Published ahead of print on June 16, 2005, doi:10.1165/rcmb.2005-0136OC
© 2005 American Thoracic Society DOI: 10.1165/rcmb.2005-0136OC Diacylglycerol-Containing Docosahexaenoic Acid in Acyl Chain Modulates Airway Smooth Muscle ToneDépartementPhysiologie, Immunologie et Neurosciences, Unité Propre de Recherche de l'Enseignement SupérieurLipides et Nutrition, Faculté des Sciences de la Vie, Université de Bourgogne, Dijon, France; and Le Bilarium, Department of Physiology and Biophysics, Faculty of Medicine, Université de Sherbrooke, Sherbrooke, Quebec, Canada Correspondence and requests for reprints should be addressed to Professeur N. A. Khan, Directeur, DépartementPhysiologie, Immunologie et Neurosciences, UPRES Lipides et Nutrition, Faculté des Sciences de la Vie, 6 Boulevard Gabriel, 21000 Dijon, France. E-mail: naim.khan{at}u-bourgogne.fr
We synthesized and assessed the role of a diacylglycerol (DAG)-containing docosahexaenoic acid (DHA), that is, 1-stearoyl-2-docosahexaenoyl-sn-glycerol (SDHG), in the contraction of guinea pig airway smooth muscle (ASM). We compared its action with 1-stearoyl-2-arachidonoyl-sn-glycerol (SAG) and 1,2-dioctanoyl-sn-glycerol (1,2-DiC8), a stable DAG analog. The three DAGs (SAG, SDHG, and 1,2-DiC8) induced reversible concentration-dependent contraction of ASM. SDHG induced higher guinea pig ASM contraction than did SAG and 1,2-DiC8. The effects of SDHG were blocked, to different extents, by nifedipine (L-type Ca2+ channel blocker). By employing GF-109203X (protein kinase C [PKC] inhibitor) and lanthanum (La3+), a nonselective cation channel blocker, we observed that SDHG evoked ASM contractile response via PKC-dependent and PKC-independent (but Ca2+-dependent) pathways. Interestingly, SAG exerted its action only by increasing [Ca2+]i and did not require PKC activation. To probe the implication of calcium mobilization, we employed thapsigargin (TG), which also induced ASM contraction in a calcium-dependent manner. SDHG and 1,2-DiC8, in a PKC-dependent manner, induced the phosphorylation of CPI-17 (myosin light chain phosphatase inhibitor of 17 kD). Furthermore, SAG and TG failed to phosphorylate CPI-17 in ASM cells. Our results suggest that different DAG species, produced during a dietary supplementation with fatty acids, could modulate the reactivity of airway smooth muscles in a PKC-dependent and -independent manner, and hence, may play a critical role in health and disease.
Key Words: 1,2-dioctanoyl-sn-glycerol calcium entry docosahexaenoic acid isometric tension protein kinase C
Airway smooth muscle (ASM) cells represent a significant proportion of all cells present in the pulmonary airways (1). Traditionally, the role of the ASM cells in airway inflammation has been regarded to be passive, contributing to the pathogenesis of asthma solely by their contractile properties (1). Several reports have shown that in addition to their contractile functions in asthma, ASM cells can undergo hyperplasia and/or hypertrophy (1, 2), leading to structural changes in the airway wall that contribute to the development of persistent airway obstruction and increased nonspecific airway hyperresponsiveness (3, 4). Additional reports from cell culturebased studies are emerging to suggest a further role for ASM in airway inflammation by acting as an important source of proinflammatory and bronchoprotective mediators (5, 6). This apparent functional diversity of ASM has prompted interest in the possibility that there is plasticity in its function that may be related to the severity of tissue remodeling process during chronic inflammation of the airway wall (6, 7). The mechanisms responsible for bronchial hyperreactivity are believed to be caused by an exaggerated response of immune system, which leads to the release of histamine, neuropeptides, and arachidonic acid (AA) metabolites like leukotrienes (LTs) and prostaglandins (PGs) from ASM cells (810).
