Author + information
- Received July 8, 2008
- Revision received May 13, 2009
- Accepted May 19, 2009
- Published online September 22, 2009.
- Brian C. Jensen, MD⁎,†,
- Philip M. Swigart, MS⁎,
- Marie-Eve Laden, MD⁎,
- Teresa DeMarco, MD†,
- Charles Hoopes, MD‡ and
- Paul C. Simpson, MD⁎,†,⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Paul C. Simpson, VA Medical Center (111-C-8), 4150 Clement Street, San Francisco, California 94121
Objectives The goal was to identify alpha-1-adrenergic receptor (AR) subtypes in human coronary arteries.
Background The α1-ARs regulate human coronary blood flow. The α1-ARs exist as 3 molecular subtypes, α1A, α1B, and α1D, and the α1D subtype mediates coronary vasoconstriction in the mouse. However, the α1A is thought to be the only subtype in human coronary arteries.
Methods We obtained human epicardial coronary arteries and left ventricular (LV) myocardium from 19 transplant recipients and 6 unused donors (age 19 to 70 years; 68% male; 32% with coronary artery disease). We cultured coronary rings and human coronary smooth muscle cells. We assayed α1- and β-AR subtype messenger ribonucleic acid (mRNA) by quantitative real-time reverse transcription polymerase chain reaction and subtype proteins by radioligand binding and extracellular signal-regulated kinase (ERK) activation.
Results The α1D subtype was 85% of total coronary α1-AR mRNA and 75% of total α1-AR protein, and α1D stimulation activated ERK. In contrast, the α1D was low in LV myocardium. Total coronary α1-AR levels were one-third of β-ARs, which were 99% the β2 subtype.
Conclusions The α1D subtype is predominant and functional in human epicardial coronary arteries, whereas the α1A and α1B are present at very low levels. This distribution is similar to the mouse, where myocardial α1A- and α1B-ARs mediate beneficial functional responses and coronary α1Ds mediate vasoconstriction. Thus, α1D-selective antagonists might mediate coronary vasodilation, without the negative cardiac effects of nonselective α1-AR antagonists in current use. Furthermore, it could be possible to selectively activate beneficial myocardial α1A- and/or α1B-AR signaling without causing coronary vasoconstriction.
Adrenergic receptors (ARs) play an important role in coronary arterial blood flow regulation. Coronary α-ARs cause vasoconstriction, whereas β-ARs cause vasodilation. The α1-ARs constrict primarily epicardial coronary arteries and large arterioles, whereas α2-ARs act mostly on the coronary microcirculation (1,2). Stimulation of α1-ARs by endogenous catecholamines produces little constriction of normal coronary arteries (2–4), but causes pronounced vasoconstriction in coronary arteries with atherosclerotic endothelium (1,2,4).
The α1-ARs exist as 3 molecular subtypes, α1A, α1B, and α1D. All 3 subtypes are activated by norepinephrine (NE) and epinephrine but differ in amino acid sequence, tissue expression, and signaling (5). In the mouse heart, cardiac myocytes express the α1A and α1B subtypes, whereas the α1D subtype is functional in coronary arteries (6–8). However, very few data exist on α1-AR subtypes in the human heart. A single small study of post-mortem tissue identified the α1A as the predominant α1-subtype in epicardial coronaries (9). The α1A is also thought to be the predominant α1-AR subtype in the human myocardium, on the basis of messenger ribonucleic acid (mRNA) assay (10). Taken together, these previous results suggest that α1-AR subtype expression is different in the human heart than the mouse heart.
The distribution of cardiac α1-AR subtypes has significant physiological impact in the mouse, where the myocardial α1A and α1B mediate adaptive and beneficial effects, including positive inotropy, physiological hypertrophy, and protection from myocyte death (6,11–13). The coronary α1D mediates vasoconstriction (7,8). In humans, nonselective blockade of all α1-subtypes can be associated with heart failure (14,15). These results and others raise the possibility that the human heart α1A and α1B subtypes should not be blocked and might even be targets for selective agonists to treat myocardial disease (12,16). Thus, it could be significant clinically if human coronary arteries express predominantly the α1D subtype, as in the mouse (7,8), rather than the α1A subtype, as reported previously (9).
