Flow modulates the transport of K⁺ through the walls of single perfused mesenteric venules in anaesthetised rats

Abstract

1. We have investigated the effects of varying flow velocity (U) upon permeability to potassium ions (P_K) of single perfused mesenteric venules in anaesthetised rats. P_K was estimated using a development of the single bolus microperfusion technique at chosen flow velocities in the range of 300 to 6000 μm s⁻¹.

2. In an initial study on 12 vessels, there was a strong positive correlation between P_K and U. This was described by the relation:

   $$P_{\text{K}}+0\cdot 0053U+8\cdot 86,$$

   , where P_K and Uare both expressed in micrometres per second (μm s⁻¹).

3. The addition of the nitric oxide (NO) synthase inhibitors (20 μmol l⁻¹) N^G-monomethyl-L-arginine (L-NMMA) and N^G-nitro L-arginine (L-NNA) to the superfusate abolished the positive correlation between P_K and U. The addition of D-NNA (20 μmol l⁻¹) did not change the relation between P_K and U where the median value for the slope of the relation was 57·7 (± 58·7 interquartile (IQR)) × 10⁻⁴ (n= 4). The addition of L-arginine (200 μmol l⁻¹) restored the relation between P_K and U where the slope of the relation was increased from 3·9 (± 16·3 IQR) × 10⁻⁴ to 69·2 (± 13·5 IQR) × 10⁻⁴ (n= 7).

4. The addition of the guanylate cyclase inhibitor LY83583 (10 μmol l⁻¹) abolished the positive correlation between P_K and U (n= 6).

5. Our data suggest that the flow modulates the potassium permeability through the walls of single perfused rat mesenteric venules via a NO-cGMP-dependent process.

The rate of delivery of molecules to a tissue depends on microvascular blood flow, the surface area of the walls of microvessels, and microvascular permeability. In the conventional analysis of blood tissue exchange (e.g. Renkin, 1959, 1984), it is assumed that the permeability is constant and independent of blood flow, and that transport is regulated by changes in flow and surface area. Recently, however, the interesting possibility has arisen that microvascular permeability might increase with microvascular flow. The evidence for this has been reported in studies carried out on the isolated perfused vessels (Yuan et al. 1992; Pallone et al. 1995) and cultured monolayers of endothelium (Jo et al. 1991; Sill et al. 1995; Waters, 1996). Recently, we have been able to demonstrate rapid and reversible changes in the potassium ion permeability (P_K) following the changes in the flow velocity (U) of the single perfused capillaries of frog mesentery (Kajimura et al. 1998; Kajimura & Michel, 1999). Our development of the single bolus microperfusion technique of Crone et al. (1978) for measuring the P_K allows not only rapid and repeated estimates of the permeability coefficient to K⁺ but also estimates of the transit time of the single bolus from which the flow velocity is accurately estimated. Using this method, we have shown that the flow-dependent component of microvascular permeability to K⁺ is independent of both microvascular pressure and the direction of gradient of [K⁺] across the vessel wall (Kajimura et al. 1998). Furthermore it can be abolished by the agents which raise intracellular levels of cAMP (Kajimura & Michel, 1999).

In this paper, we apply the technique to mammalian microvessels in situ. Our results suggest: (i) that permeability itself increases with blood flow on single mesenteric venules perfused in situ in anaesthetised rats; and (ii) that in these vessels the phenomenon can be abolished by inhibitors of nitric oxide synthase (NOS).

Preliminary reports of our findings have been presented to the Japanese Society for Microcirculation (Kajimura & Michel, 1998a), the Microcirculatory Society (Kajimura & Michel, 1998b) and the British Microcirculation Society (Kajimura & Michel, 1998c).

METHODS

General preparation

The experiments were carried out on single mesenteric venules of young Sprague-Dawley rats (100–150 g, supplied by Harlan, Bicester, UK) under Home Office Regulations. The rats were anaesthetised with an intramuscular injection of 0.1-0.2 ml of a mixture of Hypnorm (0.315 mg ml⁻¹ fentanyl citrate and 10 mg ml⁻¹ fluanisone; Janssen Pharmaceuticals Ltd, High Wycombe, UK), Hypnovel (5 mg ml⁻¹ midazolam hydrochloride; Roche Products Ltd, Welwyn Garden City, UK) and water made up in the proportions 1:1:2 by volume. Further intramuscular injections of this mixture were given before opening the abdominal cavity and during the experiment when the rat's reflex response to pinching the hind paws (which was tested between each sequence of measurements) suggested that it was necessary. The abdominal cavity was opened by a small incision to the left of rectus abdominis and a loop of intestine was carefully delivered through it. The mesentery was gently arranged over the surface of a polished Perspex pillar. This allowed transillumination of the mesenteric microvasculature. The upper surface of the mesentery was superfused continuously with Ringer solution at 37°C. The flow of superfusate was maintained at 3.5-4.0 ml min⁻¹ and this kept the layer of fluid over the tissue at approximately constant depth. The microvessels chosen for study were mostly second order postcapillary venules (diameters 18–35 μm). The tissue was observed through a stereomicroscope (Wild Heerbrugg M8) with a CCTV camera (Hitachi) attached to the camera tube. The output from the camera was displayed on videomonitors and recorded. At the end of the experiments, the animal was killed by an overdose of anaesthetic.

