(Circulation. 1997;95:1806-1812.)
© 1997 American Heart Association, Inc.
Articles |
the Clinica Medica Generale e Cardiologia, Florence, Italy.
Correspondence to Prof G.G. Neri Serneri, Clinica Medica Generale e Cardiologia, Viale Morgagni, 85, 50134 Florence, Italy.
| Abstract |
|---|
|
|
|---|
Methods and Results Twenty-nine patients with proven unstable angina, 36 with stable angina, and 30 healthy subjects were studied. Both early and short-lived (interleukin-2 receptor [IL-2R],
-chain CD25, and transferrin receptor CD71) and late antigen (HLA-DR) expression were investigated by flow cytometric analysis. Soluble IL-2R (sIL-2R) was also measured in plasma by ELISA. Lymphocyte activation was studied at day 1 of hospital admission and after 7, 15, 30, 60, and 90 days. In patients with unstable angina, the number of HLA-DR+ CD3 lymphocytes and levels of sIL-2R were higher (P<.001) than in patients with stable angina and control subjects. Both CD4+ and CD8+ lymphocytes expressed HLA-DR antigens. No differences were found among the different groups of subjects in regard to the expression of CD25 and CD71. Lymphocyte activation was more marked in patients with urgent revascularization. No relationships were found between the number of HLA-DR+ lymphocytes and either the severity of coronary angiographic lesions or the number of ischemic episodes. Observations over time showed a gradual decrease in the number of HLA-DR+ lymphocytes and sIL-2R levels from weeks 3 through 8 to 12.
Conclusions The present results indicate that (1) CD4+ and CD8+ circulating lymphocytes are activated in patients with unstable angina, and their activation state lasts 6 to 8 weeks; and (2) activation of lymphocytes is not a consequence of myocardial ischemia. These results support the immune systemmediated inflammatory nature of unstable angina.
Key Words: angina interleukins lymphocytes immune system coronary disease
| Introduction |
|---|
|
|
|---|
| Methods |
|---|
|
|
|---|
|
|
Unstable angina was defined as chest pain occurring at rest or on minimal effort (washing, speaking, hair combing) without any increase in the creatine kinase MB fraction, with ECG evidence of myocardial ischemia (transient ST-segment displacement >0.1 mV during chest pain), and with angiographic evidence of coronary artery disease. All patients with unstable angina were continuously monitored with a Holter monitor for the first 3 days.
Experimental Procedure
Citrated blood, analyzed within 1 hour, was centrifuged to remove platelet-rich plasma and resuspended in PBS. Dual-color immunophenotyping of lymphocytes was performed with the following murine MAbs conjugated with FITC: anti-CD3 (Dako), anti-CD4, anti-CD8, anti-CD20, and anti-CD14; and murine MAbs conjugated with PE: antiHLA-DR (Ortho Diagnostic Systems) and antiIL-2R
-chain (CD25) (Dako). To evaluate transferrin receptor membrane expression, an unconjugated antibody anti-transferrin receptor (CD71) stained with a second rabbit anti-mouse immunoglobulin conjugated to PE (Dako) was used. Isotype-matched control antibodies (goat anti-mouse IgG FITC and goat anti-mouse IgG PE) were used in each experiment to set the analysis gate to exclude nonspecific binding. An aliquot of 100 µL of whole blood was added to 10 µL of each monoclonal antibody: 10 µL FITC-conjugated MAbs and 10 µL PE-conjugated MAb. Blood samples incubated at 4°C for 30 minutes in the dark and the stained cells were washed in PBS. After red-cell lysis, samples were centrifuged, and then the pellet was resuspended with PBS.
Flow cytometric analysis was performed on Cytoron Absolute (Ortho Diagnostic Systems) equipped with a 15-MW air-cooled 488 argon-ion laser. Forward and side were used to identify the lymphocyte population and to gate out other cells and debris. A minimum of 2000 events were collected for sampling in the selected gate. In fluorescence analysis, electron compensation was used to remove spectral overlaps. The cells with fluorescence greater than channel 50 (of 256 channels) were considered positive. With these analysis settings, cells stained with isotype-matched control MAbs were <1%. All data are presented as percent positive cells with isotype-matched control MAb staining subtracted out.