In this study, we re-examined the α1-AR subtypes in human epicardial coronary arteries, and measured β-ARs for comparison. Our results show that the α1D is the predominant and functional coronary α1-AR subtype, whereas the α1A and α1B are expressed at very low levels. We contrast this finding with the minimal expression of the α1D in human ventricular myocardium. We also find that α1-AR levels in coronary arteries are approximately one-third the level of β-ARs, most of which are the β2 subtype.
With the approval of the University of California, San Francisco (UCSF) Committee for Human Research and with full informed consent, we obtained heart tissue from transplant recipients or unused donors provided by the California Transplant Donors Network (CTDN).
The heart was explanted after cold cardioplegia, under anesthesia and analgesia with fentanyl, midazolam, rocuronium, and isoflurane at UCSF, and with varied agents at the CTDN hospitals. The explanted heart was placed immediately in ice-cold physiologic solution, epicardial coronary arteries were dissected, cleaned rapidly of fat, flash frozen in liquid nitrogen, and stored at −80°C.
Ribonucleic acid (RNA) preparation
Coronaries were pulverized in a liquid nitrogen-cooled mortar and homogenized (Polytron) in TRIzol reagent (Invitrogen, Gibco BRL, Gaithersburg, Maryland). Myocardium was homogenized directly in TRIzol. The RNA was extracted with chloroform and isopropanol, purified on Qiagen Mini-Prep columns, and treated with DNase (Turbo DNAfree, Ambion, Austin, Texas). We found no significant RNA degradation (Agilent 2100 BioAnalyzer, Palo Alto, California).
Quantitative real-time reverse transcription (RT) polymerase chain reaction
The RT reactions used 1 μg RNA, SuperScript III Reverse Transcriptase (Invitrogen), random hexamers (Invitrogen), and oligo-dT (Roche, Nutley, New Jersey). Quantitative real-time RT polymerase chain reaction (qRT-PCR) was done in triplicate in an ABI PRISM 7900HT Sequence Detection System with 5% of the RT product, primers at 125 nmol/l, and SYBR Green Master (Roche) with ROX reference dye. Data were analyzed with SDS software version 2.3 (Applied Biosystems, Foster City, California).
Relative quantitation of PCR products used the ΔΔCT method, where arbitrary units were 2−ΔΔCT× 1,000, CT = cycles to threshold, and ΔΔCT = (mean target gene CT) − (mean CT of 2 reference genes, β-actin, and TATA-binding protein, for improved accuracy).
Multiple protocols for membrane preparation from single arteries did not yield sufficient protein for reliable binding. Therefore, 15 epicardial coronaries totaling 10.2 g wet weight were pooled from 11 patients, pulverized in a liquid nitrogen-cooled mortar, homogenized in lysis buffer (5 mmol/l Tris-hydrochloride, 5 mmol/l ethylenediaminetetraacetic acid, 250 mmol/l sucrose pH 7.4 plus phenylmethylsulfonyl fluoride), and centrifuged at 1,000 gfor 15 min. The supernatant was saved, and the pellet containing insoluble material was washed in lysis buffer and recentrifuged. The combined supernatants were centrifuged at 100,000 gfor 1 h, and the resulting pellet was homogenized in lysis buffer and centrifuged at 100,000 gfor 1 h. The resulting final membrane pellet, containing both plasma and intracellular membranes, was resuspended (50 mmol/l Tris pH 7.4, 1 mmol/l ethylenediaminetetraacetic acid) and used for α1- and β-AR binding.
