Dietary copper is known to be essential for the normal functioning of the cardiovascular system in both humans and experimental animals. The cardiovascular defects in copper deficiency are often associated with impaired activity of some of the ~30 copper-dependent enzyme systems in living organisms. Structurally, weakened heart and blood vessel walls have been attributed to reduced activity of the copper-dependent enzyme lysyl oxidase (Owen 1982) and the resultant impaired cross-linking of elastin and collagen. Many vascular functions are also dependent on copper intake. Studies of large conduit blood vessels in copper-deficient rats have reported altered vasoactive responses to catecholamines (Kitano 1980), endothelium-dependent agents (Lynch et al., 1997, Saari 1992) and altered release of prostacyclin (Nelson et al. 1992). These results suggest that defects in copper-dependent dopamine -monooxygenase, soluble guanylate cyclase and antioxidant defenses can directly or indirectly affect vascular function. The development of hypertension in copper-deficient rats (Klevay 1987, Medeiros 1987) may be the result of impaired vasoactive functions in the small resistance blood vessels, thus suggesting a role for copper in microvascular control mechanisms.

This review will summarize recent work that relates primarily to the role of dietary copper in the physiology of the microcirculation. The microcirculation is usually defined as those blood vessels distal to the conduit arteries and before the veins, including the arterioles, the capillaries and venules. The arteriole is characterized by the presence of one to several layers of vascular smooth muscle in its wall with more muscled arterioles in the proximal region of the vascular tree. The capillary is a single layer of endothelial cells with an outer sleeve of basement membrane. The venules are similar but have a few added smooth muscle cells dispersed within their walls. The main functions of the microcirculation are the transport, diffusion and exchange of materials between blood and tissue for the purposes of the delivery of nutrients, the removal of metabolic products, tissue defense and repair.

Much of the regulation of microvascular function is associated with the endothelial cells at the blood-tissue interface. The endothelium is a dynamic structure that has a major influence on blood components and vascular smooth muscle tone as well as the response to injury. The endothelium also acts as a selective barrier between the components of the blood and the extravascular space and conveys signals between the tissues and the circulating elements. To perform these functions, endothelial cells synthesize and interact with a variety of hormones and mediators; some of these are expressed constitutively, whereas others are induced only in response to certain stimuli. The three major areas of endothelial function that will be covered in this review are vascular permeability to macromolecules, platelet-endothelial interactions and vascular smooth muscle reactivity.

The microcirculation of the rat cremaster muscle was used as a model to investigate in vivo the role of copper in the regulation of microvascular control mechanisms related to the endothelium. In this model, dietary copper deficiency was induced by feeding a purified diet (Johnson and Kramer 1987) containing 0.4 µg Cu/g diet (copper-deficient diet). The purified diet was made either copper marginal by adding 1.5 or 3.0 µg Cu/g diet or copper adequate by adding 6 µg Cu/g diet. These diets yield hepatic copper concentrations of 32 ± 4.3, 105 ± 12.6, 135 ± 7.9 and 205 ± 20.5 µmol/kg, respectively, after 3 wk (Schuschke et al. 1994b and 1995b).

MACROMOLECULAR PERMEABILITY

By using the cremaster muscle as a window to observe the microcirculation, we found that copper-deficient rats had a significantly greater macromolecular leakage from postcapillary venules (Fig. 2) as a result of mast cell degranulation induced by compound 48/80 (Schuschke et al. 1989 and 1994a). This enhanced leakage to compound 48/80 was blocked by pretreatment with diphenhydramine (Schuschke et al. 1994a), which is a histamine H1 receptor blocker. Contrary to the mast cell degranulation response, simple topical administration of various concentrations of histamine or other endothelial-mediated inflammatory mediators (serotonin and bradykinin) induced macromolecular leakage that was not significantly different between the adequate and deficient groups (Schuschke et al. 1994a). These results suggest that mast cell histamine mediates the increased leakage. This increased histamine-mediated leakage does not appear to be the result of any increased sensitivity to histamine because there is no difference in the histamine concentration-response curves between dietary groups (Fig. 2). In addition, in these experiments, there was no difference between dietary groups in spontaneous base-line leakage of protein. Therefore, although the altered cross-linking of collagen and elastin in the extracellular matrix during copper deficiency (Farquharson et al. 1989, Owen 1982) may promote protein extravasation when endothelial integrity is compromised, there does not appear to be a change in endothelial integrity in the copper-deficient rat. Thus it was hypothesized that dietary copper deficiency increased the macromolecular leakage associated with mast cell degranulation by a primary effect on the mast cell rather than on the endothelial cell.

