Arginase I Attenuates Inflammatory Cytokine Secretion Induced by Lipopolysaccharide in Vascular Smooth Muscle Cells
Objective—Inflammation plays an important role in atherosclerosis. Arginase I (Arg I) promotes the proliferation of vascular smooth muscle cells; however, the effect of Arg I on inflammation remains unknown. The present study investigated the role of Arg I in inflammation in vitro and in vivo.
Methods and Results—Quantitative reverse transcription–polymerase chain reaction and Western blot analysis demonstrated that Arg I inhibited tumor necrosis factor-α production induced by lipopolysaccharide in human aortic smooth muscle cells. Inducible nitric oxide synthase substrate competition and nuclear factor-κB activation were main contributors to lipopolysaccharide-mediated inflammatory cytokine generation. However, Arg I could attenuate the function of inducible nitric oxide synthase and inhibit the subsequent nuclear factor-κB activation, leading to inhibition of tumor necrosis factor-α generation. Furthermore, upregulation of Arg I significantly decreased macrophage infiltration and inflammation in atherosclerotic plaque of rabbits, whereas downregulation of Arg I aggravated these adverse effects.
Conclusion—The results indicate the antiinflammatory effects of Arg I and suggest an unexpected beneficial role of Arg I in inflammatory disease.
Inflammation is causally linked to atherosclerosis severity and outcome. The synthesis and release of inflammatory cytokines from vascular smooth muscle cells (VSMCs) play important roles in the pathology of atherosclerosis. Some proinflammatory factors, such as lipopolysaccharide (LPS), trigger the inflammation response and promote the generation of inflammatory cytokines by VSMCs1 and are demonstrated to be involved in atherogenesis.2 Increased inflammatory cytokine activity promotes infiltration of inflammatory cells and then augments the inflammatory response.3
VSMCs in the medial wall of blood vessels are normally quiescent and express a differentiated phenotype to generate and maintain vascular tone. Under atherogenic conditions, VSMCs acquire the characteristic of a “synthetic” phenotype and may dedifferentiate, thereby secreting inflammatory cytokines and contributing to vascular pathogenesis.4 Inducible nitric oxide synthase (iNOS), an important contributor to the inflammatory response, is often induced by the exposure of smooth muscle cells (SMCs) to proinflammatory cytokines or under atherosclerotic conditions.1,5,6 Reactive oxygen species (ROS) and reactive nitrogen species are common contributors to many inflammatory diseases.7,8 Primary sources of ROS and reactive nitrogen species are several isoforms of NADPH oxidase and iNOS, which can generate superoxide anion (O2·−) and NO.9 The reaction between NO and O2·− favors the generation of peroxynitrite (ONOO−), which may contribute to the pathogenesis of inflammatory disease.10
l-Arginine is a common catalyzing substrate of NOS and arginase. Arginase is the enzyme of the urea cycle that converts l-arginine to urea. Arginase has 2 isoforms: arginase I (Arg I) and Arg II. Arg I generally expresses in the liver and also exists in extrahepatic tissues, whereas Arg II is found in several extrahepatic tissues.11 Because extrahepatic cells do not possess a complete urea cycle, the function of Arg in extrahepatic tissues has attracted attention.12 Arginase is associated with inflammatory diseases, such as asthma and bowel disease.13,14 A few studies have suggested an association of Arg I and immunity.15 Arg I contributes to immune deactivation and is involved in inflammatory responses to LPS and pathogens.16 Human aortic SMCs (hASMCs) contain constitutive Arg I, which plays an important role in SMC proliferation.17 However, the function of Arg I in inflammatory disease remains unknown. Arginase and NOS have reciprocal activities that may shift arginine metabolism to polyamine homeostasis or cytotoxic NO production, respectively. One potential mechanism for Arg I involvement in inflammation is through competition with NOS for the common substrate l-arginine.
To understand the role of Arg I in inflammatory disease, we aimed to determine the effect of Arg I on release of tumor necrosis factor-α (TNF-α) during LPS incubation in vitro and the function of Arg I in atherosclerotic plaques in vivo. Elevation of arginase activity should decrease the catalyzing substrate and inhibit the function of NOS.