To evaluate the reproducibility of the method, CD3, CD4, CD8, and CD3 HLA-DR antigen expressions by lymphocytes were measured in blood samples from 5 healthy control subjects. The intratest coefficients of variation determined by analysis of repeated (sixfold) samples from the 5 subjects were, for each antigen, 2.1%, 2.8%, 3.3%, and 6%, respectively. In 5 control subjects, blood sampling and flow-cytometric analysis were taken on 4 consecutive days to evaluate day-to-day variation.
sIL-2R was measured by commercial assay (Immunotech SA) (the sensitivity was 5 pmol/L).
Coronary Angiography
Coronary angiography was performed by the Judkins technique. The occurrence and severity of coronary angiographic lesions were evaluated from at least three projections by the score suggested by the American Heart Association.22
Statistical Analysis
Unless otherwise indicated, the results are given as median and range. The nonparametric Kruskal-Wallis test for one-way ANOVA (H test) followed by the post hoc test was used to investigate differences among the three different groups at each time. For the analysis of follow-up in patients who completed the protocol, the Friedman test was used. The Mann-Whitney test for unpaired data was used for comparisons of baseline data between subgroups of unstable angina patients (patients with revascularization in the first week versus others and patients with complete follow-up versus others). For correlation analysis, Spearman's rank correlation coefficient was used. All probability values reported are two-tailed, with values of P<.05 considered statistically significant. The confidence limits (95%) were calculated by adding 2 SD to the mean value.
| Results |
|---|
|
|
|---|
Lymphocyte Activation
Day-to-day variations in the expression of CD3, CD4, CD8, and CD3 HLA-DR antigens by lymphocytes, measured in 5 healthy control subjects on 4 consecutive days, were 1.8%, 2.5%, 3%, and 5%, respectively, for each lymphocyte subtype.
There was no significant difference among unstable angina patients, stable effort angina patients, and control subjects in the number of blood lymphocytes or monocytes (Table 2
). The percentages of CD3-, CD4-, and CD8-positive circulating lymphocytes in healthy control subjects, in patients with unstable angina, and in patients with stable effort angina, respectively, were reported in Table 2
. The percentage of CD3-positive lymphocytes from unstable angina patients that expressed little or no transferrin receptor (CD71) was 0.5%, a value similar to those observed in control subjects and in patients with stable effort angina (Table 2
). In patients with unstable angina, the percentage of peripheral CD3-positive lymphocytes that expressed IL-2R
-chain (CD25) on their surface membranes did not differ significantly from that of control subjects and from that of patients with stable effort angina (Table 2
). However, when IL-2R was measured as the circulating form, levels of sIL-2R (Fig 1
) were significantly (P<.001) higher in the group of patients with unstable angina (138 pmol/L; range, 88.4 to 217.6 pmol/L) than in both control subjects (89.4 pmol/L; range, 70.5 to 105.6 pmol/L) and patients with stable effort angina (100.6 pmol/L; range, 78.1 to 113.2 pmol/L). Twenty-three of 29 patients with unstable angina had sIL-2R values over the confidence limits of control subjects (112.4 pmol/L), whereas only 2 patients with stable angina had sIL-2R values over the confidence limits of control subjects (Fig 1
).
|
In patients with active unstable angina, the percentage of HLA-DR+ CD3 lymphocytes was, on average, significantly more elevated (P<.001) than in patients with effort stable angina (4%; range, 1.4% to 5.8%) and control subjects (2.75%; range, 1.3% to 4.5%) (Fig 2
). In all but 2 patients with unstable angina, the percentage of HLA-DR+ CD3 lymphocytes was out of the range of the control subjects (4.5%). At variance, only 11 of 36 patients with effort angina had a percentage of HLA-DR+ lymphocytes over the confidence limits of the control subjects (Fig 2
).
|
Both CD4 and CD8 lymphocytes expressed HLA-DR antigens. More precisely, in patients with unstable angina, the percentage of CD4 lymphocytes HLA-DR+ was 10.5% (range, 3.7% to 17.0%), ie, a 2.9-fold increase compared with control subjects (3.6%; range, 1.2% to 6.1%; P<.001), whereas in patients with stable effort angina, only a slight increase was observed (5.5%; range, 2.3% to 9.8%; P<.01). The percentage of CD8 lymphocytes HLA-DR+ was similarly increased (9.7%; range, 2.9% to 25.0%) and was significantly different (P<.001) from control subjects (3.7%; range, 0.7% to 6.0%) and patients with stable effort angina (4.8%; range, 1.5% to 9.5%) (Fig 2
).