Solutions and reagents

Ringer-Locke solution was used as the bathing solution for the dissection of the mesentery and the superfusates. Its composition was (in mmol l⁻¹): 130 NaCl, 4.6 KCl, 1.2 MgCl₂, 2.5 CaCl₂, 0.195 NaHCO₃, 5.5 glucose, buffered with 1.79 Na₂HPO₄, 0.21 NaH₂PO₄, 2.3 Hepes acid and 2.7 Hepes sodium salt (Sigma Chemicals). The pH was adjusted to 7.4. The perfusate contained bovine serum albumin (BSA: A-7638, Fraction V, Sigma) at a concentration of 50 mg ml⁻¹. Evans Blue (5 mmol l⁻¹) was added to colour the solution containing BSA (50 mg ml⁻¹). At this concentration 98 % of the dye should be bound to the BSA (Levick & Michel, 1973). Evans Blue perfusates were dialysed in 8000 molecular weight cut-off dialysis tubing (Spectro/Por Spectrum, CA, USA) against three 2 l changes of Ringer solution of equal osmolarity over a 24 h period at 15°C. High-K⁺ solutions (30 mmol l⁻¹ K⁺) were prepared by replacing 25.4 mmol NaCl l⁻¹ with equimolar KCl.

Reagents were prepared as concentrated stock solutions with appropriate vehicles. These stocks were kept at −20°C and used within 2 weeks.

LY83583 (CalBiochem, Beeston, Nottingham, UK) was initially dissolved in 95 % ethanol at a concentration of 10 mmol l⁻¹ as a concentrated stock. On the day of the experiment the first dilution, one-tenth dilution, was made into BSA-free Ringer solution. Then the second dilution was made using BSA-Ringer solution to bring the final concentration of LY83583 to 10 μmol l⁻¹. The final ethanol concentration of the test perfusate was 0.1 % v/v (17 mmol l⁻¹).

L-NMMA, L-NNA, D-NNA and noradrenaline (L-(−)-norepinephrine-(+)-bitartrate) were obtained from CalBiochem and Ringer solutions were used as a vehicle for these reagents.

Fabrication of a K⁺ ion-sensitive electrode

The electrodes were made according to the method described by Voipio et al. (1994). Single-barrelled pipettes (quartz with filament, o.d., 1.2 mm; i.d., 0.60 mm, Sutter Instrument Co., Navato, CA, USA) were pulled on a micropipette puller (model P-2000, Sutter Instrument Co.). The micropipettes were mounted horizontally on a brass holder, placed in a Petri dish, and baked at 200°C. After 30 min approximately 50 μl of N,N-dimethyltrimethylsilylamine (Fluka Chemicals, Gillingham, UK) were added to the Petri dish. Baking continued for a further 1 h. This silanization process made the glass surface hydrophobic and ensured good contact between the glass and the lipophilic ion exchanger. The micropipettes were then back-filled with a small amount of a liquid ion exchanger (potassium ionophore I-Cocktail A: Fluka Chemicals) and filled with an electrolyte solution (0.5 mol l⁻¹ KCl).

Double-barrelled perfusion pipettes

Double-barrelled microperfusion pipettes (Davis & Gore, 1987; McKay & Huxley, 1995) were made from θ tubing (o.d., 1.5 mm; Clark Electromedical Instruments, Reading, UK). Two small holes were made on one side of one barrel with a 400 diamond disc (Rx Honing Machine, Mishawaka, IN, USA) at distances of 6 and 14 mm from what was to become the open end of the pipette. After being ground, the glass tubes were cleaned with detergent (5 % Decon90, Decon Laboratories Ltd, Sussex, UK) and rinsed in distilled water, methanol and acetone. The hole closest to the open end of the tube and the open end itself of that barrel were filled with a plug of melted wax. This meant that the lumen of that barrel communicated with the outside through the hole, which was 14 mm from the open end. The pipettes were pulled and then bevelled with a micro-grinder (model EG-4, Narishige) to produce the tip diameter of 18–28 μm. The double-barrelled microperfusion pipette was placed in a holder, which had been manufactured so that the open end of one barrel was separated from the opening in the side of the second barrel by a silicone ring which acted as a water-tight seal between two interior compartments. As pressure could be applied to each of these compartments independently, the ejection or retention of the solution in each barrel could be controlled separately.

Calibration of the K⁺-sensitive microelectrodes

All measurements with K⁺-sensitive microelectrodes were carried out in a Faraday cage. Chlorided silver wires were used to connect the recording apparatus to either the calibration chamber or the preparation. Signals from K⁺-sensitive microelectrodes were amplified using a multipurpose microelectrode amplifier with headstage (HS-2 gain 0.0001M) input resistance above 10¹⁴Ω (Axoprobe-1A, Axon Instruments, Foster City, CA, USA). The output from the electrometer was fed through an oscilloscope (Tektronix 5A22 N differential amplifier, Harpenden, UK) into an AD converter (1401, Cambridge Electronic Design, Cambridge, UK) and into a Pentium 90 computer. Microelectrodes were calibrated before and after the experiments by solutions containing 4.6 and 30 mmol l⁻¹ K⁺. The signal measured in the solution containing 4.6 mmol l⁻¹ K⁺ was used as baseline, i.e. the voltage was set at zero. On some occasions concentrations of 4.6, 7, 10, 15, 20 and 30 mmol l⁻¹ K⁺ were used to check the linearity of the calibration curve. The time constants of the electrodes were usually less than 50 ms. Electrodes with a time constant greater than 200 ms were rejected. The response time of the recording system itself was less than 25 ms. The electrode typically gave a 46–47 mV change from 4.6 to 30 mmol l⁻¹ change in K⁺ concentration. The response time and calibration of the microelectrodes were estimated at high and low flows through the calibration cell. Both calibration and response time were independent of flow velocity over the range 800–4000 μm s⁻¹.