Over the past 30 yr, there has been a considerable interest in the therapeutic potential of fish oils in various inflammatory conditions (11). Fish oils, rich in
At the cellular and molecular level, acute regulation of ASM contraction and relaxation is typically mediated by G proteincoupled receptors (19). In the case of bronchoconstrictor agonists, receptor activation elicits phosphatidylinositol turnover that results in the formation of 1,2-diacylglycerol (DAG), which activates protein kinase C (PKC), and inositol 1,4,5,-trisphosphate, which binds to its intracellular receptors to mobilize intracellular calcium (20, 21). Contractile tone triggered by intracellular Ca2+ release is maintained by Ca2+ entry from the extracellular medium, mediated via the activation of voltage-dependent L-type calcium channels and via nonspecific cation channels (21). The intracellular free calcium binds to calmodulin, which activates the phospholipase A2, PLA2 (20). Hence, PLA2 will liberate, from the sn-2 position, DHA, which will exert myorelaxing effects (17). However, DHA may also be liberated in the conjugated form as a DAG during receptor activation in ASM cells (19). Does DAG-containing DHA exert a relaxing or a contractile effect? If DAG-containing DHA exerts opposite effects to free DHA, it may explain, in part, the ambiguous results obtained on
Chemicals 1-stearoyl-2-arachidonoyl-sn-glycero-3-phosphocholine, 1-stearoyl-2-docosahexaenoyl-sn-glycero-3-phosphocholine, phospholipase C from Bacillus cereus, carbamylcholine (CCh), nifedipine, DHA, AA, phorbol 12-myristate 13-acetate (PMA), thapsigargin (TG), DAG-lipase inhibitor (RHC 80267), and GF-109203X were purchased from Sigma (St. Louis, MO). Lanthanum chloride (LaCl3) was purchased from ICN Biomedicals Inc. (Cleveland, OH). AntiCPI-17 and antiphospho-CPI-17 antibodies were purchased from Upstate Signaling (Waltham, MA). DAG-kinase inhibitor (R59949) was procured from Calbiochem (Fontenay-sous-Bois, France).
Synthesis of Different DAG Molecular Species
Bronchial Smooth Muscle Strip Preparation and Tension Measurements
ASM Cell Culture and Measurement of Ca2+ Signaling The ASM cells were washed with NaCl/Pi (phosphate-buffered saline), pH 7.4. The composition of NaCl/Pi was as follows: 3.5 mM KH2PO4; 17.02 mM Na2HPO4; 136 mM NaCl. The cells were then incubated with Fura-2/AM (1 µM) for 60 min at 37°C in loading buffer that contained the following: 110 mM, NaCl; 5.4 mM, KCl; 25 mM, NaHCO3; 0.8 mM, MgCl2; 0.4 mM, KH2PO4; 20 mM, Hepes-Na; 0.33 mM, Na2HPO4; 1.2 mM, CaCl2, and the pH was adjusted to 7.4. After loading, the cells (2 x 106/ml) were washed three times (2,000 x g, 10 min) and remained suspended in the identical buffer. [Ca2+]i was measured according to Grynkiewicz and colleagues (29). The fluorescence intensities were measured in the ratio mode in PTI spectrofluorometer at 340 nm and 380 nm (excitation filters) and 510 nm (emission filters). The cells were continuously stirred throughout the experimentation. The test molecules were added into the cuvettes in small volumes with no interruptions in recordings. The intracellular concentration of free Ca2+, [Ca2+]i, were calculated by using the following equation: [Ca2+]i = Kd x (R Rmin)/(Fmax F) x (Sf2/Sb2), where Kd is the dissociation constant for Fura-2/calcium complex, R is the ratio emission with excitation at 340 nm divided by excitation at 380 nm, Rmin is the ratio in the presence of no Ca2+, Fmax is the ratio of saturating [Ca2+]i, and Sf2/Sb2 is the ratio of 380 nm excitation fluorescence at zero and saturating [Ca2+]i. A value of 224 nM for Kd was added into the calculations. Rmax and Rmin values were obtained by addition of ionomycin (5 µM) and MnCl2 (2 mM), respectively. All the experiments were performed at 37°C. We designed a Ca2+-free/Ca2+-reintroduction (CFCR) protocol for the experiments (30), conducted in Ca2+-free (0% Ca2+) medium. Hence, we examined the role of different DAG molecular species, DHA, and CCh on direct calcium influx. First test agents and then CaCl2 were added into the cuvette.
