We have previously reported an inhibitory effect of dietary copper deficiency on in vivo platelet thrombus formation (Schuschke et al. 1989, 1994 and 1995). By using a light-dye reaction to induce thrombus formation, we demonstrated a significant delay in thrombogenesis and the time to complete venular occlusion in the cremaster muscle microcirculation of copper-deficient rats (Schuschke et al. 1989 and 1994). Bleeding time following microvascular puncture was also significantly longer in copper deficiency (Schuschke et al. 1994 and 1995).

Despite the suppression of thrombogenesis by copper deficiency, subsequent in vitro studies demonstrated that adenosine diphosphate (ADP)⁵-induced platelet aggregation is greater for platelets from copper-deficient rats than for those from copper-adequate rats, suggesting that platelet-to-platelet interaction is enhanced (Lominadze et al. 1996). Because similar results were observed when platelets were washed with calcium-free Tyrode buffered saline solution (TBSS), the increased platelet-to-platelet interaction was not caused by any plasma factors (Lominadze et al. 1996). It does appear, however, that decreased thrombogenesis in copper deficiency is in part produced by diminished levels of platelet adhesion proteins. Under static (Lominadze et al. 1996) or flow (Lominadze, D., unpublished data) conditions, platelets from copper-deficient rats suspended in autologous plasma and added to cultured endothelial cells demonstrated decreased adherence to normal endothelial cells. These changes in platelet reactivity from copper-deficient rats were coincident with alterations in platelet von Willebrand factor (vWF ) and fibrinogen production (Lominadze et al. 1996). Addition of the purified vWF to platelets from copper-deficient rats restored impaired platelet-to-endothelial cell adhesion in a flow chamber (Lominadze, D., unpublished data).

In vivo thrombogenesis involves not only platelet-to-endothelium and platelet-to-platelet binding but rheological factors that have a modulatory effect on platelet reactivity. Shearing forces present during the flow of blood through blood vessels directly activate platelets, resulting in the release of proaggregatory mediators such as vWF and ADP and in the expression of adhesion molecule binding sites (Marcus and Safier 1993, Ruggeri 1993, Sato and Ohshima 1990, Slack et al. 1993). vWF is one of the most important of the adhesive proteins involved in the binding of activated platelets to the vessel wall; it is found in plasma, endothelial cells and in the -granules of platelets and mediates platelet adhesion to both altered endothelium and exposed subendothelial matrix (Ware and Heistad 1993). Adhesion occurs via cell surface glycoprotein receptors (Ware and Heistad 1993) and is linked to the mobilization of platelet cytosolic calcium (Hamilton and Sims 1987).

ADP is an agonist released by platelets in response to shear stress; it mediates the recruitment of platelets that aggregate upon the initial layer of adherent platelets (Marcus and Safier 1993). In addition, an increase in free calcium in the cytosol of activated platelets leads to the expression of the fibrinogen receptor on the platelet surface and to platelet-to-platelet binding (Marcus and Safier 1993). Thus, in general, shear-induced changes in platelets favor thrombogenesis.

Besides inducing platelet activation, shear forces also activate potassium channels in the endothelial cell wall, thereby leading to the release of the vasodilator and antiaggregatory nitric oxide (NO) (Hirafuju and Shinoda 1993, Slack et al. 1993), and of prostacyclin (PGI₂), which is an inhibitor of platelet aggregation (Hirafuju and Shinoda 1993) and inhibits platelet adhesion to the subendothelium (Adelman et al. 1981).

Thus shear forces affect both the platelets and the endothelial cells to produce different effects on the processes of thrombus formation. They enhance thrombus formation by activating platelets, whereas they mitigate thrombogenesis by activating endothelial cells. The objective of the present study was to determine the effect of shear forces on in vivo platelet thrombus formation during copper deficiency. Comparisons were made between acute focal injuries to individual microvessels in the rat cremaster muscle. The phenomena were studied by direct observations of the microcirculation using intravital video microscopy and by measurements of hematocrit and blood and plasma viscosity. We also determined the ability of purified vWF to restore the depressed thrombotic behavior in copper deficiency under conditions of high and low wall shear rates.