General protocol

A detailed description of the method used to measure P_K and U in a single perfused microvessel has been published previously (Kajimura et al. 1998). Briefly, each venule was cannulated with a bevelled double-barrelled micropipette made out of θ tubing. One barrel of the pipette was filled with a normal K⁺ solution (4.6 mmol l⁻¹ K⁺) and the other was filled with a high-K⁺ solution (30 mmol l⁻¹ K⁺). The tubing leading from the two barrels of the pipette was connected through an electric rotary valve (Omnifit Ltd, Cambridge, UK) to the two water manometers. This arrangement allowed alternate perfusion with the normal K⁺ solution or the high-K⁺ solution. The heights of the water columns of the two manometers were adjusted so that when the normal K⁺ solution was being perfused, the high-K⁺ solution was not and vice versa. To do this, one solution (the normal K⁺ solution) was coloured with Evans Blue (5 mmol l⁻¹), therefore making the interface between the normal and high-K⁺ solutions visible. The interface between the two solutions at the tip of the perfusion pipette was carefully monitored to prevent either the normal K⁺ solution from entering the other barrel or the high-K⁺ solution from perfusing the vessel.

After the interface was adjusted, the electric rotary valve, which functioned as a cross-over tap between two manometers, was switched so that the higher pressure was applied to the high-K⁺ solution causing it to flow through the microvessel. After 2 s, the rotary valve was returned to its initial position. The intraluminal [K⁺] was monitored by two K⁺-sensitive microelectrodes. The two microelectrodes, designated as e₁ and e₂, respectively, were located downstream from the perfusion pipette at points 280–1070 μm apart. The more proximal microelectrode, e₁, was at least 300 μm downstream from the cannulation site. Potassium indicator potentials were acquired at the rate of 200 Hz using Chart software (Cambridge Electronic Design) running on a Pentium 90 computer.

An interval between each measurement of no less than 40 s was allowed to ensure adequate washout of K⁺ from the interstitium surrounding the vessel. The superfusion rate was kept high (3.5-4 ml min⁻¹) to clear K⁺ effectively from the mesothelial surface.

The perfusion velocity, U, was calculated from the distance between the two K⁺-sensitive electrodes, e₁ and e₂, and the transit time, τ, of the high-K⁺ bolus between them. The leading edge of the high-K⁺ bolus was detected as a sharp rise in [K⁺] as it passed the intraluminal tip of each electrode, the response of e₂ being delayed after that of e₁ by an interval equal to τ (see Fig. 2).

Figure 2. Time-concentration curves with ‘high’ and ‘low’ flow.

Changes in [K⁺] recorded from two K⁺-sensitive microelectrodes separated by a distance of 770 μm, as a bolus of high [K⁺] flowed down the vessel. A, under conditions of high flow (U= 2850 μm s⁻¹) when P_K was calculated as 13.6 μm s⁻¹. B, low flow (U= 1570 μm s⁻¹) in the same microvessel when P_K was calculated as 8.2 μm s⁻¹. Flow velocities calculated from the transit time, τ, between the electrodes. Vessel radius = 14.3 μm; distance from perfusion pipette to e₁= 605 μm.

Increases and decreases in U were achieved by raising and lowering the pressure applied to the perfusion pipette. Every change in perfusion pressure involved adjustment to both manometers so that the colourless (high-K⁺) perfusate filled its barrel of the micropipette down to the tip when the vessel was being perfused with normal (Evans Blue-containing) Ringer solution. In most experiments flow was increased in a series of steps and then lowered so that measurements of P_K were made at low perfusion velocities at the beginning and end of each sequence and measurements at high velocity in the middle. In some experiments measurements of P_K at high and low U were alternated.

Calculation of diffusional potassium permeability (P_K)

Permeability was estimated using a development of the method of Crone et al. (1978). Briefly, a bolus of high-K⁺ solution flows along a single microvessel and the [K⁺] is recorded at two points by K⁺-sensitive microelectrodes (e₁ and e₂) separated by a length of the vessel over which permeability is to be determined. If C₁ and C₂ are [K⁺] at e₁ and e₂, respectively, as the bolus flows past the electrodes and C_E1 and C_E2 are the [K⁺] values in the pericapillary fluid at these points then:

(1)

where R is the radius of the microvessel and τ is the transit time of the bolus between the two electrodes.

Crone et al. (1978) assumed that the pericapillary [K⁺] was equal to the superfusate [K⁺] and did not change significantly as the bolus swept along the vessel. We have shown, however, that this is not so (Kajimura et al. 1998). In frog mesenteric capillaries and venules, the mean [K⁺] in the pericapillary space, estimated over any time during the passage of a high [K⁺] bolus, was directly proportional to the mean [K⁺] within the microvessel over the same period.