Smooth Muscle Tissue Preparation and Measurement of CPI-17 Phosphorylation
Statistical Analysis
SDHG and SAG Induce ASM Contraction In guinea-pig bronchi, exogenous addition of cumulative concentrations of SAG, SDHG, and 1,2-DiC8 produced increases in tension (Figures 1A1C). The effects were fully reversible upon washing with freshly oxygenated (95% O2, 5% CO2) Kreb's solution. However, the reversibility was slow and chaotic for the 1,2-DiC8 compound. At the dilution used, the vehicle (ethanol, 0.1% vol/vol) exerted no effect on the resting tone of guinea pig ASM (data not shown). Figure 1D shows the comparative responses for identical concentrations of different DAG molecular species. The ability of DAGs to modify guinea-pig ASM tone seems to be dependent on the nature of PUFA (AA and DHA) present in sn-2 position. Hence, 1,2-DiC8, at 0.1 and 0.3 µM, seems to be the least potent inducer of ASM contraction, whereas SDHG exhibited higher effects than SAG at all the concentrations tested (Figure 1D). The EC50 values for SAG, SDHG and 1,2-DiC8 are, respectively, 1.5 ± 0.05 µM, 1 ± 0.2 µM, and 1.2 ± 0.08 µM.
DHA Alone Induces ASM Relaxation As SDHG is a potent substrate for DAG-lipase, we hypothesized that the contraction induced by this DAG might be mediated by free DHA. Hence, we performed complementary experiments, using DHA as free fatty acid, to delineate its putative effects from those of the corresponding SDHG. Figure 2 shows that DHA, in place of evoking a contractile response, induced concentration-dependent relaxations of the basal tone, initially adjusted to 1 g. Hence, relaxing response was amplified when the resting tone was adjusted to 2 g (Figure 2). EC50 for relaxation values of DHA is 0.22 ± 0.02 µM. AA alone failed to evoke relaxation (or contraction) in guinea-pig ASM tone (n = 5, 0.011 ± 0.01 as % tension induced by CCh). Interestingly, linoleic acid and linolenic acid evoked neither contraction nor relaxation in guinea-pig ASM tone (n = 5, results not shown). Furthermore, DHA exerted no effects on the tonic responses induced by CCh (results not shown).
We also conducted experiments by 15-min pretreatments of guinea-pig bronchi with R59949 at 1 µM (31), which inhibited DAG-kinase by binding to the catalytic domain, and RHC 80267 at 50 µM (31, 32), a selective inhibitor of DAG-lipase activity, before DAG stimulation, and we observed that these inhibitors failed to diminish the action of exogenous DAGs (1 µM) on the contraction of smooth muscles (SAG: control, 15.33 ± 0.43 versus with DAG-lipase inhibitor, 14.6 ± 0.45 versus with DAG-kinase inhibitor, 14.3 ± 0.56; SDHG: control, 40.5 ± 2.6 versus with DAG-lipase inhibitor, 39.3 ± 2.55 versus with DAG-kinase inhibitor, 38.3 ± 2.50; results are expressed as % tension induced by CCh).
DAG-Induced Influx of Extracellular Ca2+ Is Involved in ASM Contraction
The efficiency of DAGs on the ASM contraction was in the following order: SDHG > 1,2-DiC8 > SAG while the recordings were performed in calcium-added medium (Table 1). The difference between 1,2-DiC8 and SDHG was not significant while the recordings were conducted in calcium-free medium (Table 1). The contractions were reversed, during washing, upon removal of DAGs and Ca2+ from the physiologic solution.
To probe the contribution of calcium influx in guinea-pig ASM contraction, we further employed TG, which inhibits Ca2+-ATPase of endoplasmic reticulum and provokes calcium influx by depleting intracellular calcium pools (33). Figure 3D shows that TG failed to induce contractions of ASM in the absence of calcium in the extracellular medium (0% calcium), whereas in calcium added-medium this agent induced a significant response, which was higher than that induced by DAGs (Table 1). To investigate the contribution of calcium influx in DAG-induced ASM contraction, we used CFCR protocol. We observed that addition of the three DAG species to 0% Ca2+ medium did not evoke any increase in [Ca2+]i (Figure 4). However, addition of DAG species in 0% Ca2+ medium, followed by addition of CaCl2 into the cuvette, induced calcium influx in ASM cells (Figure 4). CCh also induced a significant response on the increases in [Ca2+]I, whereas DHA did not evoke a significant increase in [Ca2+]i (Figure 4, insert).