MATERIALS AND METHODS

For measurement of blood and plasma viscosity, Brookfield's "cone and plate"-type viscometer was used (Model LVDV-II, with cone angle 0.8°, Brookfield Engineering Laboratories, Stoughton, MA). Plasma (_P) and blood (_B) viscosities were measured at 26-27°C at shear rates of 75, 150, 225 and 450 s¹. Because the plasma viscosity is shear rate independent (Chien et al. 1971), the obtained results were averaged. Plasma viscosity was determined as an index of plasma protein concentration (including vWF and fibrinogen), whereas blood viscosity was measured to calculate "relative" blood viscosity. Further, we calculated the relative blood viscosity (_B/_P) which indicates the contribution of the red blood cells to blood viscosity independently of the plasma viscosity effect.

The plasma was made "lipid free" by centrifugation at 18,000 × g for 30 min at 4°C and removing the lipid-free plasma from the bottom of the centrifuge tube. Cryoprecipitate formation was facilitated by 50% ice-cold ethanol mixed with the lipid-free rat plasma to a final concentration of 3% (v/v). The plasma-ethanol mixture was incubated in a refrigerated water bath at 3°C for 60 min and gently mixed every 10 min to facilitate optimal precipitation of rat vWF. The cryoprecipitate fraction was removed from other plasma proteins by centrifugation at 4°C for 15 min at 12,000 × g. The cryosupernatant fraction was decanted and the cryopellet was then dissolved in 2.1 mL of 0.05 mol/L Tris 0.15 mol/L NaCl (pH 7.4) solution.

Purification of rodent vWF was achieved by using minor modifications of the technique for purification of canine vWF (Benson et al. 1986). Briefly, the rodent cryoprecipitate fraction (~2.1 mL) was applied to the top of a 2.5 × 40 cm, agarose gel column (Bio-Gel A-50m, Bio-Rad, Richmond, CA) that was equilibrated with Tris-buffered saline (0.05 mol/L Tris 0.15 mol/L NaCl, pH 7.4) at room temperature. Rodent vWF was eluted at a flow rate of 20 mL/h in 35 5-mL fractions.

The amount of vWF in each fraction was measured by ELISA (Benson et al. 1992). As a capture antibody, rabbit anti-human vWF antibody (no. A082, Dako, Santa Barbara, CA) was used for coating the microtitrator plate well (rodent vWF will bind to this capture antibody). Plasma was carefully washed from the system and a second "sandwich" antibody (goat anti-human factor VIII-related antigen-IgG fraction; lot no. 55797, Atlantic Antibodies, Scarborough, ME) was added to the system. Swine anti -goat IgG (TAGO, Burlingame, CA) was used as the detection antibody.

Fractions 19-26 contained detectable vWF. A single major peak fraction (fraction 25, 3.95 U vWF/mL) was obtained and stored at 70°C. A broader ascending peak corresponding to higher molecular weight forms of rodent vWF that eluted between fractions 19 and 24 (1.06 U vWF/mL) was subsequently pooled and stored at 70°C for use in the in vivo experimentation. Column recovery of partially purified vWF was 73%.

After surgical preparation, the rats were positioned on a modified stage of a microscope (Nikon, Tokyo, Japan) so that the cremaster muscle, which is approximately 200-250 µm thick, could be observed by transmitted light or epi-illumination. A closed-circuit television system was used to observe the microcirculation in precapillary arterioles and postcapillary venules.

After the surgical preparation and preceding each experiment, there was a 30-min equilibration period. After the equilibration, fluorescein isothiocyanate conjugated to bovine serum albumin (FITC-BSA, 20 mg/100 g of body wt) was injected intra-arterially and allowed to circulate for 20 min. Epi-illumination with blue light (450-490 nm) from a mercury arc lamp (Miller et al. 1992) was used to photoactivate a segment of arteriole (23-27 µm in diameter) and then a segment of a postcapilary venule (26-28 µm in diameter). This technique causes microvessel platelet thrombus formation and hemostasis without endothelial denudation and exposure of the subendothelium (Miller et al. 1992, Sato and Ohshima 1990). The power density (set at 1.8 W/cm²) was measured at the focal plane of the objective by using an optometer (model 1815, Newport/Klinger, Irvine, CA).