Materials and Methods
Reagents and Antibodies
SMC medium was from ScienCell (Carlsbad, CA). LPS, dihydroethidium, aminoguanidine (AG), 1400W dihydrochloride (1400W), BAY-11-7082, l-arginine, α-isonitrosopropiophenone, and MnCl2 were from Sigma-Aldrich (St. Louis, MO). S-(2-Boronoethyl)-l-cysteine (BEC) was from Alexis Corp (San Diego, CA).
hASMCs were from the American Type Cell Collection. Cells at up to passage 4 were used and cultured in SMC medium containing 2% fetal bovine serum (FBS) and 1% smooth muscle cell growth supplement (ScienceCell) at 37°C for 24 to 96 hours. At 12 hours before each experiment, the complete medium was replaced by fresh SMC medium containing 0.5% FBS. BEC was a positive control of Arg I interference. The iNOS inhibitor 1400W, AG, and nuclear factor-κB (NF-κB) inhibitor BAY-11-7082 were applied 2 hours before LPS incubation.
Animal Model and Experimental Protocol
Sixty New Zealand White rabbits (1.5 to 2.0 kg) were randomly divided into 6 groups (n=10). After accustomization to an atherogenic diet (containing 1% cholesterol) for 2 weeks, all rabbits underwent balloon injury of the abdominal aorta and were fed 1% cholesterol. At the end of 12 weeks, physiological saline, Arg I lentivirus, pGC FU-green fluorescent protein-lentivirus (LV), rabbit ARGI-RNAi lentivirus (sequences: forward, 5′-GGUGGAUGCUCAUACUGAUTT-3′, reverse, 5′-AUCAGUAUGAGCAUCCACCTT-3′; forward, 5′-GUCCUACAGUAUUGAGAAATT-3′, reverse, 5′-UUUCUCAAUACUGUAGGACTT-3′), and pGC FU-RNAi-NC-LV were locally delivered to plaques as described.18 At the end of 16 weeks, all rabbits were anesthetized and killed for pathological studies. All animal care and experimental protocols complied with the animal management rules of the Ministry of Health of the People's Republic of China (document no. 55, 2001).
Cytokine levels in supernatant were determined by the TNF-α ELISA kit (Bender, Austria) without dilution according to the manufacturer's instructions. All operations were performed at room temperature. Mean absorbance for standards and samples was determined in duplicate. The color reaction was detected at 450 nm by use of a Varioskan Flash multifunction plate reader (Thermo Scientific).
Transient Transfection and Small Interfering RNA
The human ARGI cDNA was amplified and then cloned into a pCDNA3.1(+) expression vector at the XhoI and XmaI sites. pCDNA3.1(+)-hARGI plasmid, pCDNA3.1(+)-LacZ plasmid and hARGI-small interfering RNA (siRNA) (sequences: forward, 5′-GAUAUUGUGUAUAUUGGCUTT-3′, reverse, 5′-AGCCAAUAUACACAAUAUCTT-3′, Gene Pharma) and siRNA-N.C (forward, 5′-UUCUCCGAACGUGUCACGUTT-3′, reverse, 5′-ACGUGACACGUUCGGAGAATT-3′, Gene Pharma), as well as hARGI-siRNA paired with pCDNA3.1(+)-hARGI, were transfected in hASMCs with double-strand DNA in OptiMEM medium (Gibco) mixed with Lipofectamine 2000 (Invitrogen) according to the manufacturers' instructions. At 12 hours after transfection or interference, hASMCs were incubated in SMC medium containing 2% FBS for 4 hours. Arg I expression was measured by Western blot analysis.
Measurement of Arginase Activity
Arginase activity was measured in hASMC lysates as previously described.19 One unit of enzyme activity was defined as the amount of enzyme that catalyzes the formation of 1 μmol urea/minute.
Monocyte Chemotaxis and Migration
THP1 cells and hASMCs were cocultured in Transwell (0.4 and 8 μm pore size; Corning), with the former cells cultured in the upper well and the latter cells in the bottom well. After LPS treatment for 48 hours, coculture was performed (6 hours for 0.4-μm-pore Transwells, 12 hours for 8-μm-pore Transwells). Chemotaxis and migration of THP1 cells were determined by counting THP1 cells adhering to the bottom surface and to the external surface of the Transwell bottom, respectively.
Measurement of NO Activity
NO activity was determined by NO assay kit according to the manufacturer's instructions (BioVision). All operations were performed at room temperature. Absorbance was measured at 542 nm by use of a Varioskan Flash multifunction plate reader.
ROS and O2·− Assay
Cellular ROS levels were determined using the ROS Assay Kit (Beyotime, Haimen, China) according to the manufacturer's instructions. For cellular O2·− assay, cells were incubated with 2 μmol/L dihydroethidium for 30 minutes at 37°C and then underwent Hoechst 33342 staining for 5 minutes. Images were acquired with a Zeiss LSM 710 confocal microscope with a ×20 objective (number aperture 0.5) and an LD condenser (number aperture 0.55) (Zeiss, Germany).