The α1-AR saturation binding used 200 μg membrane protein in 1 ml/tube with 3H-prazosin (0.04 to 1.2 nmol/l, PerkinElmer, Waltham, Massachusetts); phentolamine (10 μmol/l, Sigma #P7561) defined nonspecific binding. The β-AR binding used 50 μg membrane protein/tube with 125I-cyanopindolol (CYP) (0.04 to 1.0 nmol/l, NEN Life Sciences, PerkinElmer); L-propranolol (1 μmol/l, Sigma #P8688) defined nonspecific binding (17). The α1-AR subtype proteins were assayed with competition for 3H-prazosin binding (0.5 nmol/l) by BMY-7378 (0.05 nmol/l to 500 μmol/l, Sigma #B134), an α1D-selective antagonist (18). All incubations were 60 min at 30°C. Binding data were analyzed by Prism 4.0b (GraphPad Software, Inc., San Diego, California).
Coronary artery smooth muscle cell culture and immunoblots
Human coronary arteries were digested 20 min in Hanks buffer with collagenase Type II (1 mg/ml, Worthington, Lakewood, New Jersey) and elastase (0.5 mg/ml, Worthington), and intima and adventitia were removed mechanically. Rings of media (approximately 2 mm) were cut free hand and cultured. Other rings were minced and digested for 2 h in collagenase and elastase; enzymes were inhibited with serum; and smooth muscle cells (SMCs) grew out of the minces (19). Clonetics normal human coronary artery SMCs were from Lonza (#CC2583, Walkersville, Maryland). All cells were cultured in Dulbecco's modified Eagle's medium with 10% fetal bovine serum for 8 to 48 h (rings) or 3 to 11 passages (cells). For experiments, cultures were incubated in Dulbecco's modified Eagle's medium without serum with 5 mg/ml bovine serum albumin (Sigma #A7030) for at least 12 h (rings) or 48 h (cells), pretreated without or with α1-antagonists (10 nmol/l BMY-7378; 1 μmol/l prazosin, Sigma #P7791) for 90 min (rings) or 15 min (cells) and treated with α1-agonists (1-200 nmol/l L-NE, Sigma #N5785; 10 nmol/l A61603, Tocris #1052, Ellisville, Missouri), in the presence of the β-AR antagonist L-propranolol (1 μmol/l). After 30 min (rings) or 15 min (cells), homogenates were made in radioimmunoprecipitation assay buffer with protease and phosphatase inhibitors. Ring homogenates were centrifuged at 12,000 rpm for 20 min at 4°C. The supernatant (rings) or total lysates (cells) were used (10 to 20 μg protein/lane) for immunoblotting with antibodies from Cell Signaling for total–extracellular signal-regulated kinase (ERK) 1/2 (rabbit pAb #9102) and phospho-ERK1/2 (rabbit mAb Thr202/Tyr204 #4370) and antibodies for phospho-S19/20-myosin light chain 2 (MLC), including Sigma #M6068, Cell Signaling #3671, and abcam #ab2480.
Results are mean ± SEM. Significant differences (p < 0.05) were tested with analysis of variance and Tukey's multiple comparison for more than 2 groups and Student unpaired ttest for 2 groups. The Ftest was used to compare goodness-of-fit for competition binding analysis (GraphPad Prism version 4.0). Multivariate linear regression (Stata version 9, Stata Corp., College Station, Texas) was used to determine whether clinical variables were independently associated with α1- and β-AR density. Multivariate model assumptions were checked for all regression analyses.
To test α1-AR subtype expression in human coronary arteries, we collected hearts from 19 transplant recipients and 6 unused donors. Average age was 46 years (range 19 to 70 years), and 68% were male (Table 1).Coronary artery disease (CAD) was present in 32%, who were older (p < 0.005) and had higher ejection fractions (EFs) (p < 0.05) (Table 1).