In the first series of photoactivation experiments, platelet thrombus formation was induced in arterioles and venules of six CuA and eight CuD rats. After the photoactivation of both an arteriole and a venule, a second technique for inducing a focal, acute vessel injury was used in five of the CuA and five of the CuD rats. A glass micropipet and a micromanipulator were used to puncture a 60- to 80-µm diameter venule. This technique mechanically disrupts the entire vascular wall and permits adhesion of the circulating platelets to the subendothelial collagen (Schuschke et al. 1994 and 1995). Bleeding time as an index of platelet-plugging and hemostasis was then measured.

The second series of experiments was conducted to determine the effect of vWF on the depressed thrombogenesis in microvessels. Photoactivation was again used to induce thrombus formation in five CuA and five CuD rats to establish base-line values. After the control photoactivation, another microvascular field in the same rats with comparable vessel diameter and shear rate was selected. Purified vWF (1 U/mL; 0.2 U/100 g body wt) was then injected via the carotid cannula. Ten minutes after the protein injection, the experimental field was photoactivated and vessel occlusion was quantified. After the second photoactivation, the micropuncture technique was performed, and the results were compared with the bleeding time results in the rats from in the first series of experiments in which vWF was not present.

1

Before photoactivation in each vessel, flow was adjusted if necessary to provide wall shear rates of ~2000, 500, 300 or 150 s¹. Arterioles were used for the two higher shear rates and venules for the two lower shear rates because of the relative differences in velocity and diameter in the two types of vessels. Flow adjustments were made by using a partial manual restriction of the appropriate upstream arterioles or venules with a polished glass pipet and a micromanipulator. Similarities of wall shear rates in the microvessels of CuA and CuD rats enabled us to minimize differences in the flow conditions in microvessels with similar internal diameters.

Photoactivation was not started until wall shear rate in the selected vessel remained unchanged for at least 5 min. The time from the start of photoactivation until complete vessel occlusion was determined. The vessel was considered to be completely occluded when the flow stopped for at least 30 s. The image of the cremaster muscle microvessel was displayed on a calibrated monitor, and the red blood cell column and luminal diameter of the vessel were measured from a television monitor screen at 5-min intervals.

Assuming that the value of wall shear rate was changing from its initial maximal value (before photoactivation) to zero (occluded vessel) during the time period necessary for complete vessel occlusion, we calculated the variable "shear rate change," which represents changes of wall shear rate during the time period necessary for vessel occlusion. It is expressed in s² as follows:

We have calculated the ratio of all variables (platelet thrombus initiation time, growth period and rate of vessel occlusion) for CuA rats to CuD rats at all wall shear rates used in both arterioles and venules. This was done to determine whether the effects of shear rate or abnormalities caused by copper deprivation are the result of changes in platelet-to-endothelial cell (thrombus initiation time  platelet adhesion) and platelet-to-platelet interaction (thrombus growth period  platelet aggregation) or thrombus formation in general (shear rate change). To allow for statistical comparisons, the CuA/CuD ratio for each CuD rat was calculated by comparing the individual CuD data with the mean CuA response. The means ± SD were then determined for the individual CuA/CuD ratios and compared for each shear rate.

RESULTS

Table 1. Hepatic copper and iron concentration, body weight, hematocrit, blood pressure and heart rate in copper-adequate (CuA) and copper-deficient rats with and without intravascular bolus of supplemental von Willebrand factor1,2 [View Table]

Plasma viscosity was lower in CuD than in CuA rats (Table 2). Blood viscosity decreased with increasing shear rates and was significantly less in CuD rats than in CuA rats (Table 2). The differences in blood viscosity may have been caused by lower plasma viscosity (Table 2) and lower hematocrit in CuD rats (Table 1). The relative blood viscosity (_B/_P) also was lower in the CuD than in the CuA group at all shear rates (Table 2).