This hypothesis was tested in isolated peritoneal mast cells from copper-adequate and copper-deficient rats. The spontaneous release of histamine, histamine concentration and the amount of histamine released with increasing concentrations of compound 48/80 were measured in the mast cells. No differences were found for these parameters between dietary groups (Schuschke et al. 1994c). Cremaster muscles from copper-adequate and copper-deficient rats were then fixed and stained for mast cells with toluidine blue. The copper-deficient animals had significantly more mast cells per 5-µm section compared with the copper-adequate group (Schuschke et al. 1994c). From these studies, we concluded that dietary copper deficiency increases the mast cell population but does not alter the histamine content or sensitivity to degranulation of individual mast cells in rats. The end result of this increased mast cell population is an increase in the total histamine released during degranulation.

Although the increase in tissue mast cells may be a mechanism by which acute inflammation is exaggerated during copper deficiency, the reason for this change in the mast cell population remains to be determined. In addition to mast cells, the numbers of platelets (Schuschke et al. 1994b) and neutrophils (Karimbakas and Percival 1997) have also been shown to be greater in copper-deficient rats and mice, respectively, whereas hematocrit is decreased (Johnson and Saari 1991). This increase in inflammation-related cells suggests that copper-deficient rats are "primed" for an acute inflammatory response by an as yet unknown mechanism.

PLATELET-ENDOTHELIAL INTERACTIONS

Several studies in the cremaster muscle microcirculation of copper-deficient rats have identified a significant inhibition of in vivo platelet thrombus formation (Lominadze et al. 1997, Schuschke et al. 1989, 1994b and 1995b). In these studies, focal thrombotic events were triggered by either micropuncture of a vessel, which disrupts the endothelium and exposes the underlying collagen to the circulating platelets, or by photoactivation of a light-sensitive intravascular dye, which induces platelet thrombus formation without compromising endothelial integrity. After micropuncture of a 50- to 80-µm diameter venule, bleeding time was determined as an index of platelet plugging.

The bleeding time was significantly longer (Fig. 4) and the hematocrit was significantly lower in weanling rats 1 wk after starting the copper-deficient diet compared with age-matched copper-adequate controls (Schuschke et al. 1994b). Because erythrocytes enhance platelet activity (Santos et al. 1991), the low hematocrit may be a mechanism for the inhibited hemostasis. In rats that were copper deficient for 3 and 5 wk, platelet thrombus induced by photoactivation was delayed and prothrombin time was significantly longer (Schuschke et al. 1994b), suggesting inactivity of a plasma coagulation factor as a mechanism for the inhibited hemostasis. The attenuated platelet thrombogenesis occurred even though there were a significantly greater number of circulating platelets. These results suggest a defect in the thrombotic function of copper-deficient platelets as a cause for the depressed hemostasis. Relative blood viscosity was also significantly lower in copper-deficient rats compared with the copper-adequate group (Lominadze et al. 1997). This suggests that wall shear rate may be greater during copper deficiency, causing an altered thrombotic response.

Subsequent mechanistic studies have examined the role of dietary copper in platelet thrombus formation and hemostasis in rats. Copper-marginal diets (3.0 and 1.5 µg Cu/g diet) were used to eliminate the anemia seen with the copper-deficient diet (Schuschke et al. 1995b). After rats were fed these marginal diets for 3 wk hepatic copper was significantly lower compared with the copper-adequate controls, but this occurred without the decreased hematocrit seen in the severely copper-deficient rats. Bleeding time after micropuncture was significantly longer at all diet times in the group fed 1.5 µg Cu/g diet and after 5 wks in the group fed 3.0 µg Cu/g diet (Schuschke et al. 1995b). The correlation between hepatic copper concentration and bleeding time was significant (P < 0.0001) and suggested that bleeding time increases when hepatic copper concentrations are <4 µg/g (62.9 µmol/kg) dry tissue weight. However, there was no difference in vessel occlusion time during photoactivation, prothrombin time or in the activity of copper-related plasma coagulation factors (V, VII, VIII) between the copper-adequate and copper-marginal groups (Schuschke et al. 1995b). The significantly longer bleeding time still occurred in the absence of a change in hematocrit or loss of coagulation factor activity. This obviously suggests that these are not the mechanisms causing depressed thrombogenesis in copper deficiency. Other in vivo microvascular studies have also shown that the longer time for platelet thrombus formation is independent of the effect of the blood velocity gradient within the vessels (wall shear rate) (Lominadze et al. 1997).