During the first week of observation, 8 patients underwent coronary revascularization because of the severity of refractory angina. These patients had higher percentages of HLA-DR+ lymphocytes (9.1%; range, 6.2% to 12%) and higher serum levels of sIL-2R (152.4 pmol/L; range, 116.7 to 217.6 pmol/L) than the other unstable angina patients (HLA-DR+: 7%; range, 2.4% to 12%; P<.05; sIL-2R: 130 pmol/L; range, 88.4 to 206.5 pmol/L; P<.05) (Table 3
). Conversely, the 8 patients who have concluded the follow-up had a lower percentage of HLA-DR+ lymphocytes at the first observation (5.4%; range, 2.4% to 6.5%) and lower serum levels of sIL-2R (110.3 pmol/L; range, 100 to 138.4 pmol/L) than the remaining patients (HLA-DR+: 8.3%; range, 4.65% to 12%; P<.001; sIL-2R: 148.8 pmol/L; range, 88.4 to 217.6 pmol/L; P<.01) (Table 3
).
|
No significant relationships were found between either the serum levels of sIL-2R or HLA-DR+ lymphocytes and the severity of coronary angiographic lesions (r=.10, P=.87 and r=.40, P=.07, respectively) or the time between the onset of symptoms and the first blood sampling (r=.03, P=.87 and r=.05, P=.84, respectively). Likewise, there was no relationship between the number of ischemic episodes and lymphocyte activation at admission (sIL-2R: r=.25, P=.42; HLA-DR+ lymphocytes; r=.21, P=.42).
Follow-up of Lymphocyte Activation
All the control subjects and patients with stable effort angina (data not shown) who were studied during the follow-up period (7, 15, 30, 60, and 90 days) did not show significant changes in the percentage of HLA-DR+ CD3 lymphocytes compared with the first observation. Conversely, consistent variations were observed in patients with unstable angina (Table 3
). All the 21 patients restudied after 7 days had values of HLA-DR+ CD3 lymphocytes over the confidence limits of control subjects. The median value of the percentage of HLA-DR+ CD3 lymphocytes was 7.5% (range, 4.6% to 9.5%), and in 13 of 21 patients, the percentage of HLA-DR+ CD3 lymphocytes was increased compared with the first examination. It is worth stressing that all these patients but one were completely free of ischemic episodes (Holter monitoring). After 15 days, the number of patients over the confidence limits of the control subjects was 15 of 18, and it was 9 of 12 at 30 days, 3 of 9 at 60 days, and 1 of 8 at 90 days. In the 8 patients whose follow-up was 90 days, the percentage of HLA-DR+ CD3 lymphocytes and the levels of sIL-2R in serum increased from baseline to day 7, then progressively decreased (Fr=12.2, P<.05; and Fr=22.7, P<.01). The time course of CD3+ DR+ shows that the percentage of activated lymphocytes at 30, 60, and 90 days decreased compared with the maximum of activation (which was often reached at 7 days) in
90% of patients. More precisely, the decrease was evident in 10 of 12 patients at day 30, in 8 of 9 at day 60, and in 7 of 8 at day 90 (Table 3
). The time course of sIL-2R was similar (Table 3
).
| Discussion |
|---|
|
|
|---|
Lymphocyte Activation
In a previous investigation,10 we showed that lymphocytes were necessary to induce tissue factor expression by monocytes from patients with unstable angina, thus indirectly suggesting a possible lymphocyte activation. In the present study, we provide direct evidence both that circulating lymphocytes from unstable angina patients are actually activated and that an immunological reaction precedes the occurrence of angina. Moreover, a relationship seems to exist between the intensity of the immunological reaction and the clinical outcome.
