LPS applied together with all of the three cytokines did not further reduce conductance. Data in Fig. 6A agree with our report that PTK inhibitors prevent the effect of LPS on cell-to-cell communication (24).

Data from our experiments in mice (Figs. 5C and 6B) demonstrate for the first time that 1) LPS reduced the conducted arteriolar response, 2) a wash restored the response, and 3) PTK inhibitors prevented the effect of LPS. Furthermore, data in Figs. 5C and 6B permit comparison with our cell culture work (Figs. 5A and 6A). Clearly, our in vitro and in vivo models differed. The spread of signal(s) in two dimensions of endothelial cell monolayer might not be comparable to the prevalently one-dimensional spread of signal(s) in endothelial and smooth muscle cells along the arteriolar wall (40). Yet, despite these differences, LPS, LPS washout, and PTK inhibitors had strikingly similar effects in the two models. Although it is possible that the parallel outcome for the five treatments (LPS, LPS washout, LPS + TYR A9, LPS + PP-2, and LPS + TYR A1) was coincidental, it is also possible that the outcome reflected a fundamental mechanism common to both models. One such mechanism could be the spread of signal(s) in both models via gap junctional (GJ) communication (24, 25). LPS could modulate GJ function in both models similarly, including activation of the PTK pathway. Although the mechanism of this modulation has not been clarified, PTK inhibitors in the present experiments could have prevented LPS receptor-mediated phosphorylation of cytosolic kinases or GJ proteins. Cytosolic Src kinases (e.g., pp60^src) have been shown to phosphorylate tyrosine residues in GJ proteins and reduce cell-to-cell communication (26) or, in turn, activate other tyrosine kinases (e.g., p125^Fak) which could also phosphorylate GJ tyrosine residues (21, 23).

Conducted response in arterioles in vivo has been shown to be reduced by GJ uncouplers (31) and enhanced by angiotensin II (14). To our knowledge, the physiological impact of modulation of conducted response has not been addressed. Referring to Figs. 5 and 6, it is difficult to estimate this impact, based on the LPS-induced 40-50% reduction in c_i in our in vitro model. However, this task may be easier considering the reduction in the conducted response (40-50% reduction in CR₅₀₀). To this end, we have developed a mathematical model and used the present data to estimate the modulatory effect of LPS on the resistance to blood flow (R) in an unbranched arteriole (APPENDIX). Assuming that the conducted response in our mice was of the same exponential character as that reported for the rat mesenteric arteriole (employing the same stimulation protocol) (14), our local 50% constriction was predicted to increase R in control arterioles by a factor of 4.31 (Fig. 8). The same local constriction during exposure to LPS was predicted to increase R by a factor of 2.83 (Fig. 8). Thus LPS could reduce the arteriolar ability to control resistance by ~30%. Thus it is possible that during an LPS-induced inflammatory response the effect of the documented reduced vasoconstrictive ability (13, 33) could be aggravated by the reduced ability to conduct constriction along the blood vessel length. To this end, agents aimed at restoring/enhancing communication in blood vessel wall (enhancing the ability to increase peripheral resistance) could possibly be beneficial against LPS-induced hypotension. Clearly, given the multiple effects of LPS on the vessel wall, further studies are needed to address this possibility.

The present approach of cell culture-driven in vivo work in mice may provide a framework for such future studies. Recently, Giepmans et al. (12) indicated that inhibition of GJ communication in cultured fibroblasts was caused by c-Src-mediated phosphorylation of residue Y265 on the COOH-terminal tail of GJ protein Cx43. If our in vitro model would indicate that this Y265 residue also mediates the LPS-induced reduction in communication, then a specific transgenic mouse with a mutation at this residue could be used to examine 1) lack of effect of LPS on conducted response and 2) possible attenuation of hypotension in LPS-injected mice.

In conclusion, the present study showed that 1) LPS reduced arteriolar conducted response in mouse cremaster muscle in a reversible and tyrosine kinase-dependent manner, 2) in vitro data predicted the effect of LPS in this mouse model, and 3) the degree of reduction of conducted response had the potential to appreciably affect the microcirculatory hemodynamics.

	APPENDIX

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION APPENDIX REFERENCES

The model. To assess the effect of local arteriolar diameter change on the change in resistance to blood flow (R) in the arteriole, and the modulating effect of LPS on this relationship, the following model incorporated these assumptions. 1) R (mmHg · ml¹ · min) could be computed from the Poiseuille's law, such that R = kl_A/r⁴, where l_A (mm) is the length of unbranched arteriole, r (mm) is the inner (luminal) radius of the arteriole, and k (ml · min¹ · mmHg¹ · mm³) is a constant that reflects the viscosity of the blood in the arteriole; 2) viscosity of blood, and k, do not change with r; and 3) a stimulus is applied at the midpoint of the arteriole to cause a localized radius change. This change spreads equally toward both ends of the arteriole, and the size of this change decreases with distance according to a simple exponential decay (Fig. 7).