Dietary copper deficiency produces several different and often opposing platelet effects that have been demonstrated in vitro and provide evidence for augmentation of some components of thrombus formation and for depression of others. Platelets from copper-deficient rats undergo greater actin polymerization and accumulate elevated amounts of cytoskeletal myosin when stimulated with thrombin (Johnson 1993, Johnson and Dufault 1989). The rate of dense granule secretion is increased twofold in thrombin-activated platelets from copper-deficient rats (Johnson 1993, Johnson and Dufault 1989). In addition, when copper intake is restricted in rats, there is greater platelet thromboxane synthesis (Morin et al. 1993). The increased cytoskeletal myosin along with greater dense granule secretion (including ADP and thromboxane) suggest a proaggregatory effect of dietary copper deficiency on platelet reactivity. However, intracellular calcium mobilization is inhibited in platelets from copper-deficient rats (Johnson and Dufault 1993). Because a rise in cytosolic free calcium is essential for platelet activation and consequent aggregation events, the inhibited calcium mobilization would suggest an antiaggregatory effect of copper deficiency.

The processes of platelet-to-platelet aggregation and platelet-to-endothelial cell adhesion were examined in vitro to characterize platelet function in copper-deficient rats (Lominadze et al. 1996). Platelet aggregation in platelet-rich plasma samples was induced by adding ADP and was measured in a turbidometric platelet aggregometer. Platelet adhesion was determined for platelets from copper-adequate and copper-deficient donor rats. The platelets were incubated with cultured rat endothelial cells from normal animals. Dietary copper deficiency caused greater platelet aggregation and less platelet adhesion compared with copper-adequate controls (Lominadze et al. 1996). Subsequently, similar results have been reported in vivo in the rat cremaster microcirculation (Lominadze et al. 1997). The time to thrombus appearance (adhesion) was significantly longer, whereas the thrombus growth after thrombus appearance (aggregation) occurred more quickly in copper-deficient rats compared with copper-adequate rats.

The greater platelet aggregation suggests that altered platelet-to-platelet interactions are not responsible for the decreased thrombogenicity seen in copper-deficient rats. However, these results clearly indicate that there is a role for copper in normal platelet aggregation. A possible mechanism for enhanced platelet aggregation is the concentration of fibrinogen available for platelet-to-platelet binding. The higher platelet fibrinogen concentration in the copper-deficient rats (Lominadze et al. 1996), combined with a normally higher affinity of platelet binding sites for fibrinogen than for the adhesion molecule von Willebrand factor (vWF)⁴ (Gralnick et al. 1984), suggests that platelet-to-platelet aggregation would increase in copper deficiency.

The decreased in vitro platelet adhesion to endothelial cells provides a possible mechanism for the reduced thrombogenesis and hemostasis we observed in vivo. Because it was possible that the presence or absence of plasma coagulation factors (including copper-related V and VIII) could have been responsible for the attenuated adhesion, a similar series of experiments was performed using washed platelets suspended in Tyrode buffered saline solution. Again, we found that adhesion of platelets from copper-deficient rats to cultured endothelial cells was lower than the adhesion of platelets from copper-adequate rats (Lominadze et al. 1996). This suggests that plasma components do not have a significant effect on the adhesion process during the described conditions. Thus, proteins that are present in the plasma and participate in general platelet-to-endothelial interactions do not appear to have a significant effect on the reduced platelet adhesion seen during copper deficiency. Therefore the copper-deficient platelet itself is most likely the site of this defect in hemostasis.

There are several adhesion molecules that regulate the adhesive properties of the platelets and vessel wall to prevent serious bleeding. Of these, only vWF is specific for platelets (Sixma 1987). vWF is synthesized in both endothelial cells (Wall et al. 1980) and in megakaryocytes where it is subsequently stored in platelet -granules (Ruggeri and Zimmerman 1987). The -granules contain ~20% of the total circulating vWF and are known to contain the larger molecular weight multimers (Gralnick et al. 1985, Hoyer and Shainoff 1980), which are important for platelet adhesion because larger multimers interact preferentially with platelets (Gralnick et al. 1981). Platelets from copper-deficient rats contain less vWF than platelets from copper-adequate rats (Lominadze et al. 1996), and the addition of the high molecular weight fraction of vWF to the blood restores platelet thrombosis and hemostasis function to normal in the microcirculation (Fig. 4), suggesting that the receptor-binding of vWF is not affected by deficient copper nutriture. Thus the decreased platelet vWF could be the reason for the prolonged bleeding time in copper-deficient rats (Schuschke et al. 1994b and 1995).