α1-AR subtype mRNA levels
To quantify α1-subtype mRNAs, we validated a qRT-PCR approach, with primers that span the single intron in each α1-AR subtype gene. Primer pairs were designed with Primer3 (version 0.4.0) and BLAST (Basic Local Alignment Search Tool) and chosen for comparable reaction efficiencies (20). Specificity of the α1-subtype primers was confirmed with: 1) PCR with human α1-AR complementary deoxyribonucleic acid; 2) a dissociation step in all PCR reactions; and 3) sequencing of the PCR products. Amplification of genomic deoxyribonucleic acid was excluded by: 1) use of intron-spanning primers; 2) DNase treatment of RNA; and 3) end point PCR reactions with no-RT templates as negative controls.
In human epicardial coronary arteries, the α1D was 85% of total α1-mRNA (Fig. 1A).The α1B (11%) and the α1A (4%) were markedly less abundant than the α1D (p < 0.001 for each). In LV myocardium, by contrast, the α1D was only 21% of the total α1-mRNA, and the α1A (63%) was the most abundant (Fig. 1B). The absolute level of α1D mRNA in coronary arteries was almost twice that in LV (p = 0.01), whereas the absolute level of the α1A was 30-fold higher in LV than in coronary arteries. As a control, there was no difference between coronary arteries and myocardium in the qRT-PCR cycles-to-threshold for the reference genes, β-actin, and TATA-binding protein. Levels of α1-subtype mRNAs were the same in 4 right and 4 left anterior descending coronaries (data not shown). There were no differences in α1-subtype mRNA levels in coronaries collected at UCSF versus the CTDN hospitals, suggesting no important effects due to anesthetic and analgesic agents (data not shown).
In summary, the α1D is 85% of total α1-AR subtype mRNA in human epicardial coronary arteries but is significantly less abundant in human LV myocardium.
α1-AR subtype protein levels
To test α1-AR subtype protein levels we used radioligand binding with 3H-prazosin. We could not use immunohistochemistry or immunoblot, because none of the 10 α1-AR antibodies that we tested is specific for α1-ARs (21).
We used pooled membranes from 11 patients for binding. Patient characteristics were similar in these 11 patients and the entire patient population (Table 1), and the α1-subtype mRNA levels in the pooled samples were similar to the levels of the entire group (data not shown). Saturation binding identified 8.7 fmol/mg protein of total α1-ARs in coronary artery membranes, with a Kd 0.03 nmol/l, and specific binding 70% of total at the 3H-prazosin Kd (Fig. 2).The level of α1-AR binding in coronaries was roughly twice that in LV myocardium (20).
To test whether the α1D subtype protein was also predominant in coronaries, as with mRNA, we did competition binding with the α1D-selective antagonist BMY-7378. The BMY-7378 competition yielded a 2-site binding curve (p = 0.002 vs. 1-site model), with 75% high-affinity sites (Kd 13 pmol/l) and 25% low-affinity (2.6 μmol/l) (Fig. 3A,Table 2).In ventricular myocardium, BMY competition gave a 1-site curve with low BMY-7378 affinity, indicating minimal or no α1D binding (Fig. 3B).
In summary, the α1D is 75% of total α1-AR subtype protein in human epicardial coronary arteries but is much less abundant in myocardium. The coronary levels of α1D mRNA and protein agree very well (Table 2).
α1-AR signaling in coronary SMCs
To test whether the α1D was functional in human coronary SMCs, we used immunoblot to quantify phosphorylation (activation) of ERK, which is a target for the α1D in rat aortic SMCs (22) and is involved in activation of MLC kinase in smooth muscle (23). We used SMCs from Lonza, coronary media rings, and primary isolates from the coronary medial SMC layer (Fig. 4).A qRT-PCR for α-smooth muscle actin and smooth muscle myosin heavy chain confirmed SMC identity, and α1D mRNA was the predominant α1-subtype, with less α1B and no α1A (data not shown). Low concentrations of NE (mean 27 nmol/l), in the presence of propranolol to block β-ARs, induced a 1.8-fold increase in phospho-ERK, and activation was abrogated to an equal extent by a low concentration of BMY-7378 (10 nmol/l), the α1D-selective antagonist, and prazosin, the nonselective α1-antagonist (Fig. 4). The α1A-selective agonist, A61603 (10 to 100 nmol/l), did not activate ERK (data not shown). Phospho-MLC was barely detectable with 3 different antibodies and was not useful as a read-out (data not shown).