Table 2. Plasma and blood viscosity in copper-adequate (CuA) and copper-deficient (CuD) rats1 [View Table]

The initial diameters and wall shear rates for the vessels used in the first photoactivation series are given in Table 3. Photoactivation of the intravascular FITC-BSA caused platelet thrombus formation and eventual vessel occlusion in arterioles and venules of both CuA and CuD groups. The time period for appearance of platelets adhered to the vessel wall (thrombus initiation time) was significantly prolonged in both arterioles and venules of CuD rats compared with those of CuA rats (Fig. 1). In arterioles, the ratio of the time period for platelet thrombus initiation in the CuA group to that in the CuD rats (0.7 ± 0.1) at shear rates of ~2000 s¹ was not different from the ratio (0.6 ± 0.1) at shear rates of ~500 s¹. In venules, the ratio of the platelet thrombus initiation in the CuA group to that in the CuD rats (0.3 ± 0.02) at shear rates of ~300 s¹ was not different from the ratio (0.3 ± 0.03) at shear rates of ~150 s¹. However, the ratio of the platelet thrombus initiation time in the CuA group to that in the CuD rats was significantly higher in arterioles (0.6 ± 0.1) with relatively low (~500 s¹) wall shear rates compared with that in venules (0.3 ± 0.02) with relatively high (~300 s¹) wall shear rates (Fig. 1).

Table 3. Initial diameters and shear rates in arterioles and venules used for photoactivation-induced thrombogenesis in copper-adequate (CuA) and copper-deficient (CuD) rats1 [View Table]

Platelet thrombus growth time (Fig. 2) after thrombus initiation was significantly shorter in CuD compared with CuA rats in arterioles at low (~500 s¹) wall shear rates and in venules at both high (300 s¹) and low (150 s¹) wall shear rates. In arterioles, the ratio of platelet thrombus growth time in the CuA group to that in the CuD rats (1.1±0.1) at shear rates of ~2000 s¹ was significantly lower than the ratio (1.8 ± 0.4) at shear rates of ~500 s¹. Despite differences in the platelet thrombus growth time in the CuD group, the ratios of the platelet thrombus growth time in the CuA group to that in the CuD group were not different at the higher (1.7 ± 0.1) and at the lower (1.9 ± 0.2) wall shear rates within venules (Fig. 2).

The shear rate change (Fig. 3) was significantly less in both arterioles and venules of the CuD rats compared with the CuA group at all wall shear rates used. In arterioles, the ratio of the shear rate change in CuA rats to that in the CuD rats (1.5 ± 0.1) at shear rates of ~2000 s¹ was not different than the ratio (1.9 ± 0.3) at shear rates of ~500 s¹. In venules, the ratio of the shear rate change in the CuA rats to that in the CuD rats (2.1 ± 0.1) at shear rates of ~300 s¹ was not different than the ratio (2.9 ± 0.4) at shear rates of ~150 s¹. The ratios of the shear rate change in the CuA group to that in the CuD group also were not different in arterioles (1.9 ± 0.3) and in venules (2.1 ± 0.1) with similar (low in arterioles and high in venules) wall shear rates (Fig. 3).

The initial diameters and wall shear rates for the vessels used in the second photoactivation series before and after the injection of vWF are given in Table 3. Light dye-induced platelet thrombus formation in arterioles and venules of CuD rats was restored after systemic injection of the purified vWF. Platelet thrombus initiation time (Fig. 4) was significantly shortened after the addition of vWF in arterioles and venules of the CuD group. In addition, after the vWF injection there was no difference between CuA and CuD values in either arterioles or venules.

Platelet thrombus growth time following thrombus initiation was significantly shorter in the venules of CuD rats than in those of CuA rats, and the addition of vWF did not change the growth time of either CuA or CuD groups (Fig. 5). However, the shear rate change (Fig. 6) was significantly increased by vWF in arterioles of CuD rats, whereas no significant change was observed in the CuA group.

