Journal of Applied Physiology
Vol. 82, No. 2, pp. 389-398, February 1997
PULMONARY CIRCULATION AND LUNG FLUID BALANCE

Dirofilaria immitis: heartworm infection alters pulmonary artery endothelial cell behavior

Maria Mupanomunda¹, Jeffrey F. Williams², Charles D. Mackenzie³, and Lana Kaiser¹

¹ Department of Physiology, ² Department of Microbiology, and ³ Department of Pathology, Michigan State University, East Lansing, Michigan 48824-1101

ABSTRACT

INTRODUCTION

MATERIALS AND METHODS

RESULTS

DISCUSSION

ACKNOWLEDGEMENTS

FOOTNOTES

REFERENCES

ABSTRACT

Mupanomunda, Maria, Jeffrey F. Williams, Charles D. Mackenzie, and Lana Kaiser. Dirofilaria immitis: heartworm infection alters pulmonary artery endothelial cell behavior. J. Appl. Physiol. 82(2): 389-398, 1997.The pathogenesis of filariasis has generally been attributed to either physical presence of the adult parasites or the host's immune response to the parasites. However, the spectrum of filariasis cannot be entirely explained by these causes, and other mechanisms must be operative. It is now evident that factors released by filarial parasites likely contribute to the pathogenesis of filarial diseases. Adult heartworms (Dirofilaria immitis) reside in the right heart and pulmonary artery, so the pulmonary artery should be exposed to the highest concentration of filarial factors. We tested the hypothesis that endothelium-dependent relaxation is altered in the in vitro pulmonary artery from heartworm-infected dogs. Relaxation responses to endothelium-dependent vasodilators (methacholine, bradykinin, substance P, and A-23187) and the non-endothelium-dependent vasodilator nitroglycerin and contractile responses were measured in rings of pulmonary artery from control and heartworm-infected dogs. Endothelium-dependent relaxation was assessed in the presence and absence of inhibitors of nitric oxide synthase, cyclooxygenase, and guanylate cyclase. Responses to methacholine, substance P, and A-23187, but not to bradykinin, nitroglycerin, norepinephrine, or KCl, were depressed in pulmonary artery from heartworm-infected dogs when compared with control, suggesting that changes in endothelial cell and not vascular smooth muscle behavior are involved in altered relaxation. The mechanism of endothelium-dependent relaxation in control pulmonary artery appears to involve nitric oxide in the case of methacholine and both nitric oxide and a cyclooxygenase product in the case of bradykinin and A-23187. The mechanism of endothelium-dependent relaxation in pulmonary artery from heartworm-infected dogs was not clearly elucidated. These data provide no evidence that heartworm infection globally influences either endothelial cell receptor function or the vascular smooth muscle guanylate cyclase guanosine 3,5-cyclic monophosphate system, making it likely that changes in intracellular signaling are primarily responsible for the observed alteration of endothelium-mediated relaxation. Alteration of endothelial cell function by filarial parasites may be an important component in the pathology associated with filariasis.

canine heartworm; nematode; endothelium-dependent responses; filariasis; pathogenesis; methacholine; A-23187; substance P; bradykinin; nitroglycerin

INTRODUCTION

STUDY OF THE PATHOGENESIS of filarial diseases is complicated by the narrow host specificity of the human filarial pathogens and the parasites' complex life cycle (5, 10, 13-19, 28, 36). Microfilariae are live-borne young; in heartworm disease and lymphatic filariasis, the microfilariae circulate in the blood and enter a mosquito after it bites an infected mammal. Filarial parasites require an insect vector for both maturation of the larvae and transmission of the disease. After maturation, when the insect bites another mammal, the infective larva enters the skin, migrates to a suitable environment, and develops into a mature parasite. The mature parasites then produce microfilariae, and the cycle is completed.

Worldwide, over 400 million people and countless animals are infected with filarial parasites, yet human disease is generally limited to the tropics. This is due to the lack of suitable insect vectors in more temperate regions. In addition, the major human filarial pathogens Wuchereria bancrofti and Onchocerca volvulus infect only humans. In the United States, the parasite responsible for canine heartworm disease, Dirofilaria immitis, is the most prevalent filarial parasite. Canine heartworm disease is a major cause of morbidity and mortality in domestic dogs, and drugs targeted toward prevention of heartworm have found some utility in prevention of human disease. The human pathogens share many similarities with D. immitis, and these common behavioral, taxonomic, and biochemical characteristics make D. immitis an ideal animal model for the study of the pathogenesis of both human and animal filariasis.