SDHG-, but Not SAG-, Stimulated ASM Tone Is Partially PKC-Dependent To determine whether the effects of DAGs on ASM could be mediated via PKC activation (24), we examined their effects in the absence or presence of GF-109203X, a potent PKC inhibitor that interacts with ATP binding to the enzyme (34). GF-109203X inhibits both calcium-dependent and calcium-independent isoforms of PKC (35). Incubation of ASM strips for 15 min with GF-109203X reduced the action of SDHG and 1,2-DiC8. GF-109203X was more effective on 1,2-DiC8 than on SDHG (Figure 5). Surprisingly, SAG-induced tension was poorly sensitive to GF-109203X pretreatment (Figure 5).
SDHG Acts via Calcium and PKC, Whereas SAG Acts via Calcium and Independently of PKC Activation on ASM Tone Nifedipine, a Ca2+ channel blocker, was used to evaluate the contribution of L-type calcium channels (26). Nifedipine (1 µM) completely relaxed the sustained responses induced by SAG, SDHG, and 1,2-DiC8 (Table 2). We also employed La3+ at low concentration (3 µM) to block partially the calcium channels (30) and to evaluate the relative contribution of PKC activation in DAGs-induced guinea-pig ASM contraction. La3+ at 30 µM inhibits completely the opening of nonselective action and L-type calcium channels, whereas at low concentrations (3 µM), this agent will partially inhibit these calcium channels (36).
We observed that La3+ at 3 µM exerted a partial inhibitory effect on tonic responses (Table 2). In the presence of La3+ and SAG, addition of GF-109203X failed to further stimulate the ASM tone relaxation as compared with assays in the absence of the PKC inhibitor (Table 2). However, addition of GF-109203X to assays containing SDHG and 1,2-DiC8 potentiated the La3+-induced relaxation (Table 2).
PKC Inhibition in Calcium-Free Medium Completely Abolished the Action of SDHG, but Not of SAG, on ASM Contraction
While the experiments were performed in calcium-containing medium (100% calcium), addition of SDHG before or after TG exerted an additive response on ASM contractions (Figures 7A and 7B). It is noteworthy that addition of SAG after TG failed to induce an additive response (Figure 7D). SAG alone exerted a usual response, and addition of TG to these assays further evoked an additional response on ASM contractions (Figure 7C).
SDHG, but Not SAG, Induces the Phosphorylation of CPI-17 in a PKC-Dependent Manner CPI-17 represents a novel smooth musclespecific regulatory protein, and its phosphorylation depends on PKC activation (37). Figure 8 shows that CPI-17 protein is phosphorylated by PMA, SDHG, and 1,2-DiC8, but not by SAG, in ASM cells. Furthermore, the action of PMA, SDHG, and 1,2-DiC8 on CPI-17 phosphorylation is abolished by a pretreatment with GF-109203X. Interestingly, TG failed to induce the phosphorylation of CPI-17 protein in ASM cells (Figure 8).
It has been reported that fish oils, containing DHA, exert either beneficial effects (1316) or no improvements in chronic obstructive pulmonary diseases (COPD) and asthma (10, 13, 38). Hence, we have proposed a hypothesis which states that the myorelaxing effects of this agent may be due to its free fatty acid form, and that contractile actions may be contributed by its conjugated DAG form. Several studies have shown that myorelaxing effects are evoked by DHA (1316); however, nothing is known on the contractile effects of this -3 fatty acid. Indeed, fish oil diets result in several fold increases in phospholipids and, consequently, DAG levels, containing DHA (39, 40). We have also recently reported that in vitro incubation of T-lymphocytes with AA or DHA, upon activation, gives rise to the high production of DAG containing these fatty acids (which, in turn, modulate cell signaling via a PKC-dependent and PKC-independent pathways) (41). We therefore assessed the properties of DAG-containing DHA (i.e., SDHG) to induce guinea-pig ASM contraction in comparison to SAG and 1,2-DiC8, a cell-permeant and stable analog of DAG. We observed that the three DAG molecular species induced a concentration-dependent ASM contraction regardless the nature of fatty acid present in the acyl chain of DAG. However, SDHG-induced increase in tension is more important than that induced by SAG and 1,2-DiC8. It has been reported that DAG can be metabolized, in isolated smooth muscle cells from rabbit aorta, by the DAG-lipase, thus leading to the release of free fatty acid (42). To ascertain whether the action of SDHG is mediated via the release of free DHA, we conducted experiments on DHA. Surprisingly, DHA, in place of inducing a contraction in ASM tone, evoked