Real-Time Polymerase Chain Reaction
The primers for iNOS were as follows: forward, 5′-TCATCCGCTATGCTGGCTAC-3′, reverse, 5′-CTCAGGGTCACGGCCATTG-3′; and for Arg I, forward, 5′-GTGGACAGACTAGGAATTGGC-3′, reverse, 5′-TCCAGTCCGTCAACATCAAAAC-3′. SYBR Green real-time polymerase chain reaction and quantitative assays involved the use of a sequence detector system (Bio-Rad). The relative expression of genes was obtained by the 2-ΔΔCt calculation method.
Westen Blot Analysis
Protein expression was assayed with cell or tissue lysates of the same protein content (BCA method, Bio-Rad). Proteins were separated by 10% SDS-PAGE, transferred to polyvinylidene difluoride membranes, and then incubated with antibodies for mouse anti-Arg I (1:4000, BD, BD Biosciences), rabbit anti-iNOS (1:1000, Santa Cruz Biotechnology), mouse anti-TNF-α (1:2000, Abcam), and mouse or rabbit anti-β-actin (1:2000, Abcam), followed by horseradish peroxidase–conjugated goat anti-rabbit (1:5000) or goat anti-mouse IgG (1:5000, both from Santa Cruz Biotechnology). The bands were developed by use of enhanced chemiluminescent reagent (Amersham Pharmacia Biotech), detected by AlphaImager HP, and quantified by using Alpha View software (Alpha).
Immunocytochemistry and Immunohistochemistry
Cells were incubated with antibodies for mouse anti-Arg I (1:100, BD) and rabbit anti-iNOS (1:100, Santa Cruz Biotechnology) or rabbit monoclonal anti-NF-κB p65 (1:50, Cell Signaling) overnight at 4°C, and then subsequent secondary antibodies (all 1:2000, Invitrogen, CA) were added. Before image acquisition, a drop of Prolong Gold antifade reagent with 4′,6-diamidino-2-phenylindole (Invitrogen) was needed.
For tissue sections, mouse anti-α-actin (1:100, Chemicon), mouse anti-TNF-α (1:100, Abcam), mouse anti-Ram11 (1:100, Dako), and mouse anti-iNOS (1:100, Santa Cruz Biotechnology) antibodies were incubated overnight at 4°C. Addition of secondary antibody and color development were followed by manufacturer's instructions (Jingmei, Shenzhen, China). Data were analyzed by use of ImagePro Plus 5.0 (Media Cybernetics).
ONOO− generation was assayed by flow cytometry. Briefly, cells were permeabilized with FACS Permeabilizing Solution 2 (BD), blocked for 30 minutes, and then incubated with mouse monoclonal nitrotyrosine antibody for 1 hour at 37°C. Alexa 488–conjugated goat anti-mouse IgG (1:1000, Invitrogen) was incubated for 20 minutes. Cell concentration was adjusted to 1×106 cells/mL for fluorescence-activated cell sorting analysis (FACSCalibur, BD). Data were analyzed by use of CellQuest Pro software (BD).
Data analysis involved use of SPSS v16.0 (SPSS Inc., Chicago, IL) and was the result of at least 3 independent experiments. Time- and dose-dependent data were assessed by Pearson bivariate correlation (2-tailed). Other data were assessed by 1-way ANOVA, followed by the Tukey t test (2-tailed). A value of P<0.05 was considered statistically significant.
LPS Increases TNF-α Production via iNOS Activity in hASMCs
To confirm the cells we used were hASMCs, cells were immunostained with anti-α-SM-actin antibodies. Results showed that all the cells were hASMCs (Supplemental Figure I, available online at http://atvb.ahajournals.org).
To test whether LPS could stimulate the generation of inflammatory cytokines, the level of TNF-α in the cell culture medium was measured. High concentrations of LPS caused significant TNF-α production (Figure 1A). Furthermore, a rapid increase in TNF-α release was detected with a short period of LPS incubation (>2 hours) (Figure 1B). To illustrate the role of iNOS in TNF-α release, we pretreated cells with 2 iNOS inhibitors, AG and 1400W, before LPS incubation. The release of TNF-α was abolished by ≈70% to 80% in the presence of the inhibitors (P<0.01) (Figure 1C).