The bleeding time following micropuncture of second-order venules was significantly prolonged in CuD rats compared with that in the CuA rats (Fig. 7). Addition of the vWF significantly shortened bleeding time in CuD rats, whereas there was no significant change in CuA rats (Fig. 7).

DISCUSSION

The values of the shear rates used in the viscometry study were typical for rat cremaster muscle microvessels (venules). At low shear rates, relative blood viscosity can be used as an index of the degree of red blood cell aggregability (Chien et al. 1966). Differences in blood viscosity or relative blood viscosity between CuA and CuD groups at the low shear rates (75 s¹) were similar to those at higher shear rates. This suggests that alterations of the blood rheological variables in copper deficiency are the result mainly of lower hematocrit (Table 1) and possibly the higher plasma fibrinogen concentration (unpublished data) seen during copper deficiency in rats. Decreased hematocrit may reduce the proaggregatory effect of red blood cells (Santos et al. 1991), but we have observed depressed thrombogenesis in rats fed a copper-marginal diet when hematocrit was not decreased (Schuschke et al. 1995).

The increased platelet thrombus initiation time in both arterioles and venules of the CuD compared with the CuA group shows that platelet adhesion is decreased during copper deficiency in rats (Fig. 1). Greater wall shear rates (~2000 to ~500 s¹) in arterioles were associated with an increase in the time necessary for platelet adhesion to occur compared with platelet adhesion at the lower wall shear rates (~300 to ~150 s¹) observed in venules (Fig. 1). These results are in agreement with other studies (Rugerri 1993, Turitto and Baumgartner 1987) that suggest that the high wall shear rates retard the platelet adhesion process.

The ratio of thrombus initiation time in CuD to CuA rats was not different between arterioles at higher and lower shear rates or between venules at higher and lower shear rates. However, there was a difference in the ratio of thrombus initiation time between arterioles at lower shear rates and venules at higher shear rates (Fig. 1), suggesting that platelet adhesion is influenced by wall shear rate and/or by functional differences between arteriolar and venular endothelium.

Although the time for platelet thrombus growth after thrombus initiation was not different in arterioles at high (~2000 s¹) wall shear rates in the CuA group compared with that in CuD group (Fig. 2), thrombus growth time was significantly lower in arterioles at lower (~500 s¹) shear rates and in venules of the CuD rats compared with that in the CuA group (Fig. 2). These results are in agreement with our previous study (Lominadze et al. 1996), which shows that platelet aggregation is increased during copper deficiency. This could be the result of increased platelet fibrinogen concentration in CuD rats (Lominadze et al. 1996) particularly because there is a greater effect of fibrinogen on thrombogenesis at low shear rates (de Groot and Sixma 1987, Ross et al. 1995). These results also suggest that the platelet aggregation is not influenced by changes in wall shear rate or functional differences between arteriolar and venular endothelium in copper deficiency.

Copper deficiency decreased the shear rate change in both arterioles and in venules (Fig. 3). This variable represents the overall thrombotic process. The absence of differences in the ratios of the rate of vessel occlusion in the CuA to that in the CuD group at higher and lower wall shear rates in both arterioles and venules (see results section) suggests that the altered platelet thrombus formation during copper deficiency is not influenced by wall shear rate and/or functional differences between arteriolar and venular endothelium. Thus it must be the result of impaired platelet-to-endothelial cell interactions. The results of the current study support the concept that reduced platelet adhesion and not altered response to shear force stimuli is the mechanism behind our previous in vivo observations of decreased thrombus formation and hemostasis in copper-deficient rats (Schuschke et al. 1989, 1994 and 1995).