a concentration-dependent relaxation. Furthermore, DHA-induced relaxation was ineffective in CCh precontracted tissues. Our observations on the DHA-induced relaxation can be substantiated by the findings of Morita and coworkers and those of Engler and colleagues (17, 43) who have demonstrated that, DHA as a free fatty acid, induces a direct vasorelaxant response in SHR aorta. However, the vasoralaxant effect of DHA is not related to an alteration in calcium, because DHA failed to induce significant increases in free intracellular calcium in Fura-2loaded smooth cells. It is also noteworthy that free AA completely failed to induce any modifications in guinea-pig ASM tone. We can further rule out the possibility that free AA and DHA, contributed by the actions of DAG-lipase, may be responsible for the actions of DAGs, because RHC 80267, a DAG-lipase inhibitor, evoked no inhibition of SAG- and SDHG-induced contractions of ASM. Similarly, an inhibitor of DAG-kinase, R59949, also did not influence the SAG- and SDHG-induced contractions of ASM. These observations demonstrate that DAGs exert their actions via their molecular configuration (i.e., per se actions). DAGs are known to activate PKC, which may evoke sustained contraction of smooth muscle cells (44). To assess the PKC-mediated actions of SDHG and 1,2-DiC8, GF-109203X, a PKC inhibitor, was employed (24). GF-109203X partially reversed the tonic responses triggered by SDHG and 1,2-DIC8; however, it failed to significantly reverse the SAG-evoked responses. We also conducted experiments in 0% calcium medium containing GF-109203X. We observed that the contractile responses of SDHG and 1,2-DiC8, but not of SAG, were completely blocked by the presence of this PKC inhibitor. These observations suggest that SDHG and 1,2-DiC8 act, in part, by PKC activation, whereas SAG exerts its action independently to PKC activation as far as the ASM tone contraction is concerned. The exact PKC-dependent mechanism of action of SDHG and 1,2-DiC8, implicated in muscle contraction is not known; however, direct phosphorylation of myosin light chains has been described as a link between PKC activation and smooth muscle contraction (4548). Myosin light chain phosphatase (MLCP) plays a critical regulatory role in myosin phosphorylation and smooth muscle contraction. MLCP inhibition leads to an increase in both MLC phosphorylation and contractile force of smooth muscle without any changes in Ca2+ (45). CPI-17, purified as a novel myosin phosphatase inhibitor protein, is phosphorylated by PKC, and thereby inhibits in vitro and in situ MLCP and induces muscle contraction (37, 49). Indeed, we observed that SDHG, 1,2-DiC8, and PMA phosphorylated the CPI-17 and GF-109203X completely abolished their actions on CPI-17 phosphorylation. These observations clearly demonstrate that SDHG and 1,2-DiC8 (also PMA) induce phosphorylation of CPI-17 in a PKC-dependent manner and, consequently, evoke contractions in smooth muscles, as has also been suggested by Eto and coworkers (37, 49). Besides, the involvement of the PKC pathway in response to SDHG and 1,2-DiC8 is consistent with our previous reports in which we have demonstrated that the nature of fatty acid in sn-2 position of DAGs plays an important role in their ability to activate PKC (24, 28). The preceding paragraph shows that SAG exerts its action in a PKC-independent manner. SDHG and 1,2-DiC8 may also act, in part, in a PKC-independent manner, as their effects were not completely abolished by the PKC inhibitor in water bath containing 100% Ca2+. We therefore conducted experiments in the absence and re-addition of calcium into the incubation chamber to elucidate the role of extra- and intracellular calcium concentrations in DAG-induced ASM contraction. We observed that SAG, SDHG, and 1,2-DiC8 triggered transient contractile responses in the absence of extracellular Ca2+, and that addition of exogenous CaCl2 into the assay chamber further evoked and sustained DAG-induced contractile responses, suggesting a Ca2+ entry from the extracellular medium. TG exerted higher ASM contractile response than the three DAGs. TG failed to trigger a contractile response in absence of extracellular Ca2+; however, addition of exogenous CaCl2 into the assay chamber resulted into a sustained ASM contraction, indicating that TG-induced calcium influx, termed store operated calcium (SOC) influx, is participating in the muscle contraction. These observations corroborate the findings of Quinn and colleagues (50), who have recently shown that TG did not induce any contractions in the