We conclude that the α1D is functional in human epicardial coronary SMCs, whereas there is no evidence for the α1A.
β-AR mRNA and protein levels
We measured β-AR mRNAs and protein in the same coronary arteries, to compare with α1-ARs. By qRT-PCR (20) the β2 was the predominant β-subtype mRNA (99% of total β-AR mRNA) (Table 2). Total α1-AR mRNA was 37% of total β-AR mRNA (p < 0.0001) (Table 2).
The β-AR proteins were quantified by saturation binding, because the available β-AR antibodies are not specific in our hands (12). Saturation binding with 125I-CYP, a nonselective β-AR antagonist, identified 25.2 fmol/mg protein of total β-ARs in coronary artery membranes, with a Kd 0.16 nmol/l and specific binding 41% of total at the 125I-CYP Kd (Fig. 2B). Given the preponderance of β2-mRNA in coronary tissue, β-AR competition binding was not done. Total α1-AR binding was 35% of total β-AR binding, in excellent agreement with the mRNA values (Table 2).
In summary, the β2 was the predominant β-AR subtype in epicardial coronary arteries, and α1-AR mRNA and binding were only one-third of β-ARs.
Impact of clinical variables on coronary α1- and β-AR mRNA levels
The qRT-PCR results were analyzed to determine whether clinical variables affected the expression of AR subtypes in human coronaries. Human noncoronary arterial and prostate α1-ARs are said to increase with age (9,24), and α1-mediated vasoconstriction is more prominent in CAD (1–4). However, we found that age, EF, β-agonist exposure, CAD (Fig. 5),and sex (data not shown) had no effect on coronary artery α1-subtype mRNA levels. Interestingly, α1D and total α1-mRNA levels were 35% lower in coronary arteries of patients using β-blockers (p = 0.04). This association persisted after adjusting for age, sex, CAD, and EF (p = 0.03) (Fig. 5B). Among β-blockers, α1D mRNA levels appeared similar among patients taking metoprolol (1 patient), nadolol (1 patient), or carvedilol (8 patients) (Fig. 5B).
The levels of coronary β-AR mRNAs, which were almost entirely β2, did not vary with age, EF, β-blocker or β-agonist use, or sex (data not shown).
In summary, age, sex, CAD, EF, and β-agonists had no significant effect on α1- or β-subtype mRNA levels in coronary arteries. Beta-blocker use was associated with a significant decrease in α1D and total α1-mRNA levels.
This study reports that the α1D subtype is the predominant α1-AR in human epicardial coronary arteries, comprising approximately 80% of total α1-AR mRNA and protein. These data also reveal that coronary α1-AR levels are only one-third of β-ARs. This is the most extensive characterization of α1-AR subtypes in coronary tissue in any species.
Our results disagree with those of a previous investigation that identified the α1A as the predominant α1-subtype in human coronary arteries (9). In our study, the α1A in the coronaries was only 4% of total α1-mRNA in coronaries and was absent in isolated SMCs, and the combined α1A and α1B were only 25% of total α1-binding. The discrepancy might be explained by the prior study's small sample size (5 arteries from an unspecified number of patients), the use of post-mortem tissue, or a qualitative RNA assay (RNase protection) (9).
Of particular importance is our finding that the α1-AR subtype profiles in the human coronary arteries and ventricular myocardium were quite different. Previous limited evidence suggests that α1A subtype mRNA is predominant in both coronary arteries and myocardium (9,10). In contrast, we found that the α1D was predominant in the coronary arteries but was much less abundant in the myocardium, where α1A mRNA was predominant. In the mouse heart, myocardium has the α1A and α1B subtypes (6), whereas the α1D subtype is functional in coronary arteries (7,8). Thus, rodents and humans might have similar α1-AR subtype expression in coronary arteries and myocardium (7,8,17,20,25–27) (Table 3),contrary to prior claims (9), and studies done in mouse models might therefore be applicable to human cardiac α1-AR biology.