Effect of LPS on Coregulation of Arg I and iNOS
To analyze which isoform of Arg is involved in the LPS stimulation in hASMCs, we determined the expression of both Args. Only Arg I was constitutively expressed in hASMCs, and no Arg II was detected with or without LPS incubation (data not shown). To illustrate the effects of LPS on Arg I and iNOS, the mRNA level, protein content, and catalytic activity of both enzymes were determined. The induction of mRNA and protein levels of Arg I lagged behind that of iNOS (Figure 1D and 1E). Induced Arg I activity declined after 48 hours, whereas NO release continued to increase slowly. Moreover, we observed a slight increase in NO release after Arg I activity decreased. To validate the competitive relationship between iNOS and Arg I, we pretreated cells with l-arginine during LPS incubation. Both Arg I activity and NO release were increased after 12 hours and continued to increase even at 72 hours (Figure 1F).
Arg I Suppresses TNF-α Release and Inhibits Monocyte Chemotaxis and Migration
hASMCs were transiently transfected with human Arg I–expressed plasmid and bacterial β-galactosidase (LacZ)–expressed plasmid, human Arg I siRNA, negative siRNA, and human Arg I siRNA paired with Arg I–expressed plasmid. LacZ-expressed plasmid represented unrelated cytosolic protein and negative siRNA an irrelevant interference, respectively. Transfection efficiencies were >80% in our experiments (Supplemental Figure II). hASMCs transfected with human Arg I cDNA contained a high quantity of Arg I protein as compared with controls (488.76±41.7% higher, P<0.01), whereas Arg I level in cells transfected with Arg I siRNA was barely detected (22.65±9.07%, P<0.01 versus control); cells with LacZ transfection and controls did not differ in Arg I level (P>0.05 versus control). Similar results were demonstrated between cells transfected with negative siRNA and controls (P>0.05 versus control). Transfection with human Arg I siRNA and Arg I–expressed plasmid produced a lower level of Arg I than did transfection with Arg I–expressed plasmid alone (44.05±9.13% lower, P<0.01) (Supplemental Figure IV). The inhibition levels by BEC or si-ARGI on arginase activity was presented (Supplemental Figure V).
To determine the effect of Arg I on the generation of inflammatory cytokine activity induced by LPS, release of TNF-α was determined. Arg I elevation reduced LPS-induced TNF-α secretion by 52.56±5.78% (P<0.01, versus LPS alone), whereas Arg I inhibition or knockdown increased TNF-α secretion to 113.80±3.75% and 111.38±5.33%, respectively (P<0.01, versus LPS alone) (Figure 2A). To further validate the influence of Arg I on inflammation, chemotaxis and migration of monocytes were determined. Arg I elevation in hASMCs reduced LPS-induced monocyte chemotaxis and migration by 52.83±6.49% and 40.07±12.81%, respectively (P<0.01, versus LPS alone). Opposite results were observed when cells were under Arg I inhibition or knockdown (Figure 2B to 2E).
Arg I Inhibits NO Release Without Decreasing iNOS Expression
The data presented above suggest that Arg I inhibits LPS-induced TNF-α release and that iNOS participates in the process. We speculated that Arg I inhibited inflammation through l-arginine competition, which leads to iNOS dysfunction. However, many enzymatic pathways are compartmentalized in the cell to avoid direct competition for common substrates. To show the localization of these 2 enzymes, we performed immunocytofluorescence analysis. The 2 enzymes were largely colocalized in hASMCs with LPS incubation for 48 hours (Figure 3A). Subsequently, the influence of Arg I on iNOS expression and NO release was determined. In the presence of LPS, Arg I did not significantly reduce the expression of iNOS in hASMCs (all P>0.05 versus LPS alone) (Figure 3B and 3C). However, NO release was decreased by 22.51±5.64% with Arg I elevation but increased to 130.03±7.46% and 126.10±10.57% with Arg I inhibition and knockout, respectively (both P<0.01 versus LPS alone) (Figure 3D).
Arg I Decreases ROS, O2·−, and ONOO− Generation in hASMCs
One possible mechanism of NO produced by iNOS provoking inflammation is its reaction with O2·− to produce ONOO−. O2·− may generate from iNOS uncoupling or cellular ROS. To determine the source of O2·−, ROS and O2·− production were both determined. LPS-induced increase in ROS was attenuated to 54.77±14.48% by Arg I elevation but was augmented to 133.98±10.50% and 132.97±13.52% by Arg I inhibition and knockdown, respectively (all P<0.01 versus LPS alone) (Figure 4A and 4B). Arg I elevation attenuated O2·− production induced by LPS by 41.74±12.23% (P<0.01 versus LPS alone), whereas Arg I inhibition or knockdown augmented O2·− generation to 127.70±16.03% and 124.81±14.05%, respectively (both P<0.05 versus LPS alone) (Figure 4C and 4D). Reduced ONOO− generation was also observed with Arg I overexpression (27.74±3.87%, P<0.01 versus LPS alone), whereas converse results were seen with Arg I inhibition or knockdown. (159.49±28.59% and 157.24±36.99%, respectively; P<0.01 versus LPS alone) (Figure 4E and 4F).