Thus at e₁:

(2)

where C₀ is the initial value of [K⁺] in the perfusate and superfusate and α is a constant which in frog mesentery has a value of approximately 0.5 (Kajimura et al. 1998). If the relation in eqn (2) is used to describe the difference in potassium concentration across the vessel wall, and the fluxes of potassium at every point between e₁ and e₂ are integrated, eqn (1) can be rewritten in terms of C₁, C₂, C₀, r, α and τ, all of which can be measured, i.e.:

(3)

(Kajimura et al. 1998).

Since α has only been estimated in microvessels of the frog mesentery, the determination of α was the first step in the measurement of P_K in rat mesenteric venules.

To determine the value of α, one K⁺-sensitive electrode was placed beneath the mesothelium just outside the microvessel and directly opposite the other electrode, which was placed inside the microvessel lumen. In this way changes in [K⁺] in the microvessel (ΔC) and those in the pericapillary fluid (ΔC_E) during brief perfusions of K⁺-rich solutions could be monitored simultaneously at different flow velocities.

Statistical analysis

Average values are reported as the median ± interquartile (IQR) except when stated otherwise. Linear regression was used to fit lines to the data for P_K and U. Analysis of covariance (Weatherburn, 1946) was used to examine the relations between P_K and U under control conditions and in the presence of L-NNA. To compare and contrast the average values of slopes and intercepts between other groups of experiments and control data, the Wilcoxon signed rank test (paired comparison) and Mann-Whitney U test (unpaired comparison) were used. To compare slopes and intercepts within a single experiment, Student's t test was used. Level of significance was set at < 5 %.

RESULTS

Determination of α

Figure 1 shows the changes in [K⁺] recorded by two K⁺-sensitive microelectrodes, one inside and the other immediately outside a venule as a bolus of high [K⁺] flowed down the vessel. The rapid rise in [K⁺] inside the vessel is followed by a smaller but still substantial rise in the [K⁺] of the perivascular fluid. The mean increment in the perivascular [K⁺], ΔC_E, over the 2 s after the arrival of the K⁺-rich bolus at the intravascular electrode, was approximately half the mean value of the increment in [K⁺] of the perfusate, ΔC. A similar relation between ΔC_E and ΔC was found in 35 measurements of this kind carried out in three different vessels which were similar in type and size (radii of 14, 14 and 15 μm) to those used in the subsequent experiments of this investigation. In each vessel, determinations were made over a wide range of flow velocities (approximately 1000–3000 μm s⁻¹). These results are summarised in Fig. 1B and the pooled data can be described by the expression:

(4)

where the error qualifying the regression coefficient is the standard error. The intercept of the relation between ΔC_E and ΔC does not differ significantly from zero (t test). From these data it is reasonable to conclude that the value of α is approximately constant in rat mesenteric venules over the range of flows investigated and that its value can be taken as 0.48.

Figure 1. Changes in [K⁺] inside and just outside a microvessel during a 2 s perfusion of a bolus of high [K⁺].

A, recordings of [K⁺] within (in) and outside (out) the venule. B, plot of the mean increment of [K⁺] inside the microvessel lumen (ΔC) against the mean increment of [K⁺] in the pericapillary fluid (ΔC_E) during the passage of a 2 s bolus of a high [K⁺] solution. Repeated measurements were made on three capillaries at different flow velocities in the range 1000–3000 μm s⁻¹; the different symbols represent different microvessels (n= 35, r= 0.87, P < 0.0001).

Effect of varying flow on P_K

Figure 2 shows the changes in [K⁺] recorded from two K⁺-sensitive microelectrodes placed in a single microvessel, 770 μm apart, as a bolus initially containing 30 mmol l⁻¹ K⁺ flowed along the vessel. The recordings shown in Fig. 2A and B were made when flow velocities were 2850 and 1570 μm s⁻¹, respectively. The signal from e₂ was seen to rise later than that from e₁, the delay representing the transit time (τ) between the electrodes. The difference in [K⁺] recorded at e₁ and e₂ showed that a significant amount of K⁺ left the vessel between these points. When P_K was calculated from the data in Fig. 2A and B it was found that the recordings in Fig. 2A yielded a value of 13.6 μm s⁻¹ whereas that in Fig. 2B gave a value of 8.2 μm s⁻¹. Thus P_K estimated from the data in Fig. 2B was only 60 % of its value when estimated from the data in Fig. 2A where the vessel was perfused at almost twice the flow rate. Further estimates of P_K on this same vessel over a range of flows revealed a clear correlation between P_K and U. These data are shown in Fig. 3A, where the correlation coefficient, r, has a value of 0.88 (n= 14, P < 0.001).

Figure 3. The relationship between P_K and flow velocity: a single experiment on a venous microvessel.

A, determinations of P_K plotted against corresponding values of flow velocity (U) made on a single venous microvessel. The line is the regression line, P_K= 0.0042 U+ 2.14 (n= 14, r= 0.88, P < 0.001). B, relations between the pressure applied to the perfusing barrel (P_p, ○) and the non-perfusing barrel of the double-barrelled micropipette (P_e, •); * indicates where two points overlie one another.