Technical aspects of this study warrant emphasis. We studied coronaries from a large number of patients of both sexes without and with CAD or heart failure and assessed the effects of these variables on α1-AR expression. We took extensive measures to validate our qRT-PCR approach. We quantified α1-subtype proteins by competition binding, because binding is for now the only accurate method to measure α1-AR proteins (21).
We also measured β-AR subtypes in the coronaries and found that α1-AR levels were only one-third of β-ARs. However, these lower α1-AR levels do not negate the physiological significance of the α1D. The contractile response to NE in isolated human epicardial coronary arteries is constriction at low concentrations (nmol/l) and relaxation at high concentrations (μmol/l) (28). The α1D has the highest NE affinity of any subtype (29), and thus constriction at low NE concentrations is consistent with an α1D-response. The larger population of coronary β-ARs could mediate relaxation with μmol/l NE.
Indeed, our experiments in epicardial coronary SMCs revealed that the α1D mediated activation of ERK by low concentrations of NE (Fig. 4). ERK is activated by the α1D in rat aortic SMCs (22), and ERK phosphorylation facilitates activation of MLC kinase in smooth muscle, thus contributing to the adrenergic contractile response (23). These findings suggest that the α1D is both abundant and functional in human epicardial coronary arteries.
We analyzed coronary α1- and β-AR subtype mRNA levels by age, CAD, EF, β-blockers, and β-agonists (Fig. 5) and sex (data not shown). The only association we found was a decrease in the α1D and total α1-ARs in patients treated with β-blockers, possibly implying that α1D expression in coronary vascular cells is increased by β-stimulation. In fact, in human monocytes, β2-stimulation induces α1D mRNA and protein (30).
Coronary α1-subtype mRNA levels did not differ in patients with CAD versus without CAD. This was noteworthy, because α1-ARs cause pronounced vasoconstriction in atherosclerotic coronary arteries but have little effect in normal coronaries (1–4). Thus, increased α1D levels in coronary vascular SMCs might not explain augmented α1-vasoconstriction in CAD. Instead, a small population of endothelial cell α1-ARs mediating endothelium-dependent vasodilation (31) could be lost in CAD.
Important clinical implications derive from the predominance of the α1D subtype in human coronary arteries. Cell and animal models show that the cardiac myocyte α1A and α1B subtypes have significant adaptive and protective roles (6,11–13,16). Clinical trials also suggest α1-mediated cardioprotection, because nonselective antagonism of all α1-subtypes was associated with a 2-fold excess of heart failure in the doxazosin arm of the ALLHAT (Antihypertensive and Lipid-Lowering treatment to prevent Heart Attack Trial), and a trend towards increased mortality in the prazosin arm of the V-HeFT (Vasodilator-Heart Failure Trial) (14,15).
Despite the ALLHAT trial results, 13.4 million prescriptions were dispensed in 2002 for mostly nonselective α-blockers (32), primarily to treat symptoms from prostate hypertrophy. However, the α1D-selective antagonist, naftopidil, is effective in relieving prostate symptoms (33). In light of our results, it seems possible that more selective antagonism of the α1D subtype might relax both coronary and prostate smooth muscle, without blocking beneficial signaling mediated by the myocardial α1A and α1B subtypes. Furthermore, the adaptive and protective roles of the myocardial α1A and α1B raise the intriguing possibility of activating one or both of these subtypes selectively to treat myocardial disease (16). The low levels of the α1A and α1B in human coronary arteries would make α1A- or α1B-agonist–induced coronary vasoconstriction unlikely.