Arg I Inhibits TNF-α Release via NF-κB Activation
The localization of NF-κB subunit P65 was determined by immunocytofluorescence analysis. LPS triggered the NF-κB P65 subunit translocation from the cytoplasm to nucleus. However, iNOS inhibitors (AG and 1400W) and BAY-11-7082 suppressed P65 nuclear translocation. In addition, reduction of nuclear P65 translocation was seen with Arg I elevation, opposite to si-ARGI or BEC treatment (Figure 5A).
To determine a possible association of NF-κB activation and magnitude of TNF-α response to LPS, we pretreated cells with the selective iNOS inhibitor 1400W and the NF-κB inhibitor BAY-11-7082. In the presence of the 2 inhibitors, LPS-induced TNF-α release was attenuated by 24.68±4.92% and 49.50±3.14%, respectively (both P<0.01 versus LPS alone) (Figure 5B).
Arg I Attenuated Atherosclerotic Plaque Inflammation
Arg I regulation in vivo was accomplished through locally delivering ARGI-LV or si-ARG-LV. Arg I was upregulated to 282.17±53.85% by delivering ARGI-LV, whereas it was downreguated to 49.67±12.24% and 57.33±16.24% by si-ARGI-1-LV and si-ARGI-2-LV, respectively. (Supplemental Figure VI).
Immunohistochemistry results in aortaventralis of rabbit demonstrated that treatment with Arg I produced less TNF-α, which was also confirmed by Western blot analysis (Figure 6A and 6B). The opposite results from si-ARG groups were also interconfirmed by the 2 experimental approaches. Moreover, the results obtained in vivo agreed with the observations in cultured hASMCs (Figure 6A to 6C). To explore the contribution of SMCs to the release of TNF-α under Arg I regulation, we calculated the area ratio between TNF-α and SMCs. The area ratio was attenuated by Arg I upregulation to 20.17±7.81% but augmented by Arg I downregulation to 354.5±86.44% and 373.67±107.22%, respectively (all P<0.01 versus control group) (Figure 6D).
Arg I upregulation showed less macrophage content (31.67±9.05%, P<0.01, versus control group) and less iNOS expression (59.4±12.6%, P<0.01, versus control) in plaque (Figure 6E and 6F). Strikingly, si-ARGI increased macrophage content ≈80% to 100% (both P<0.01 versus control) and iNOS expression ≈35% (both P<0.05 versus control) (Figure 6E and 6F). However, the groups did not differ in blood serum TNF-α content (data not shown), which suggests that locally derived Arg I did not have a systemic anti-inflammation effect.
Coronary artery disease and atherosclerosis are considered inflammatory processes.20 Although Arg I could be a new candidate gene of atherosclerosis resistance,21 no reports exist of the effects of Arg I in these diseases. As well, the fundamental mechanism still remains unexplored. We demonstrated that hASMCs constitutively express Arg I, and Arg I inhibited the inflammation both in LPS-induced hASMCs and in artherosclerotic plaque in rabbits. This knowledge may help in the control of VSMC inflammation or pathophysiological developments, such as atherosclerosis.
VSMCs exist in a extremely diverse range of phenotypes. Under physiological conditions, the predominant phenotype is contractile VSMCs, which have a major function in vasodilation and blood flow regulation.22 However, the synthetic phenotype is present when VSMCs are subjected to injury. The synthetic phenotype of VSMC increases the capacity to generate extracellular matrix protein or inflammatory cytokines23 and contributes to vascular remodeling. SMC-derived Arg I enhances polyamine generation and promotes VSMC proliferation but has no effect on cell migration. In the synthetic phenotype, migration and proliferation of VSMCs are key elements in atherosclerotic lesion development. However, VSMC proliferation appears to have paradoxical effects: it both promotes atherosclerotic lesions and stabilizes atherosclerotic plaque. The role of Arg I in atherosclerosis still remains unclear. To address this question, we systematically explored the effect of Arg I on inflammation in vitro and in vivo.