absence of extracellular calcium in guinea-pig gallbladder smooth muscles, whereas in calcium-containing medium this agent exerted a sustained response. Moreover, TG solely acts via SOC influx and it does not exert its action via CPI-17 as this agent, in our study, failed to induce the phosphorylation of this protein. Furthermore, in calcium-containing medium, addition of TG before or after SDHG exerted an additive response on the contraction of ASM. Although TG evoked a contractile response when this agent was added after SAG, the latter failed to exert an additive response while it was added after the former. These observations suggest that TG-induced SOC influxassociated contractile response overlaps with the SAG-induced response. In fact, TG induces opening of a large number of calcium channels, including calcium-releaseactivated calcium (CRAC) and calcium-releaseactivated nonselective cation (CRANC), including L-type channels (50). Which channels are the ones opened by SAG (also by SDHG and 1,2-DiC8) are not known. However, DAGs do open calcium channels, as is evidenced by our study conducted on Fura-2loaded smooth muscle cells to assess the increases in [Ca2+]i in 0% Ca2+ buffer where exogenous CaCl2 was added into the cuvette. Hence, in 0% Ca2+ buffer, addition of SAG, SDHG, and 1,2-DiC8 alone did not induce any increases in [Ca2+]i in ASM cells. However, addition of the three DAG species, followed by addition of exogenous CaCl2, induced increases in [Ca2+]i, suggesting that DAG species evoke calcium influx in these cells and the PKC-independent sustained responses of DAGs may be contributed by calcium influx in ASM cells. Indeed, a major mechanism of smooth muscle contraction involves membrane potential depolarization and activation of voltage-dependent calcium channels (51). This leads to increased calcium entry and activation of calcium-dependent contractile mechanisms, such as enhancement of MLCK activity (52). It has been proposed that DAGs display Ca2+ releasing properties in various cell types (53). However, the enigma of SAG-induced, PKC-independent, ASM contraction in 0% calcium remains to be resolved. To trace out the route of calcium influx in ASM, we employed nifedipine, an inhibitor of voltage-dependent L-type Ca2+ channels. We observed a large inhibition of ASM contraction, recorded upon nifedipine addition, after DAGs challenges. This observation was unexpected, because in ASM, large nifedipine inhibition has only been reported on KCl-induced tension (54). This activation of voltage-dependent L-type Ca2+ channels suggest that DAGs operate, in part, via the modulation of the membrane potential. This hypothesis is supported by previous report indicating that 1-oleoyl-2-acetyl-sn-glycerol (OAG) depolarizes the plasma membrane of ASM cells (24) and T-lymphocytes (52). Moreover, the tonic responses induced by DAGs were partially inhibited by La3+, a known inhibitor of nonselective cation, including L-type channels. We show categorically that SDHG and DHA act differently; hence, the former induces contraction, whereas the latter evokes relaxation of ASM tone. Because diacylglycerols have been generally considered as PKC activators and it has been difficult to distinguish the biological effects between SDHG and SAG, our study strikingly shows that SDHG acts, in part, via PKC, whereas SAG operates completely in a PKC-independent manner. During PKC-dependent mechanism, the phosphorylation of CPI-17 will take place and the phosphoCPI-17 will inhibit the myosin phosphatase, which will be responsible for ASM contractile response. Furthermore, regardless of the nature of DAG, these agents mediate activation of voltage-dependent L-type Ca2+ channels and, hence, will contribute to guinea-pig ASM contraction.
As far as the physiologic relevance of these observations is concerned, we can state that, during the dietary intake of
We can sum up that the effects of
The authors thank Dr. G. Dupuis for specific advice. Their sincere thanks are due to Prof. M. Bardou (Université de Bourgogne, Faculté de Pharmacie) and his technical staff for extending their laboratory facilities.
This work was supported by a contingent grant from Region Bourgogne to N.A.K. and a CIHR grant MOP-57677 to E.R., who is a National FRSQ Scholar and a member of the Health Respiratory Network of the FRSQ, Canada. Conflict of Interest Statement: None of the authors have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. Received in original form April 12, 2005 Received in final form June 14, 2005
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||