We present here the most extensive characterization of human coronary ARs to date. We show that the α1D is the predominant α1-AR subtype in human epicardial coronary arteries—not the α1A as reported previously (9)—and that the α1D is low in myocardium. The α1A and α1B are present in coronaries at very low levels, and total α1-AR levels are one-third the level of β-ARs, most of which are the β2 subtype. The tissue distribution of α1-subtypes in the human heart is similar to the rodent heart (Table 3). These results are clinically relevant to the widespread use of α1-antagonists and to the potential development of α1A- and/or α1B-AR–selective agonists.
The authors thank the CTDN for unused donor hearts and Celia Rifkin and the staff in UCSF operating rooms 9 and 10 for help with transplant hearts. Sanjiv Shah, MD, did the multivariate analysis.
Dr. Simpson received funding from the Veterans Administration and the National Institutes of Health. Dr. Jensen was the recipient of a Young Investigators Award from the GlaxoSmithKline Research and Education Foundation for Cardiovascular Disease and has received support from the University of California, San Francisco Foundation for Cardiac Research. Dr. Laden received a fellowship from the Sarnoff Cardiovascular Research Foundation, and was a medical student at Duke University, Durham, North Carolina, when this work was done. Dr. DeMarco has served as a speaker/consultant for Actelion, Gilead, Boston Scientific, Cardiokinetics, and Medtronic. Dr. Jensen is currently affiliated with the Division of Cardiology, University of North Carolina School of Medicine, Chapel Hill, North Carolina. Dr. Laden is currently affiliated with the Department of General Surgery, Stanford University, Stanford, California.
- Abbreviations and Acronyms
- adrenergic receptor
- coronary artery disease
- California Transplant Donors Network
- ejection fraction
- extracellular signal-regulated kinase
- left ventricle/ventricular
- myosin light chain
- messenger ribonucleic acid
- quantitative real-time reverse transcription polymerase chain reaction
- ribonucleic acid
- smooth muscle cell
- University of California, San Francisco
- Received July 8, 2008.
- Revision received May 13, 2009.
- Accepted May 19, 2009.
- American College of Cardiology Foundation
- Jones C.J.,
- DeFily D.V.,
- Patterson J.L.,
- Chilian W.M.
- Heusch G.,
- Baumgart D.,
- Camici P.,
- et al.
- Chilian W.M.,
- Boatwright R.B.,
- Shoji T.,
- Griggs D.M. Jr.
- Baumgart D.,
- Haude M.,
- Gorge G.,
- et al.
- Graham R.M.,
- Perez D.M.,
- Hwa J.,
- Piascik M.T.
- Turnbull L.,
- McCloskey D.T.,
- O'Connell T.D.,
- Simpson P.C.,
- Baker A.J.
- Chalothorn D.,
- McCune D.F.,
- Edelmann S.E.,
- et al.
- Rudner X.L.,
- Berkowitz D.E.,
- Booth J.V.,
- et al.
- Price D.T.,
- Lefkowitz R.J.,
- Caron M.G.,
- Berkowitz D.,
- Schwinn D.A.
- Huang Y.,
- Wright C.D.,
- Merkwan C.L.,
- et al.
- Chan T.,
- Dash R.,
- Simpson P.C.
- Rokosh D.G.,
- Stewart A.F.,
- Chang K.C.,
- et al.
- Owens G.K.,
- Loeb A.,
- Gordon D.,
- Thompson M.M.
- Jensen B.C.,
- Swigart P.M.,
- DeMarco T.,
- Hoopes C.,
- Simpson P.C.
- Xin X.,
- Yang N.,
- Eckhart A.D.,
- Faber J.E.
- Stewart A.F.,
- Rokosh D.G.,
- Bailey B.A.,
- et al.
- Simpson P.
- Toda N.
- Schwinn D.A.,
- Johnston G.I.,
- Page S.O.,
- et al.
- Nishino Y.,
- Masue T.,
- Miwa K.,
- Takahashi Y.,
- Ishihara S.,
- Deguchi T.