Figure 3B shows the values for the pressures applied to the perfusion micropipette in order to achieve the range of flows which were investigated. P_e is the pressure applied to the non-perfusing barrel of the pipette and since it is sufficient to prevent fluid from entering or leaving that barrel, its value should equal that in the vessel at the site of cannulation. Thus P_e is a good estimate of the maximum pressure within the vessel during the perfusion. The difference between P_p and P_e represents the pressure head necessary for the different values of U to be achieved through the perfusion side of the micropipette. At the tip of the micropipette inside the microvessel, the pressure in the fluid leaving the perfusion barrel should be approximately the same as that in the non-perfusing side, the kinetic energy of the fluid being small in absolute terms. If regression lines were drawn through the data for P_p and P_e in Fig. 3B, the two lines would intersect close to the pressure axis (U= 0) at a value between 16 and 17 cmH₂O. From this it can be seen that increments in perfusion pressure within the vessel of 1 cmH₂O (as indicated by the changes in P_e) increase U by approximately 1000 μm s⁻¹. This would be consistent with Poiseille flow down a smooth tube with a radius equal to that of the venule (14.3 μm) and 2.5 mm in length.

A clear positive correlation between P_K and U was found in every one of 12 vessels from 12 different animals investigated in this way. The results from all these experiments are shown in Fig. 4 with additional details also given under ‘Control’ in Table 1. It can be seen that there was considerable variation in the slopes and the intercepts of the relations between P_K and U. In every case, however, r was greater than 0.78 and P was less than 0.01. When the data from all 12 experiments were pooled (n= 150), the correlation was preserved (r= 0.532, P < 0.0001). The regression line through the pooled data was described by the relation:

(5)

Analysis of covariance revealed that the correlation between P_K and U within individual experiments was highly significant (P < 0.001) and that there was highly significant variation in the regression coefficients between experiments (P < 0.001).

Figure 4. The relations between P_K and U in 12 rat venules.

Each diagram (a–l) shows the relations between P_K and U in a single mesenteric venule. Each venule was taken from a different rat. In all experiments there is a significant positive correlation between P_K and U. The values for the correlation coefficients (r) and the number of measurements in each experiment were: a, r= 0.85, n= 8; b, r= 0.91, n= 8; c, r= 0.91, n= 15; d, r= 0.87, n= 14; e, r= 0.90, n= 18; f, r= 0.98, n= 16; g, r= 0.91, n= 9; h, r= 0.78, n= 13; i, r= 0.90, n= 8; j, r= 0.95, n= 17; k, r= 0.86, n= 12; l, r= 0.93, n= 14.

Table 1.

Values for the slopes of the relation between P_K and U and other data for venules under control conditions and in the presence of L-NNA

			Range of U			Range of Pe
	SlopePK/ U× 10−4	Radii(μm)	Max (μm)	Min (μm)	Separation(μm)	Max (cmH2O)	Min (cmH2O)
Control
1	113	13	2625	1292	420	24.6	20.9
2	27	14	4667	1458	700	22.0	14.8
3	73	12	3909	956	430	19.0	13.0
4	42	14	3080	875	770	21.6	17.6
5	23	14.5	3308	1024	860	21.6	14.2
6	100	13	5667	680	510	24.3	19.6
7	42	9.5	5667	3579	680	20.1	16.4
8	38	15	3182	972	700	22.8	17.1
9	67	16	2853	1516	970	25.3	17.4
10	65	17.5	2680	493	670	19.5	15.7
11	40	12.5	2381	446	500	16.0	13.5
12	94	16	2154	308	280	28.2	21.9
L-NNA
1	17.3	13	5467	2000	820	25.1	17.1
2	5.8	14	5176	1419	880	26.0	21.6
3	−13.5	17.5	2080	634	520	28.0	20.0
4	3.8	10	4458	1081	1070	24.4	18.6
5	−23.1	16	2591	671	570	35.7	26.4
6	3.8	10	5182	1018	570	27.5	20.2
7	1.9	13	4400	1048	440	22.8	17.9
8	19.2	13	2474	1424	470	33.6	27.1
9	−9.6	13	3273	482	720	24.4	19.6
10	40.4	16	2600	419	420	14.7	12.6

The relation between P_K and U was independent of the order in which the points were determined. Increases in P_K with increases in U appeared to be complete within 20 s, as judged by repeated estimates of P_K at the same value of U. The changes were also reversed with similar rapidity. Further details of these experiments are given in Table 1.

Effect of NO synthase inhibitors on the relation between P_K and U

To determine whether the mechanism of flow-dependent changes in P_K involved release of nitric oxide (NO) within the endothelium or surrounding tissue, we examined the effects of the NO synthase (NOS) inhibitors L-NMMA and l-NNA, on the relation between P_K and U. The addition of either NOS inhibitor (20 μmol l⁻¹) significantly reduced the slope of the relation (2 vessels with L-NMMA and 11 vessels with L-NNA). Figure 5 shows the results of an experiment on a single venule in which P_K was determined over a range of U, first under control conditions and then after L-NNA (20 μmol l⁻¹) had been present in the superfusate for 20 min. Under the control conditions P_K strongly correlated with U (n= 8, r= 0.90, P < 0.01). The vessel was then superfused with Ringer solution containing L-NNA (20 μmol l⁻¹) for 20 min and P_K was re-measured over a similar range of U. The addition of L-NNA reduced the correlation coefficient between P_K and U to −0.27 which was no longer significant at 5 % value. The two slopes were significantly different (P < 0.01, t test). Very similar results to those shown in Figure 5 were obtained in two venules where the relation between P_K and U was determined first under control conditions and then after superfusion with L-NMMA (20 μmol l⁻¹).

Figure 5. Effect of L-NNA on the relation between P_K and U.