Data from this study indicate that both CD4+ and CD8+ circulating lymphocytes from patients with active unstable angina but not from patients with stable effort angina have an enhanced expression of HLA-DR antigens, thus indicating that T lymphocytes are activated during the acute phase of unstable angina. The activation of T cells is further supported by the elevated serum concentration of the sIL-2R. In the majority of patients (13 of 21 and 14 of 21, respectively), the number of lymphocytes expressing HLA-DR antigens and the plasma levels of sIL-2R were higher at day 7 than at day 1. This further increase in the markers of immune activation probably reflects a progression in the activation of an immunological mechanism that in some patients is not at maximum at the presentation of clinical symptoms. Variations in the rises of serum sIL-2R have been reported in diseases in which immunological activation plays an important role in the pathogenesis, such as rheumatoid arthritis and Crohn's disease.27 28 29 The peak values of the markers of immune activation do not necessarily coincide with the clinical exacerbation of symptoms, especially in diseases such as unstable angina in which the medical treatment is usually very effective.
The number of lymphocytes that expressed IL-2R and transferrin receptor was not significantly different between patients with unstable angina and control subjects or patients with stable effort angina. This apparent discrepancy between the expression of HLA-DR antigen and the absence of expression of IL-2R and transferrin receptor might be due to the transient expression of these receptors on the surface membrane of activated cells.30 Namely, IL-2R and transferrin receptors are early and very short-lived activation markers of T cells, because they are expressed within 2 to 24 hours after stimulation and persist for only a few days.31 32 HLA-DR antigens are expressed several days after stimulation, when DNA synthesis is initiated and cell proliferation occurs, and remain expressed for several weeks.31 32 33 Thus, the absence of expression of IL-2R and transferrin receptor in the presence of HLA-DR antigen expression seems to indirectly suggest that the immunological reaction responsible for the lymphocyte activation preceded the clinical presentation of angina. However, only the results obtained by a specifically designed prospective study can demonstrate this issue. The sIL-2R remains in serum much longer than IL-2R expressed on lymphocytes and represents a sensitive and long-lived marker of circulating lymphocyte activation.30 34
Clinical outcome seems to be related to the intensity of the immunological reaction. Patients who needed revascularization procedures during the first week of the observation period had, as a group, a more marked immunological reaction as evaluated by the significantly higher percentage of circulating HLA-DR+ lymphocytes and higher serum levels of sIL-2R than the other patients with unstable angina. Conversely, the 8 patients who completed the follow-up had a low percentage of HLA-DR+ CD3 lymphocytes and low serum levels of sIL-2R at the first observation. Observation over time of this last group of patients seems to suggest that the acute immunological reaction wanes in about 6 to 8 weeks. However, because of the small number of patients, the time course of lymphocyte activation and especially the relationship between intensity of the immunological reaction and clinical outcome should be considered with caution, and further studies are needed.
In contrast to other clinical conditions, such as bronchial asthma, in which only CD4 lymphocytes were activated,35 we found activation of both CD4 and CD8. This different activation most likely reflects a difference in the factor(s) responsible for lymphocyte activation in the two conditions. The CD4 and CD8 activation in unstable angina is consistent with the observation that the activated lymphocytes found in human atherosclerotic plaques are both CD4 (helpers) and CD8 (mainly suppressors) of polyclonal origin.36 37 38
Recently, normalization of sIL-2R levels in plasma has been reported after percutaneous transluminal coronary angioplasty in patients with stable angina.39 Unfortunately, in the present study we did not investigate the changes of lymphocyte activation after revascularization procedures.
Significance and Mechanisms of the Lymphocyte Activation
Lymphocyte activation during the active phase of unstable angina did not appear to be an epiphenomenon of the angina, because it is also demonstrable in patients temporarily without anginal attacks because of antianginal treatment. Moreover, no relationship was found between the number of ischemic episodes and the number of lymphocytes that expressed surface activation markers. The progressive disappearance of activation antigens, which were no longer detectable 6 to 12 weeks after the beginning of clinical symptoms, indicates that lymphocyte activation is a distinct feature of the active phase of unstable angina and not a consequence of coronary artery disease per se. The existence of an acute inflammatory state in unstable angina is supported by other clinical studies, including elevated levels of C-reactive protein in serum,11 40 activation of circulating monocytes resulting in increased tissue factor10 12 and thromboxane A2 formation,13 increased urinary excretion of various leukotrienes,41 and increased expression of CD11/CD18 adhesion molecules on monocytes and granulocytes from coronary sinus blood. Recent histopathological studies have demonstrated that abundant inflammatory infiltrates formed by monocyte-macrophages and lymphocytes, which express activation antigens, are characteristically present in patients with unstable angina or myocardial infarction.25 26 42 43 44 In the present study, systemic markers of lymphocyte activation were investigated, but they may not necessarily overlap completely with the situation in the atherosclerotic plaque.