In in vitro experiments, Arg I reduced inflammatory pathways by competing for the common substrate l-arginine with iNOS. Excess NO produced by iNOS reacts with O2·−, leading to the production of ONOO−. And ONOO− can cause oxidative damage and have a potent effect on NO− mediated inflammation. Both NO and O2·− generation, as well as ONOO− generation, were decreased with Arg I elevation during LPS incubation. However, these findings do not address whether O2·− generation was due only to iNOS uncoupling or to other sources, and then ROS generation was determined. The increase in ROS generation induced by LPS was also abolished by Arg I, the reason for which remains unknown. Subsequent mechanism research demonstrated that Arg I weakened LPS-induced TNF-α level by preventing NF-κB translocation. The effect of correlation of degradation of I-κB with NF-κB translocation suggests that Arg I inhibits LPS-induced inflammation by preventing I-κB degradation and then preventing NF-κB activation and subsequent TNF-α release.
From our data, we developed a model describing the role of Arg I in inflammatory cytokine secretion in hASMCs.
In this model, proinflammatory iNOS was induced by LPS, thus resulting in increasing NO and O2·− generation, which favors ONOO− generation. In addition to originating from iNOS uncoupling, other sources of O2·− may also contribute to the generation of O2·−. Excess ONOO− may contribute to NF-κB activation and then inflammatory cytokine secretion. Elevated Arg I may suppress the inflammatory pathway though competitively inhibiting iNOS.
Atherosclerosis development is a complex process. Arg I was reported to enhance polyamine generation and VSMC proliferation17 and therefore was considered to have a proatherosclerotic effect. Our subsequent in vivo research in an atherosclerotic New Zealand rabbit challenged this viewpoint. To determine the role of Arg I in atherosclerosis, we used a local delivery viral vector method. This gene-targeting approach has been used successfully by our laboratory and elicits effective gene regulation.24 Arg I elevation in vivo decreased TNF-α expression and macrophage content in plaques, both of which aggravated pathological changes of atherosclerosis. The findings are in agreement with the observations in hASMCs under LPS incubation. A decrease in the hASMC content of TNF-α with Arg I elevation was also observed. This result is in agreement with a previous study in primary VSMCs,17 suggesting that Arg I has different roles under physical and pathological conditions. As well, Arg I overexpression attenuated the expression of iNOS in vivo, which was different from the observation in vitro. The reduction in macrophages, which also expresses iNOS in atherosclerotic conditions, was the main contributor to the decrease.
In the present study, we illustrated the role of Arg I in inflammatory cytokine secretion in hASMCs and in atherosclerotic plaques in rabbits. Further research is needed to consider Arg I a possible target gene in treating and preventing inflammatory diseases.
Sources of Funding
This work was supported by the National 973 Basic Research Program of China (No. 2009CB521900), the State Program of National Natural Science Foundation of China for Innovative Research Group (Grant 81021001), and the National Natural Science Foundation of China (Grant 30871037).
We thank Dr Yongxin Zou (Genetic Institutes, Shandong University) for providing THP1 cells.
- Received November 21, 2010.
- Accepted May 12, 2011.
- © 2011 American Heart Association, Inc.
- Moreno PR,
- Falk E,
- Palacios IF,
- Newell JB,
- Fuster V,
- Fallon JT
- Xia Y,
- Dawson VL,
- Dawson TM,
- Snyder SH,
- Zweier JL
- Ckless K,
- Lampert A,
- Reiss J,
- Kasahara D,
- Poynter ME,
- Irvin CG,
- Lundblad LK,
- Norton R,
- der Vliet A v,
- Janssen-Heininger YM
- North ML,
- Khanna N,
- Marsden PA,
- Grasemann H,
- Scott JA
- Horowitz S,
- Binion DG,
- Nelson VM,
- Kanaa Y,
- Javadi P,
- Lazarova Z,
- Andrekopoulos C,
- Kalyanaraman B,
- Otterson MF,
- Rafiee P
- Wei LH,
- Wu G,
- Morris SM Jr.,
- Ignarro LJ
- Chen WQ,
- Zhang L,
- Liu YF,
- Chen L,
- Ji XP,
- Zhang M,
- Zhao YX,
- Yao GH,
- Zhang C,
- Wang XL,
- Zhang Y
- Teupser D,
- Burkhardt R,
- Wilfert W,
- Haffner I,
- Nebendahl K,
- Thiery J
- Owens GK
- Doran AC,
- Meller N,
- McNamara CA