Fig 6. Canine pulmonary arteries stained for factor VIII with alkaline phosphatase-labeled antibody. A: pulmonary artery from a control dog with an intact layer of evenly distributed endothelial cells and a relatively continuous layer of strongly positive staining reaction to factor VIII (magnification ×630). B: pulmonary artery from a heartworm-infected dog also showing an intact layer of endothelial cells and factor VIII staining (magnification ×630). C and D: high-power magnifications (×1,130) of tissues shown in A and B, respectively.
[View Larger Version of this Image (83K GIF file)]

Fig. 7 Pulmonary arteries, with intact or denuded endothelial cells, from control and heartworm-infected dogs stained for factor VIII by using alkaline phosphatase-labeled antibody (magnification ×630). A and B: representative pulmonary artery from control and heartworm-infected dogs, respectively, showing intact endothelial layers still present after experimentation. After denudation, vessels show only remnants of both endothelial cells and factor VIII staining. C and D: pulmonary artery from control and heartworm-infected dogs, respectively.
[View Larger Version of this Image (72K GIF file)]

The pathogenesis of filariasis has generally been attributed to either physical obstruction by the adult parasites or the host's immune response to adult parasites or microfilariae (5, 10, 25, 28, 31-33). However, the entire spectrum of filariasis cannot be explained by these mechanisms, and it is now evident that factors released by filarial parasites probably contribute to the pathogenesis of filarial diseases (13-21, 23, 34, 35). We now know that filarial parasites alter endothelium-dependent relaxation in vivo and in vitro (13-19, 23, 34, 35). In the in vivo femoral artery of heartworm-infected dogs, endothelium-dependent relaxation is seasonally depressed, with the maximal depressant effect observed in the spring (18). However, relaxation of in vitro femoral artery from heartworm-infected dogs is not different from control, suggesting that the changes seen in vivo are ephemeral and require continuous exposure to circulating filarial factors (35). In heartworm-infected dogs, adult D. immitis reside primarily in the right heart and pulmonary artery. Because of the proximity of the pulmonary artery to the adult parasites and their secreted products, we hypothesized that heartworm infection induces changes in pulmonary artery endothelial cell behavior that can be demonstrated in vitro. Experiments were designed to test the hypothesis that endothelium-mediated responses are altered in the in vitro pulmonary artery from heartworm-infected dogs.

MATERIALS AND METHODS

Animal Model

Isolation and Preparation of Vessels

Experimental Protocols

Dose-response relationships to norepinephrine (10 to 4.52 log M) were done to determine the dose of norepinephrine resulting in 60% maximum norepinephrine constriction. This concentration (5.52 log M) was then used to preconstrict rings in all experiments. High-potassium PSS (KCl 120 mM) was added to the tissue bath at the end of each experiment.

Cumulative dose-response relationships were established in separate rings to the endothelium-dependent vasodilators (2, 8, 13-19, 24, 34, 35, 37) methacholine (10 to 4.52 log M), bradykinin (10 to 4.52 log M), substance P (10 to 4.52 log M), and A-23187 (9.72 to 5.24 log M) and to the non-endothelium-dependent vasodilator nitroglycerin (11 to 5.52 log M). The calcium ionophore, A-23187, which causes endothelium-dependent relaxation without binding to receptors (8, 17), was used to rule out an effect of D. immitis on receptors. To determine the role of endothelial cells in relaxation, dose-response relationships to methacholine, bradykinin, substance P, and A-23187 were established in pulmonary arteries from control and heartworm-infected dogs, with and without endothelial cells. Denudation was considered successful if rings relaxed <10% to methacholine (5.52 log M).