View larger version (12K): [in this window] [in a new window]   Fig. 7.   Schematic diagram outlining the features of the mathematical model in APPENDIX. Heavy dashed line depicts the arteriole before local stimulation. A local stimulus introduced via a pipette at position i = 0 constricts the arteriole locally and via cell-to-cell communication also along the vessel length (heavy solid line). The shape of the constricted arteriolar lumen is approximated by a series of cocentric cylindrical sleeves (total numbers of sleeves is N), as detailed in APPENDIX. The length of each sleeve is x, rc is the control radius, rsi represents the sleeve radii, and L is the length of half of the arteriole.

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

4

View larger version (15K): [in this window] [in a new window]   Fig. 8.   Plots of the normalized resistance to flow, Rnorm, vs. normalized arteriolar diameter, S, according to Eq. 8 in APPENDIX (N in Eq. 8 is set to 1,000). Rnorm is shown for S ranging from 0.5 to 2 (from 50% local constriction to 100% dilation) and for two C values (Ccontrol and CLPS) estimated from the present in vivo data. C represents the degree of conducted response normalized to the arteriolar length (details in APPENDIX). According to the model in APPENDIX, LPS treatment was predicted to attenuate the arteriolar ability to elevate (e.g., at S = 0.5) and to lower (e.g., at S = 2) the resistance to flow offered by the arteriole.

Application of present experimental data to model. In the following, we wanted to determine C in Eq. 8 to plot R_norm versus S. In particular, we aimed to estimate C under the present control (C_control) and LPS superfusion conditions (C_LPS). According to Gustafsson and Holstein-Rathlou (15), the conducted response can be described by the relationship D_x = D₀ e^(x/) (D is arteriolar diameter). If our diameter responses measured locally (x = 0) and at x = 500 µm upstream obeyed this relationship, then

(9)

Assuming further that the prestimulation diameters at local and 500-µm upstream locations were of the same value (D_local,pre = D_500,pre, Fig. 3A), and dividing both sides of Eq. 9 by this value, the equation could be rewritten as D₅₀₀(%) = D₀(%) e^(500/). Because CR₅₀₀(%) was defined as 100% × D₅₀₀(%)/D_local(%), Eq. 9 could further be rewritten as

(10)

Let us consider an arteriole at 10 s after the onset of the stimulus (Fig. 3). At this time, the arteriole constricted to its minimum diameter locally and almost to its minimum diameter at the 500-µm upstream location (Fig. 3B), both under control and LPS conditions. Thus the average CR₅₀₀(%) values at the 10-s time point are approximately equal to the CR₅₀₀(%) values shown in Fig. 5C (based on the maximal constrictions). Using the data in Fig. 5C (CR_500,control = 45% and CR_500,LPS = 21%) and Eq. 10, we computed _control = 626 µm, and _LPS = 320 µm. In a subset of our experiments we measured the average length of unbranched arterioles to be 1,900 ± 370 µm (n = 6) giving the average L = 950 µm. Setting 950 µm to be one "length unit", _control was then 0.66 units and _LPS was 0.34 units, whereas the normalized degree of conduction C_control = 0.66 and C_LPS = 0.34.

	ACKNOWLEDGEMENTS

The authors thank A. Bihari, M. Keet, and L. Cheng for technical help, and Dr. H. Ladak for help with Matlab programming of the mathematical model in APPENDIX.

	FOOTNOTES

The Heart and Stroke Foundation of Ontario and the Canadian Institutes of Health Research provided financial support.

Address for reprint requests and other correspondence: K. Tyml, Dept. of Medical Biophysics, Univ. of Western Ontario, London, Ontario, Canada N6A 5C1 (E-mail: ktyml{at}lhsc.on.ca).

The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Received 9 February 2001; accepted in final form 6 June 2001.

	REFERENCES

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION APPENDIX REFERENCES

1.   Ayala, A, Kisala JM, Felt JA, and Chaudry IH. Does endotoxin tolerance prevent the release of inflammatory monokines (IL-1, Il-6, TNF-) during sepsis? Arch Surg 127: 191-197, 1992[Abstract/Free Full Text].

2.   Baudry, N, Rasetti C, and Vicaut E. Differences between cytokine effects in the microcirculation of the rat. Am J Physiol Heart Circ Physiol 271: H1186-H1192, 1996[Abstract/Free Full Text].