Paired measurements of P_K in a single microvessel before and after the addition of L-NNA to the superfusate are shown as a function of U. Eight determinations of P_K were made under the control conditions (○). P_K strongly correlated with U (continuous line, r= 0.90, P < 0.01). The same microvessel was then superfused with L-NNA (20 μmol l⁻¹) for 20 min and another nine determinations of P_K were made (•). The addition of L-NNA abolished the correlation between P_K and U (dashed line, r=−0.27, P > 0.05). The two slopes were significantly different (P < 0.01, t test).

In a further 10 experiments in 10 different rats, measurements of P_K were made over a range of values of U in the presence of L-NNA (20 μmol l⁻¹). In five of these vessels L-NNA was added to both the perfusate and the superfusate. In the other five vessels, L-NNA was added to the superfusate alone. The results from these 10 experiments are shown in Fig. 6.

Figure 6. The relations between P_K and U in the presence of L-NNA.

Each diagram (a–j) shows the relations between P_K and U in a single venule from a different rat when the superfusate contained L-NNA at a concentration of 20 μmol l⁻¹. In only two experiments was there a significant positive correlation between P_K and U. The values for the correlation coefficients and the number of measurements in each experiment were: a, r= 0.80, n= 14; b, r= 0.44, n= 11; c, r=−0.46, n= 12; d, r= 0.23, n= 15; e, r=−0.60, n= 12; f, r= 0.25, n= 12; g, r= 0.10, n= 10; h, r= 0.38, n= 8; i, r=−0.43, n= 10; j, r= 0.91, n= 8.

It can be seen that in the presence of L-NNA a significant positive correlation between P_K and U was present in only 2 of the 10 experiments. In the remaining 8 experiments, the correlation was either not significant or negative. When all the data shown in Fig. 6 were pooled, however, a significant correlation between P_K and U was found (r= 0.45, P < 0.001, n= 112). The regression line through these data was described by the expression:

(6)

Analysis of covariance indicated that there was significant variation of P_K with U within individual experiments (P < 0.05) but the regression coefficient in the presence of L-NNA differed from its value under control conditions to a highly significant degree (P < 0.0001). Further details of the 10 experiments shown in Fig. 6 are summarised in Table 1. It is worth adding that no change of vessel diameter was observed during the application of the NOS inhibitors.

Further control experiments were carried out on an additional four vessels where the relations between P_K and U were investigated after D-NNA, an inert isomer of L-NNA, had been added to the superfusate in the same concentration as the L-NNA (20 μmol l⁻¹). In these experiments, the positive correlation between P_K and U was maintained with the average slope of the relation for the four vessels being 57.7 (± 58.7) × 10⁻⁴. This was not different from the value of the slope in the control group of vessels (Mann-Whitney U test).

L-Arginine restores the positive correlation between P_K and U in the presence of L-NNA

In further experiments on seven vessels the relation between P_K and U was determined first after 20 min exposure to L-NNA and then remeasured in the presence of L-NNA and L-arginine (L-Arg), a precursor of NO. Figure 7A shows an example of an experiment of this kind. During the addition of L-NNA (20 μmol l⁻¹) to the superfusate, there was no correlation between P_K and U (n= 10, r=−0.43). The vessel was then superfused with Ringer solution containing L-Arg (200 μmol l⁻¹) in addition to L-NNA (20 μmol l⁻¹) for 22 min. The strong positive correlation between P_K and U was restored (n= 10, r= 0.91). The two slopes were significantly different (P < 0.001, t test). In each of seven vessels investigated, the addition of L-Arg restored the positive correlation between P_K and U where median values of the slope of the relation were significantly increased from 3.9 (± 16.3) × 10⁻⁴ to 69.2 (± 13.5) × 10⁻⁴ (n= 7, P < 0.01, Wilcoxon signed rank test) (Fig. 7B).

Figure 7. Inhibition of the flow-dependent component of P_K by L-NNA can be reversed by L-arginine.

A, results from a single venule showing there was no correlation between P_K and U in the presence of L-NNA (20 μmol l⁻¹) in the superfusate (•, n= 10, r=−0.43). The vessel was then superfused with Ringer solution containing L-Arg (200 μmol l⁻¹) in addition to L-NNA (20 μmol l⁻¹) for 22 min. The strong positive correlation between P_K and U was restored (□, n= 10, r= 0.91, P < 0.001). The two slopes were significantly different (P < 0.001, t test). B, median values of the slope of the relation between P_K and U are plotted for experiments conducted in the presence of L-NNA and in the presence of L-NNA and L-Arg. The addition of L-Arg restored the positive correlation between P_K and U (n= 7, P < 0.01, Wilcoxon signed rank test).

Effect of the guanylate cyclase inhibitor LY83583 on the relation between P_K and U

To test if the mechanism to induce the flow-dependent K⁺ permeability involves activation of guanylate cyclase and production of cGMP, the relation between P_K and U was determined after vessels had been perfused with LY83583 (10 μmol l⁻¹) for 20–25 min. Figure 8A shows one experiment of this kind. In each of the six experiments, there was no significant correlation between P_K and U. The median value for the slope between P_K and U was 7.7 (± 32.7) × 10⁻⁴ which was significantly less than the slope of control vessels summarised in Table 1 (Mann-Whitney U test, P < 0.01) (see Fig. 8B).

Figure 8. Effect of LY83583 on the relation between P_K and U.