We have previously reported a significantly greater total number of platelets in the plasma of CuD compared with CuA rats (Schuschke et al. 1994). This would initially suggest a hyperthrombotic response. However, Johnson and Dufault (1993) demonstrated that calcium mobilization is reduced in platelets from copper-deficient rats. Because a rise in free calcium in the cytosol is a critical event in platelet activation (Marcus and Safier 1993), a decreased release of Ca²⁺ from intracellular stores would have an anti-aggregatory effect, even though the total number of platelets is increased. This would be consistent with our current data showing that both the thrombus initiation time is increased and growth rate is depressed by dietary copper restriction (Figs. 1 and 2).

Depressed production or binding of the adhesion glycoprotein vWF may be the mechanism by which platelet-endothelial adhesion is attenuated. vWF is synthesized in megakaryocytes and stored in platelet -granules (Ruggeri and Zimmerman 1987) until activation. It is also synthesized and secreted from cultured endothelial cells (Jaffe et al. 1974). We have previously shown that platelets from CuD rats contain significantly less vWF than platelets from CuA rats (Lominadze et al. 1996) and that, independent of plasma and associated coagulation factors, platelets from CuD rats have decreased ability to adhere to cultured endothelial cells under both static conditions (Lominadze et al. 1996) and at low wall shear rates (unpublished data) than do platelets from CuA rats.

In the current study, purified vWF significantly improved the impaired platelet thrombus formation in CuD rats (Figs. 4 and 6), suggesting that the decreased platelet and plasma vWF during copper deficiency (Lominadze et al. 1996) result in the impaired in vivo and in vitro thrombogenesis seen during dietary copper deficiency in rats (Lominadze et al. 1996, Schuschke et al. 1989, 1994 and 1995). The results also confirm that although vWF has an effect on platelet-to-endothelial cell adhesion at low wall shear rates (de Groot and Sixma 1987, Ross et al. 1995), the greater effect is on the platelet adhesion process at higher wall shear rates such as those seen in arterioles (Ruggeri 1993, Turitto and Baumgartner 1987).

The portion of the platelet thrombus growth time (platelet aggregation) after thrombus initiation was not altered in arterioles or venules of CuD rats by treatment with vWF (Fig. 5). These results confirm the widely accepted belief that the vWF is involved mainly in platelet-to-endothelial cell adhesion, which is the first stage of the thrombus formation process.

We have previously suggested that a change in the subendothelial collagen is a possible explanation for the reduced platelet thrombus formation in the copper-deficient rat (Schuschke et al. 1989). Because the copper-dependent enzyme, lysyl oxidase, (Owen 1982) is responsible for cross-linking elastin and collagen, there may be a structural change in these proteins and a loss of normal platelet-collagen adherence. Evidence that dietary copper deficiency causes disruption of the subendothelial structure, including the basement membrane and collagen, has been presented for the heart (Davidson et al. 1992) and the lung (Akers and Saari 1993). A change in the subendothelial collagen is a plausible explanation for the delayed platelet thrombus formation in response to micropuncture but is not likely in the response to photoactivation because the light-dye reaction does not involve endothelial denudation and exposure of the collagen (Miller et al. 1992). Additionally, restoration of the bleeding time to normal values in CuD rats treated with vWF (Fig. 7) suggests that the decreased concentration of the adhesive protein vWF seen during copper deficiency in rats (Lominadze et al. 1996) is more likely responsible for the decreased thrombogenesis than is an alteration of subendothelial collagen.

In summary, dietary copper deficiency significantly depresses thrombogenesis in both arterioles and venules of the rat microcirculation. This depression is apparently independent of vascular wall shear forces or structural/functional differences between arterioles and venules. Restoration of thrombogenesis in the CuD group to CuA control values by the addition of vWF supports the hypothesis that a defect in the platelet-to-endothelium adhesion mechanism is responsible for the reduced thrombogenesis in vivo as previously suggested by in vitro studies (Lominadze et al. 1996). Further, the current study confirms a greater role of vWF-mediated platelet adhesion at higher wall shear rates (Ruggeri 1993, Turitto and Baumgartner 1987; Fig. 4) and demonstrates that vWF binding in CuA rats is at an optimal level that is not increased by supplementation of the adhesive protein.

ACKNOWLEDGMENTS