3.   Beny, JL. Electrical coupling between smooth muscle cells and endothelial cells in pig coronary artery. Pflügers Arch 433: 364-367, 1997[Web of Science][Medline].

4.   Burger, AM, Kaur G, Alley MC, Supko JG, L, Malspeis Grever MR, and Sausville EA. Tyr AG17, [(3,5-Di-tert-butyl-4-hydroxybenzylidene)-malononitrile], inhibits cell growth by disrupting mitochondria. Cancer Res 55: 2794-2799, 1995[Abstract/Free Full Text].

5.   Christ, GJ, Spray DC, El-Sabban M, Moore LK, and Brink PR. Gap junctions in vascular tissues. Evaluating the role of intercellular communication in the modulation of vasomotor tone. Circ Res 79: 631-646, 1996[Abstract/Free Full Text].

6.   De Wit, C, Roos F, Boltz SS, Kirchhof S, Kruger O, Willecke K, and Pohl U. Impaired conduction of vasodilation along arterioles in connexin40-deficient mice. Circ Res 86: 649-655, 2000[Abstract/Free Full Text].

7.   Dickenson, JM, and Hill SJ. Potentiation of adenosine A₁ receptor-mediated inositol phosholipid hydrolysis by tyrosine kinase inhibitors in CHO cells. Br J Pharmacol 125: 1049-1057, 1998[Web of Science][Medline].

8.   Dietrich, HH, Kajita Y, and Dacey RGJ Local and conducted vasomotor responses in isolated rat cerebral arterioles. Am J Physiol Heart Circ Physiol 271: H1109-H1116, 1996[Abstract/Free Full Text].

9.   Doyle, MP, Duling BR, McGahren ED, and Dora KA. Acetylcholine induces conducted vasodilation by nitric oxide-dependent and -independent mechanisms. Am J Physiol Heart Circ Physiol 272: H714-H721, 1997[Abstract/Free Full Text].

10.   Ertel, W, Morrison MH, Wang P, Zheng F, Ayala A, and Chaudry IH. The complex pattern of cytokines in sepsis. Ann Surg 214: 141-148, 1991[Web of Science][Medline].

11.   Frame, MD, and Sarelius IH. L-Arginine-induced conducted signals alter upstream arteriolar responsiveness to L-arginine. Circ Res 77: 695-701, 1995[Abstract/Free Full Text].

12.   Giepmans, BNG, Hengeveld T, Posma FR, and Moolenaar WH. Interaction of c-Src with gap junction protein connexin-43: role in the regulation of cell-cell communication. J Biol Chem 276: 8544-8549, 2001[Abstract/Free Full Text].

13.   Gunnet, CA, Chu Y, Heistad DD, Loihl A, and Faraci FM. Vascular effects of LPS in mice deficient in expression of the gene for inducible nitric oxide synthase. Am J Physiol Heart Circ Physiol 275: H416-H421, 1998[Abstract/Free Full Text].

14.   Gustafsson, F, and Holstein-Rathlou NH. Angiotensin II modulates conducted vasoconstriction to norepinephrine and local electrical stimulation in rat mesenteric arterioles. Cardiovasc Res 44: 176-184, 1999[Abstract/Free Full Text].

15.   Gustafsson, F, and Holstein-Rathlou NH. Conducted vasomotor responses in arterioles: characteristics, mechanisms and physiological significance. Acta Physiol Scand 167: 11-21, 1999[Web of Science][Medline].

16.   Hanke, JH, Gardner JP, Dow RL, Changelian PS, Brissette WH, Weringer EJ, Pollok BA, and Connelly PA. Discovery of a novel, potent, and Src family-selective tyrosine kinase inhibitor. Study of Lck- and FynT-dependent T cell activation. J Biol Chem 271: 695-701, 1996[Abstract/Free Full Text].

17.   Hollenberg, SM, Cunnion RE, and Zimmerberg J. Nitric oxide inhibition reverses arteriolar hyporesponsiveness to catecholamines in septic rats. Am J Physiol Heart Circ Physiol 264: H660-H663, 1993[Abstract/Free Full Text].

18.   Hungerford, JE, Sessa WC, and Segal SS. Vasomotor control in arterioles of the mouse cremaster muscle. FASEB J 14: 197-207, 2000[Abstract/Free Full Text].

19.   Kurjiaka, DT, and Segal SS. Conducted vasodilation elevates flow in arteriole networks of hamster striated muscle. Am J Physiol Heart Circ Physiol 269: H1723-H1728, 1995[Abstract/Free Full Text].

20.   Kumer, SC, Damon DN, and Duling BR. Patterns of conducted vasomotor response in the mouse. Microvasc Res 59: 310-315, 2000[Web of Science][Medline].