A, the points represent estimates of P_K made at various values of U on a single mesenteric venule when LY83583 (10 μmol l⁻¹) had been added to the perfusate. The continuous line represents the regression through the points and it can be seen that there is no correlation between P_K and U (⋄, n= 10, r=−0.12, P > 0.05). For comparison the dashed line represents the median relation seen in control vessels. B, median values of the slope of the relation between P_K and U are plotted for the control group and for six vessels perfused with LY83583 (n= 6, P < 0.01, Mann-Whitney U test).

In all experiments where the vessels were treated with LY83583, the perfusate contained ethanol at a concentration of 17 mmol l⁻¹. To test whether the presence of ethanol itself influenced the relation, additional control experiments were carried out where the perfusate contained ethanol at a concentration of 17 mmol l⁻¹ in the absence of LY83583. In each of the three microvessels, a clear positive correlation between P_K and U was found and the median value of the slopes was 63.0 (±85.5) × 10⁻⁴ which was not different from that determined in the absence of ethanol as indicated by the control vessels summarised in Table 1 (n= 3, Mann-Whitney U test).

The effects of various agents upon the flow-dependent component of P_K, as measured by the slope of the relation between P_K and U, are summarised in Fig. 9.

Figure 9. Summary of effects of the agents on the flow dependency of P_K in rat mesenteric microvessels.

Median values for the slopes of the relations between P_K and U are given as a measure of the flow dependency of P_K. □, median slope determined from measurements in 12 vessels under control conditions; ▪, pooled data for NOS inhibitors (20 μmol l⁻¹, n= 13); , L-Arg (200 μmol l⁻¹) in the presence of L-NNA (20 μmol l⁻¹, n= 7); , LY83583 (10 μmol l⁻¹, n= 6). * P < 0.01 compared with the control (Mann-Whitney U test).

DISCUSSION

The two principal findings in this paper are: first that, as in frog microvessels, the permeability of rat mesenteric venules to K⁺ varies with the rate of perfusion of the vessel; and second, that the flow dependence of P_K in rat mesenteric venules (unlike this phenomenon in frog microvessels) is lost when the tissues are treated with inhibitors of NO synthase and guanylate cyclase.

Flow-dependent permeability

The increase in P_K with U which we have found in rat mesenteric venules appears to be very similar in magnitude to the relation which we reported for the flow dependence of P_K in frog mesenteric microvessels. The value of α, the factor relating changes in the perivascular [K⁺] to the [K⁺] in the vessel lumen, is almost identical in frog and rat mesenteric microvessels. Since α is determined by the permeability of the microvascular wall and the diffusion coefficient for K⁺ in the perivascular interstitium the similar values of α reflect the similar values of P_K and the similar ultrastructure of the perivascular interstitium for the two vessel types.

The correlation between P_K with U in frog and rat mesenteric microvessels strengthens the view that flow dependence of permeability is a general phenomenon. Increasing the shear stress at the free (luminal) surface of monolayers of cultured endothelial cells has been reported to increase their permeability to fluid and to a range of molecules (e.g. Jo et al. 1991; Sill et al. 1995; Waters, 1996). While these observations are open to the criticism that cultured endothelial cells do not adhere to the underlying surface as well as endothelia do to their basement membranes in situ, increased permeability in response to increased luminal shear stress has also been described to occur in two very different types of isolated perfused microvessel. Thus Yuan et al. (1992) have reported increased permeability to serum albumin in isolated perfused venules from the pig heart in response to increases in the perfusion rate of these vessels. Flow-dependent increases in permeability to Na⁺ and raffinose have also been described in isolated perfused descending vasa recta from rat kidney by Pallone et al. (1995) and by Turner & Pallone (1997).

At present it is not clear whether increased flow velocity increases the permeability of microvessels to all hydrophilic molecules or whether increased flow increases the permeability of some vessels to a restricted group of solutes. Experiments on perfused hearts and skeletal muscles of dogs indicate that whereas the transport of molecules larger than glucose reaches a maximum at high rates of perfusion, the transport of smaller molecules may not do so (e.g. Renkin, 1984; Watson, 1995). While the conventional interpretation of such observations is that the permeabilities to smaller molecules are so high in these capillary beds that their transport is always flow limited, the findings are also consistent with permeabilities to the smaller molecules being flow dependent whereas those to the larger molecules are not. If the permeability of muscle capillaries to glucose increased with flow, it could contribute significantly to the very high rates of glucose uptake into skeletal muscle in severe exercise (Chapler & Stainsby, 1968).

One striking feature of our results from both frog and rat microvessels is the speed at which P_K is changed to a new value following a change of flow velocity. Although in most of our experiments an interval of 40 s intervened between successive measurements to ensure that K⁺ was adequately washed from the perivascular interstitium, some measurements of P_K were made 20 s after flow had been changed. It was found in these experiments that the change in P_K was complete at this time. These rapid changes in permeability with flow appear to be rather different from those seen when permeability is increased by mediators of acute inflammation. Whereas some of the mediators increase venular permeability rapidly, the increase is usually short-lived in spite of the continuing presence of the mediator (e.g. Korthuis et al. 1982; Michel & Kendall, 1997). We have found no evidence so far for attenuation of flow-induced increases in P_K.