Further studies demonstrated that plasma vWF was also significantly lower in copper-deficient rats (Lominadze et al. 1996). Because platelet vWF is contained in storage granules until the platelet is activated, plasma vWF concentrations likely reflect spontaneous endothelial release of the protein. Although our results suggest that plasma proteins do not have an important role in the decreased thrombogenesis in copper-deficient rats (Lominadze et al. 1996), the combined decrease of both platelet and plasma vWF suggests that the synthetic machinery for this protein is impaired during copper deficiency. The mechanism of this impairment and the role of copper in the synthetic pathway of von Willebrand factor are not known at this time.

Reduced activity of the copper-dependent enzyme lysyl oxidase may also contribute to the alteration of in vivo thrombus formation. This inactivation results in the abnormal cross-linked structure of the subendothelial collagen (Farquharson et al. 1989, Owen 1982). The denatured collagen may have altered binding affinity for adhesion molecules compared with the native protein structure (Davis 1992). Thus, altered collagen conformation may account for the significant change in bleeding time in the copper-marginal rats, whereas there was no difference with the photoactivation technique where collagen is not exposed. However, an altered collagen is not a likely mechanism for the reduced platelet adhesion to the cultured endothelial cells where there is no subendothelial collagen present. Furthermore, this mechanism does not explain the delayed adhesion observed in copper-deficient rats during photoactivation experiments in which there is no exposure to subendothelial collagen (Schuschke et al. 1989 and 1994b).

Lynch and Klevay (1992) have shown that dietary copper status may change hemostasis by altering the activity of plasma factors V and VIII. Although the studies presented in this review suggest a defect in platelet-to-endothelium adhesion, they do not eliminate the possibility of an effect of decreased factor V and VIII activity. The in vivo microvascular studies on hemostasis have been designed to examine only the initial, platelet-dependent phase of thrombogenesis. They have not addressed coagulation factor-dependent clot formation and stabilization, and therefore we cannot draw any conclusions on the effect of copper restriction on the function of coagulation factors.

VASCULAR SMOOTH MUSCLE REACTIVITY

Altered contractile and dilator responses have been observed in the blood vessels of copper-deficient rats. In the larger conduit vessels, these changes include reduced endothelium-dependent smooth muscle relaxation in rat aortas (Lynch et al. 1997, Saari 1992) and augmented aortic contraction to norepinephrine in rats (Kitano 1980). In the smaller resistance vessels, exaggerated vasoconstriction to angiotensin II has been reported in isolated whole lungs from copper-deficient rats (Allen and Saari 1994). In the rat cremaster muscle, there was no difference between copper-deficient and copper-adequate groups in the response of small (10- to 25-µm diameter) arterioles to norepinephrine, but nitric oxide (NO)-mediated vascular smooth muscle relaxation is inhibited (Schuschke et al. 1995a).

Arteriole vasodilation to acetylcholine, calcium ionophore A23187 and sodium nitroprusside, are inhibited by dietary copper deficiency (Schuschke et al. 1992). Despite the attenuated dilation to stimulation of the NO pathway by these agonists (Fig. 5), dilation to either second messenger analogs dibutyryl cGMP or dibutyryl cAMP and to the phosphodiesterase inhibitor papaverine are not inhibited in the copper-deficient rat (Schuschke et al. 1995a). These results suggest that the vascular smooth muscle relaxation mechanisms are intact and the dilator capacity of the arterioles is not altered. The inhibition of NO-mediated vasodilation without any effect on the arteriolar constrictor response to norepinephrine (Schuschke et al. 1995a) or other vasodilator pathways (Schuschke et al. 1995a and 1997) suggests that the depressed dilation is a pathway-specific effect of copper depletion.

Several possibilities exist concerning how copper deficiency may cause impairment of the NO-mediated smooth muscle relaxation. The first involves increased free radical activity associated with the reduction of antioxidant enzyme activity that occurs in copper-deficient rats (L'Abbé and Fischer 1984, Taylor et al. 1988). Evidence that oxidants are present and may induce damage during copper deficiency in rats is evidenced by greater concentrations of hepatic lipid hydroperoxides (Balevska et al. 1981), an enhanced production of breath ethane (Saari et al. 1990) and an enhanced production of peroxidation products in oxidatively stressed mitochondria of various organs (Fields et al. 1984, Paynter 1980). In addition, the cardiac hypertrophy and anemia associated with copper deficiency in rats have been reversed by antioxidants (Johnson and Saari 1989, Saari 1989). The presence of oxidants is important because they likely contribute to NO degradation.