21.   Kurata, WE, and Lau AF. p130gag-fps disrupts gap junctional communication and induces phosphorylation of connexin 43 in a manner similar to that of pp60v-src. Oncogene 9: 329-335, 1994[Web of Science][Medline].

22.   Lam, C, Tyml K, Martin C, and Sibbald W. Microvascular perfusion is impaired in a rat model of normotensive sepsis. J Clin Invest 94: 2077-2083, 1994.

23.   Lau, AF, Kurata WE, Kanemitsu MY, Loo LW, Warn-Cramer BJ, Eckhart W, and Lampe PD. Regulation of connexin 43 function by activated tyrosine protein kinases. J Bioenerg Biomembr 28: 359-368, 1996[Web of Science][Medline].

24.   Lidington, D, Ouellette Y, and Tyml K. Endotoxin increases intercellular resistance in microvascular endothelial cells by a tyrosine kinase pathway. J Cell Physiol 185: 117-125, 2000[Web of Science][Medline].

25.   Little, TL, Beyer EC, and Duling BR. Connexin 43 and connexin 40 gap junctional proteins are present in arteriolar smooth muscle and endothelium in vivo. Am J Physiol Heart Circ Physiol 268: H729-H739, 1995[Abstract/Free Full Text].

26.   Loo, LW, Berestecky JM, Kanemitsu MY, and Lau AF. pp60src-mediated phosphorylation of connexin 43, a gap junction protein. J Biol Chem 270: 12751-12761, 1995[Abstract/Free Full Text].

27.   Mehrke, G, and Daut J. The electrical response of cultured guinea-pig coronary endothelial cells to endothelium-dependent vasodilators. J Physiol (Lond) 430: 251-272, 1990[Abstract/Free Full Text].

28.   Ouellette, Y, Lidington D, Naus CG, and Tyml K. A new in vitro model for agonist-induced communication between microvascular endothelial cells. Microvasc Res 60: 222-231, 2000[Web of Science][Medline].

29.   Rivers, RJ. Components of methacholine-initiated conducted vasodilation are unaffected by arteriolar pressure. Am J Physiol Heart Circ Physiol 272: H2895-H2901, 1997[Abstract/Free Full Text].

30.   Rosenkrantz-Weiss, P, Sessa WW, Milstien S, and Kaufman S. Regulation of nitric oxide synthesis by proinflammatory cytokines in human umbilical vein endothelial cells. J Clin Invest 93: 2236-2243, 1994.

31.   Segal, SS, and Duling BR. Conduction of vasomotor responses in arterioles: a role for cell-to-cell coupling? Am J Physiol Heart Circ Physiol 256: H838-H845, 1989[Abstract/Free Full Text].

32.   Steinhousen, M, Endlich K, Nobiling R, Parekh N, and Schutt F. Electrically induced vasomotor responses and their propagation in rat renal vessels in vivo. J Physiol (Lond) 505: 493-501, 1997[Abstract/Free Full Text].

33.   Titheradge, MA. Nitric oxide in septic shock. Biochim Biophys Acta 1411: 437-455, 1999[Medline].

34.   Tyml, K, and Cheng L. Heterogeneity of red blood cell velocity in skeletal muscle decreases with increased flow. Microcirculation 2: 181-193, 1995[Medline].

35.   Tyml, K, Yu J, and McCormack DG. Capillary and arteriolar responses to local vasodilators are impaired in a rat model of sepsis. J Appl Physiol 84: 837-844, 1998[Abstract/Free Full Text].

36.   Welsh, DG, and Segal SS. Endothelial and smooth muscle conduction in arterioles controlling blood flow. Am J Physiol Heart Circ Physiol 274: H178-H186, 1998[Abstract/Free Full Text].

37.   Wilson, JX, Dixon SJ, Yu J, Nees S, and Tyml K. Ascorbate uptake by microvascular endothelial cells of rat skeletal muscle. Microcirculation 3: 211-221, 1996[Medline].

38.   Xia, J, and Duling BR. Electromechanical coupling and the conducted vasomotor response. Am J Physiol Heart Circ Physiol 269: H2022-H2030, 1995[Abstract/Free Full Text].

39.   Xia, J, Little TL, and Duling BR. Cellular pathways of the conducted electrical response in arterioles of hamster cheek pouch in vitro. Am J Physiol Heart Circ Physiol 269: H2031-H2038, 1995[Abstract/Free Full Text].

40.   Yashiro, Y, and Duling BR. Integrated Ca²⁺ signaling between smooth muscle and endothelium of resistance vessels. Circ Res 87: 1048-1054, 2000[Abstract/Free Full Text].