The presence of activated T lymphocytes in unstable angina patients implies antigenic stimulation, but the nature of such antigen(s) remains to be investigated. The fact that circulating activated T lymphocytes are detectable only for a few weeks suggests an acute formation of antigen(s) able to cause a transient activation. Many possible candidate antigens may be advocated.20 LDLs undergo oxidative modification in vivo,45 46 and they have been found to be able to activate lymphocyte cocultures in vitro.47 Conversely, some recent studies in three different strains of mice (C57 BL/6, C3H/HeJ, and BALB/c) show that the inflammatory response to oxidized LDL is controlled by the induction of a set of genes.48 49 Therefore, we can speculate that formation of neoantigens arising from lipoprotein oxidation in the plaque may induce lymphocyte activation only in a subset group of patients with coronary atherosclerosis.
However, precisely designed studies are needed to investigate the mechanism(s) responsible for lymphocyte activation.
Role of the Acute Inflammatory Reaction in the Pathogenesis of Unstable Angina
An important issue to be discussed is whether the acute inflammatory reaction, which differentiates unstable angina from stable effort angina, is only a marker of the instability and represents the vascular response to endothelial disruption and thrombus formation, or it represents a primary peculiar coronary pathological process in which lymphocyte and monocyte-macrophage activation, leading to the formation of different inflammatory mediators, causes myocardial ischemia and the occurrence of unstable angina. Inflammatory infiltrates are present in almost all patients with unstable angina,25 26 42 43 44 50 whereas thrombus and/or plaque disruption is demonstrable in
22% to 40% of patients.25 43 44 50 51 This same frequency of coronary thrombus (39%) was reported by a recent angioscopic study in patients with unstable angina.52 Thus, the acute inflammatory reaction that occurs in unstable angina seems more likely to be a primary pathological coronary process rather than the expression of the vascular response to endothelial disruption and thrombus formation. We wish to stress that the development of an acute inflammatory reaction does not negate the prevalent concept that platelet aggregation and thrombus formation play a relevant role in the pathogenesis of unstable angina or that changes in vascular tone contribute to the development of unstable angina.6 On the contrary, the acute exacerbation or activation of an until then chronic, quiet inflammatory process resulting in the formation of different types of adhesion molecules,26 53 inflammatory cytokines,43 54 55 or lytic proteases56 57 can lead to endothelial dysfunction and favor platelet aggregation, even in the absence of endothelial disruption, or compromise the integrity of the arterial wall, with consequent thrombus formation. The occurrence of an acute inflammatory reaction resulting in the formation of various activated factors can explain not only the different pathomorphological lesions found in unstable angina25 43 52 but also several characteristics peculiar to unstable angina, such as the recurrent nature of the disease,4 the poor correlation of the extent of stenosis with the outcome of clinical events,58 the increased sensitivity to vasoconstrictor stimuli,59 and the high incidence of complications after angioplasty or directional atherectomy.60 61 62 63
Thus, the acute inflammatory reaction, usually superimposed on a chronic, inflammatory atherosclerotic lesion, may represent the common ground of both the different morphological aspects and the various mechanisms responsible for the occurrence of the instability of angina.
| Selected Abbreviations and Acronyms |
|---|
|
| Acknowledgments |
|---|
Received August 5, 1996; revision received November 18, 1996; accepted November 22, 1996.