To determine the mechanism of endothelium-dependent relaxation, the effects of inhibiting cyclooxygenase and the nitric oxide (NO)-guanylate cyclase-guanosine 3,5-cyclic monophosphate (cGMP) system were examined in rings from control and heartworm-infected dogs. Cumulative dose-response relationships to methacholine and A-23187 were established in the presence and absence of the NO synthase (NOS) inhibitors N -nitro-L-arginine methyl ester (L-NAME; 10 µM) or N ^G-monomethyl-L-arginine (L-NMMA; 100 µM) (29); an inhibitor of guanylate cyclase, methylene blue (10 µM) (10); and inhibitors of cyclooxygenase, mefenamic acid (10 µM) or indomethacin (10 µM) (9). Cumulative dose-response relationships to bradykinin were established in the presence and absence of L-NAME (10 µM), aminoguanidine (100 µM), a relatively selective inhibitor of inducible NOS, mefenamic acid (10 µM), or superoxide dismutase (SOD; 150 units/ml) and catalase (860 units/ml) (30). Inhibitors were added 30 min before preconstriction with norepinephrine (13, 14, 16, 17, 19, 22, 35).

To determine whether availability of L-arginine is limiting for methacholine relaxation, cumulative dose-response relationships to methacholine were done in rings from control and heartworm-infected dogs in the presence and absence of either L-arginine (1 mM or 100 µM) or D-arginine (1 mM), added 30 min before the rings were preconstricted (26).

Experiments were designed so that all dose-response relationships (± blockers) to the same agonist, in rings from the same dog, were done at the same time, thus obviating the need for time controls (22, 34, 35). When comparisons were made between treated and untreated rings, experiments were designed so that one ring from each dog served as the untreated ring, and one additional ring was used for each treatment. For example, in the bradykinin series for heartworm-infected dogs, a dose-response relationship to bradykinin was established on the first ring from dog 1; in the second ring, a dose-response relationship to bradykinin was established in the presence of mefenamic acid; and in the third ring from the same dog, a dose-response relationship to bradykinin was established in the presence of L-NAME. Blockers were incubated with the rings for 30 min. Untreated rings remained under identical conditions, without the presence of blockers, for 30 min. All of the dose-response relationships were started and completed at the same time.

Rings were weighed (wet and dry weight) after some experiments. Wet weights were determined after blotting the rings on filter paper, and dry weights were obtained after air drying the rings at room temperature for 24 h.

Histology

Drugs and Chemicals

Statistical Analysis

RESULTS

Effect of Heartworm on Contractile Responses

Effect of Heartworm on Endothelium-Dependent and -Independent Relaxation

Effect of Heartworm on Mechanism of Relaxation

Histology

Weights

DISCUSSION

Our results suggest that alterations of vascular reactivity in pulmonary artery from heartworm-infected dogs, previously attributed to endothelial cell damage and/or vascular smooth muscle hyperplasia (27, 28), are likely due to changes in the behavior of endothelial cells. The data provide no evidence that heartworm infection globally influences either endothelial cell-receptor function or the vascular smooth muscle guanylate cyclase-cGMP system, making it likely that changes in intracellular signaling are primarily responsible for the observed alteration of endothelium-mediated relaxation.

One could argue that it is not endothelial cell but rather vascular smooth muscle that is the primary vascular target of heartworm infection. This contention is not supported by our results. Heartworm infection did not influence the ability of the pulmonary artery to constrict to either KCl or norepinephrine, suggesting that the contractile elements of vascular smooth muscle are not different between groups. Furthermore, consistent with our hypothesis that vascular smooth muscle function and the guanylate cyclase-cGMP system are normal in pulmonary artery from heartworm-infected dogs, nitroglycerin relaxation is not altered by heartworm infection. Relaxation to the endothelium-dependent vasodilators methacholine, substance P, and A-23187 is profoundly depressed in pulmonary artery from heartworm-infected dogs, whereas relaxation to bradykinin is not different from control. In addition, histamine causes a marked constriction in control pulmonary artery; however, in the same vessel from heartworm-infected dogs, histamine is an endothelium-dependent vasodilator (34). Our interpretation of these data is that heartworm infection does not cause a pandepression of endothelial cell function but rather selectively alters endothelial cell behavior. The change does not appear to involve either the vascular smooth muscle contractile machinery or guanylate cyclase-cGMP system.