Superoxide anion (O₂) is known to inactivate NO (Rubanyi and Vanhoutte 1986) and has been shown to inhibit cGMP-mediated relaxation of vascular smooth muscle in rats (Cherry et al. 1990). Superoxide dismutase (SOD) is the metabolizing enzyme of O₂, but because cytoplasmic Cu, Zn-SOD is a copper-dependent enzyme (Owen 1982), O₂ concentrations should be higher in copper-deficient animals. Elevated O₂ would lead to enhanced destruction of NO and an impairment of endothelium-dependent vasodilation (Fig. 5). We found that the dilator response of small arterioles to acetylcholine was significantly improved in the copper-deficient group after exposure to exogenous Cu, Zn-SOD (Fig. 6). The SOD, however, had no effect on the dilator response of the copper-adequate group and there was no difference between dietary groups in response to acetylcholine in the presence of SOD. The data suggest that during copper deficiency, excess superoxide anion may degrade NO, specifically decreasing NO-dilator capability. The direct inactivation of NO by vascular superoxide has also been proposed on the basis of in vitro studies in aortic rings in which NO-mediated smooth muscle relaxation was inhibited in copper-deficient rats (Lynch et al. 1997, Saari 1992) and in rat aortic rings treated with copper chelators to inactivate Cu, Zn-SOD (Omar et al. 1991, Plane et al. 1997).

5

Another possible effect of copper deficiency on NO-mediated relaxation involves the interaction of NO with soluble guanylate cyclase (GC-S). Copper and iron are transition metals and are components of this enzyme (Gerzer et al. 1981), which converts GTP to cGMP (Fig. 5). The iron is the metal component of the heme moiety of the GC-S. NO binding to the heme site is responsible for a conformational change in the GC-S and activation of the catalytic site of the enzyme (Schmidt et al. 1993). If copper is a functional cofactor with the iron in the NO-heme binding site, then NO may no longer be able to activate the GC-S when copper is inadequate. Alternatively, iron metabolism is known to be altered in dietary copper deficiency (Gubler et al. 1952) and may be a mechanism by which both the NO-heme binding is prevented and the NO activation of GC-S is depressed in copper-deficient animals.

Separate from the effects on NO-binding, copper depletion may also depress the activity of the GC-S enzyme, which is independent of heme content (Gerzer et al. 1981). Hydrogen peroxide, which activates GC-S by a NO-independent mechanism (Burke and Wolin 1987), was used to test the activity of the GC-S (Schuschke et al. 1995a). In these studies, the microvessel dilation to hydrogen peroxide was not different in the copper-deficient group compared with the copper-adequate group. Thus, these data suggest that the general activity of the GC-S is not affected by dietary copper deficiency. Further, the normal dilator response to the phosphodiesterase inhibitor papaverine in copper-deficient rats (Saari 1992, Schuschke et al. 1992) suggests that cGMP generation by GC-S is not inhibited by copper deficiency. Therefore, our results indicate that if copper is a functional component of GC-S, its role is likely at the NO-binding site but it is not a requisite for the basal activity of the enzyme. However, because the administration of Cu, Zn-SOD restored the dilation to acetylcholine, it is unlikely that altered NO-heme binding is the primary mechanism for the depressed vasodilation.

The results in the microcirculation suggest that the inactivation of cytosolic Cu, Zn-SOD by restriction of dietary copper results in the depression of NO-mediated vascular smooth muscle relaxation. Even though copper is a component of GC-S and copper deficiency may alter the heme binding site for NO on the enzyme, the restoration of dilation to acetylcholine by exogenous SOD suggests that it is the Cu, Zn-SOD enzyme that is most sensitive to copper restriction in the NO signaling system. Further, normal dilator responses to cGMP and cAMP (Schuschke et al. 1995a) suggest that the specific dilator mechanisms to these two second messengers are intact, and the normal response to norepinephrine (Schuschke et al. 1995a) demonstrates that there is not a generalized vasodepressive effect of copper deficiency. Because constitutive NO is important in basal vascular smooth muscle tone, this inhibition of NO-mediated dilation may contribute to the development of hypertension reported in copper-deficient rats (Klevay 1987, Medeiros 1987).

Copper may also contribute to altered vascular tone and responsiveness by several other mechanisms. The aortic production of the dilator prostacyclin is attenuated (Nelson et al. 1992), but the arteriolar sensitivity to this agonist appears to be greater during copper deficiency (Schuschke et al. 1997). The subendothelial collagen structure is also known to have a role in arteriole dilation particularly during injury (Mogford et al. 1996). This mechanism suggests an indirect role for lysyl oxidase in smooth muscle function. Preliminary evidence also suggests an upregulation of mRNA for the vasoconstrictor neuropeptide Y during copper deficiency in rats (Rutkoski-Mariner and Levenson 1997). Thus, dietary copper may regulate peripheral vascular resistance by altering the balance of vasoconstrictor and vasodilator signals being received by the vascular smooth muscle.

SUMMARY