| References |
|---|
|
|
|---|
B-like transcription factor activation in response to an atherogenic diet in mice. J Clin Invest. 1993;91:2572-2579.This article has been cited by other articles:
![]() |
T Adachi, T Naruko, A Itoh, R Komatsu, Y Abe, N Shirai, H Yamashita, S Ehara, M Nakagawa, C Kitabayashi, et al. Neopterin is associated with plaque inflammation and destabilisation in human coronary atherosclerotic lesions Heart, December 1, 2007; 93(12): 1537 - 1541. [Abstract] [Full Text] [PDF] |
||||
![]() |
G. Caligiuri and A. Nicoletti Lymphocyte responses in acute coronary syndromes: lack of regulation spawns deviant behaviour Eur. Heart J., November 1, 2006; 27(21): 2485 - 2486. [Full Text] [PDF] |
||||
![]() |
S Ranjit and L Dazhu Potential role of dendritic cells for progression of atherosclerotic lesions. Postgrad. Med. J., September 1, 2006; 82(971): 573 - 575. [Abstract] [Full Text] [PDF] |
||||
![]() |
R. Sharma and D.-Z. Li Role of Dendritic Cells in Atherosclerosis Asian Cardiovasc Thorac Ann, April 1, 2006; 14(2): 166 - 169. [Abstract] [Full Text] [PDF] |
||||
![]() |
A. Tedgui and Z. Mallat Cytokines in Atherosclerosis: Pathogenic and Regulatory Pathways Physiol Rev, April 1, 2006; 86(2): 515 - 581. [Abstract] [Full Text] [PDF] |
||||
![]() |
G. Caligiuri, E. Groyer, J. Khallou-Laschet, A. A. H. Zen, J. Sainz, D. Urbain, A.-T. Gaston, M. Lemitre, A. Nicoletti, and A. Lafont Reduced Immunoregulatory CD31+ T Cells in the Blood of Atherosclerotic Mice With Plaque Thrombosis Arterioscler. Thromb. Vasc. Biol., August 1, 2005; 25(8): 1659 - 1664. [Abstract] [Full Text] [PDF] |
||||
![]() |
S. Kohsaka, V. Menon, A. M. Lowe, M. Lange, V. Dzavik, L. A. Sleeper, J. S. Hochman, and for the SHOCK Investigators Systemic Inflammatory Response Syndrome After Acute Myocardial Infarction Complicated by Cardiogenic Shock Arch Intern Med, July 25, 2005; 165(14): 1643 - 1650. [Abstract] [Full Text] [PDF] |
||||
![]() |
H. Methe, S. Brunner, D. Wiegand, M. Nabauer, J. Koglin, and E. R. Edelman Enhanced T-Helper-1 Lymphocyte Activation Patterns in Acute Coronary Syndromes J. Am. Coll. Cardiol., June 21, 2005; 45(12): 1939 - 1945. [Abstract] [Full Text] [PDF] |
||||
![]() |
A. Ceriello Acute hyperglycaemia: a 'new' risk factor during myocardial infarction Eur. Heart J., February 2, 2005; 26(4): 328 - 331. [Abstract] [Full Text] [PDF] |
||||
![]() |
A. Ceriello Insulin, Glycemic Control, and C-Reactive Protein During Myocardial Infarction Diabetes Care, December 1, 2004; 27(12): 3017 - 3018. [Full Text] [PDF] |
||||
![]() |
M. Madjid, A. Zarrabi, S. Litovsky, J. T. Willerson, and W. Casscells Finding Vulnerable Atherosclerotic Plaques: Is It Worth the Effort? Arterioscler. Thromb. Vasc. Biol., October 1, 2004; 24(10): 1775 - 1782. [Abstract] [Full Text] [PDF] |
||||
![]() |
E. Falk Widespread Targets for Friendly Fire in Acute Coronary Syndromes Circulation, July 6, 2004; 110(1): 4 - 6. [Full Text] [PDF] |
||||
![]() |
R. Marfella, M. Siniscalchi, K. Esposito, A. Sellitto, U. de Fanis, C. Romano, M. Portoghese, S. Siciliano, F. Nappo, F. C. Sasso, et al. Effects of Stress Hyperglycemia on Acute Myocardial Infarction: Role of inflammatory immune process in functional cardiac outcome Diabetes Care, November 1, 2003; 26(11): 3129 - 3135. [Abstract] [Full Text] [PDF] |
||||
![]() |
O. J de Boer, A. E Becker, and A. C van der Wal T lymphocytes in atherogenesis--functional aspects and antigenic repertoire Cardiovasc Res, October 15, 2003; 60(1): 78 - 86. [Full Text] [PDF] |
||||
![]() |
S.-P. Yang, L.-J. Ho, Y.-L. Lin, S.-M. Cheng, T.-P. Tsao, D.-M. Chang, Y.-L. Hsu, C.-Y. Shih, T.-Y. Juan, and J.-H. Lai Carvedilol, a new antioxidative {beta}-blocker, blocks in vitro human peripheral blood T cell activation by downregulating NF-{kappa}B activity Cardiovasc Res, September 1, 2003; 59(3): 776 - 787. [Abstract] [Full Text] [PDF] |
||||