Both the mechanism and the magnitude of vascular responses of pulmonary artery are heterogeneous, with the mechanism of cholinergic relaxation being most studied and most consistent (4, 8, 37). In this study, the mechanism of endothelium-dependent relaxation of control pulmonary artery appears to involve NO, in the case of methacholine, and both NO and a cyclooxygenase product, in the case of A-23187 and bradykinin. The obligatory role of endothelial cells in relaxation to these agonists in the in vitro canine pulmonary and intrapulmonary artery has been well established (2, 4, 8); however, the mechanism of relaxation appears more variable and may involve NO, prostanoids, and hyperpolarizing factor. Despite the use of widely different concentrations of indomethacin and NOS inhibitors, cholinergic relaxation in canine pulmonary and intrapulmonary artery likely involves the NO-guanylate cyclase-cGMP system, but not cyclooxygenase products. In our study, bradykinin relaxation appears to be mediated by both NOS and cyclooxygenase products. This is in contrast to an earlier report on canine intrapulmonary artery strips preconstricted with serotonin and exposed to a lower concentration of indomethacin that "relaxations induced by a median effective concentration or by higher concentrations of bradykinin were not inhibited by addition of indomethacin or any of the other pharmacological antagonists" (2). However, in phenylephrine-preconstricted rings of canine pulmonary artery, indomethacin and captopril significantly enhanced the sensitivity to bradykinin; the effect of indomethacin alone was not reported (24).

The mechanism of endothelium-dependent relaxation of pulmonary artery from heartworm-infected dogs is more difficult to decipher. The lack of effect of inhibitors in vessels with significant depression of relaxation may merely reflect the marked depression of relaxation and not provide insights into the pathway involved. Consistent with this hypothesis are our data showing that, although L-NAME and methylene blue depress methacholine relaxation, L-NMMA does not. Maximum methacholine relaxation in untreated control rings was >50% in both the L-NAME and methylene blue groups, but <40% in the L-NMMA group. The depression of methacholine relaxation with L-NAME, but not with L-NMMA, could be due to the former's ability to inhibit muscarinic receptors (1); however, this is not consistent with the ability of L-NAME to inhibit bradykinin relaxation in pulmonary artery of heartworm-infected dogs. Alternatively, the mechanism of relaxation of pulmonary artery from heartworm-infected dogs may be different than control. Our bradykinin data, where magnitude and mechanism of relaxation of heartworm-infected and control animals are the same, suggest that the mechanism of relaxation is also the same.

Pulmonary hypertension and right-sided heart failure could clearly result from pulmonary artery obstruction caused by adult parasites and/or intimal proliferation; however, it is difficult to ascribe more subtle cardiopulmonary changes to obstructive lesions. Decreased exercise performance seen in heartworm-infected racing greyhounds has traditionally been attributed to physical obstruction; however, decreased exercise performance has been noted in dogs with one or two adult parasites (3), a number insufficient to obstruct pulmonary outflow. In dogs naturally infected with heartworms, there is no correlation between the number of heartworms and pulmonary vascular resistance, which further supports our contention that host-parasite interactions, rather than worm burden alone, are important in the pathology associated with heartworm infection (6). Furthermore, beagles infected by surgical transplantation of 14 adult parasites rapidly developed elevated pulmonary vascular resistance when exercised daily. This increase in pulmonary vascular resistance was of greater magnitude and occurred more rapidly than in cage-confined beagles infected by surgical transplantation of 50 adult parasites. Thus mechanisms other than physical obstruction by adult parasites or the host's immunological response to the parasites (25, 31-33) must be operative. Filarial factors and filarial-induced inflammatory mediators may explain, in part, the pathology associated with filariasis and the wide spectrum of clinical disease. Immune mechanisms, physical obstruction, and alteration of endothelial cell function by filarial parasites do not have to be mutually exclusive contributors to the disease process. In any one patient, the degree to which each of these factors leads to the development of pathology may vary, and pharmacological interventions may need to be tailored appropriately.

ACKNOWLEDGEMENTS

We thank Jim Evans, Scot Crandall, Kathy Campbell, and S. Ptu for technical assistance and Greg Fink for his help in statistical analysis of the data.

FOOTNOTES

Address for reprint requests: L. Kaiser, Dept. of Physiology, Michigan State University, East Lansing, MI 48824-1101.

Received 24 January 1996; accepted in final form October 7, 1996.

REFERENCES