Since changes in U were achieved by changing the perfusion pressure, it might be imagined that some (at least) of the change in P_K calculated from eqn (3) might be the consequence of a change of R, the vessel radius, rather than the result of a true change of permeability. Quite apart from our failure to detect any change in vessel diameter with changes in U, a consideration of the distensibility properties of mesenteric vessels indicates that the influence of changes in R upon the calculation of P_K would be very small. In Table 1, it is seen that the largest change in P_e in any one experiment was from 17.4 to 25.3 cmH₂O, and the mean change was from 16.8 to 22.0 cmH₂O. Compliance curves of rat mesenteric venules reported by Swayne et al. (1989) indicate that the radius of a vessel would increase from 14.1 to 14.22 μm when the pressure is increased from 16.8 to 22 cmH₂O, i.e. by 0.85 %. By contrast P_K increased by 100 % in most experiments shown in Fig. 4. Thus it seems safe to regard the effects of distensibility upon the estimates of P_K reported in this paper as negligible.

Nitric oxide as an intracellular signal in flow-dependent permeability

Since the positive correlation between P_K and U can be abolished in rat mesenteric vessels by superfusing the tissue with the NO synthase inhibitors L-NNA or L-NMMA, it appears that a critical step between a change of flow through the vessel and a change in permeability is the release of NO from L-arginine by the action of NO synthase. This conclusion is strengthened by the various control experiments and by our demonstration that LY83583, which is an inhibitor of guanylate cyclase, the target for the NO signal, also abolishes flow-dependent changes of P_K.

These findings are entirely consistent with those of Yuan et al. (1992) who reported that flow-induced increases in permeability of isolated pig venules to albumin could be abolished by L-NMMA. Our findings in rat mesenteric venules contrast with our earlier results from frog mesenteric microvessels. Here we found that the flow-dependent component of P_K was unaffected by L-NMMA. It is interesting that we have found essentially the same dependence of P_K on U in mesenteric microvessels of frogs and rats yet while the mechanism appears to involve NO signalling in rats, that in frogs does not (Kajimura & Michel, 1999). The difference between the two species is not because frog endothelium does not use NO as an intracellular signal. He et al. (1997) have shown that agonist-induced increases in the permeability of frog microvessels, like those in mammalian microvessels, can be blocked by NO synthase inhibitors. Furthermore, Rumbaut et al. (1995) reported that NOS inhibitors lowered the baseline permeability of some frog microvessels. It would appear that different signalling mechanisms have evolved for flow-dependent increases in permeability in frogs and at least two mammalian species (rat and pig).

Possible mechanisms

While the activation of guanylate cyclase by NO appears to be an essential step between the changes in flow or shear stress at the endothelial surface and the change in P_K, neither the mechanisms which transduce the stimuli to the endothelium nor those which increase permeability are understood. Hypotheses for the transduction process have been proposed in relation to flow-induced vasodilatation and vascular remodelling. Thus(Davies, 1995, for review) has suggested that shear at the endothelial luminal surface is transmitted by the cytoskeleton to the abluminal surface of the cell where mechanotransduction occurs at integrins which hold the cell to focal adhesion sites on its basement membrane. Consistent with this hypothesis is the report of Muller et al. (1997) who have found that in isolated arterioles from pig heart, flow-induced vasodilatation can be inhibited by treatment of the tissue with a peptide containing an RGD sequence. It is known that peptides of this type compete for integrin-binding sites such as the focal adhesion sites, and inhibition of flow-induced vasodilatation by the peptide is competitive.

It has also been suggested that shear is sensed at the luminal endothelial surface in caveolae (Schnitzer, 1997). Working on the isolated perfused rat lung, Rizzo et al. (1998) have found that increased perfusion rates activate caveolar endothelial NOS, inducing it to dissociate from caveolin and associate with calmodulin. The membranes of caveolae are rich in cholesterol and by manipulating the cholesterol content of the membranes in cultured endothelial cells, it has been found that shear stimuli activate the extracellular signal-regulated kinase (ERK) of the mitogen-activated protein kinase (MAP kinase) family (Park et al. 1998). An earlier report also suggested that the shear stress is sensed at the luminal surface of endothelium. Hecker et al. (1993) found that removal of the sialic acid groups from the endothelial glycocalyx by neuraminidase selectively inhibits shear-stress-dependent release of NO in isolated rabbit femoral arteries. It is surprising that this observation has not been pursued further.

The link between the activation of guanylate cyclase and an increase in permeability is also obscure. It is known that increases in microvascular permeability can be inhibited by raising the levels of cAMP in endothelial cells and one proposal (He et al. 1997) for the effects of a NO-cGMP-dependent process on permeability is that cGMP might promote the breakdown of cAMP by activating a cGMP-dependent phosphodiesterase (Ogawa et al. 1992; Koga et al. 1995). Alternatively it has been suggested that raised levels of cGMP might increase permeability by activating protein kinase G (PKG). In support of this proposal is the finding that inhibition of PKG prevents both histamine and vascular endothelial growth factor (VEGF) from increasing the permeability to albumin of pig coronary venules where both these mediators normally increase permeability through a NO-cGMP-dependent mechanism (Wu et al. 1997; Huang & Yuan, 1997).

Conclusion

The permeability of rat mesenteric venules to K⁺ increases linearly with the rate at which they are perfused in situ. The flow-dependent component of permeability can be inhibited reversibly by inhibitors of NO synthase and this inhibition can be reversed by addition of L-arginine to the superfusate. At present, both the mechanisms of flow or shear detection by the endothelium and those responsible for increasing the permeability are obscure.