| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
1Research Center for Cardiovascular Diseases, Brown Foundation Institute of Molecular Medicine for the Prevention of Human Diseases, University of Texas Health Science Center at Houston, Houston; 2Department of Carcinogenesis, University of Texas M. D. Anderson Cancer Center, Smithville; 3Encysive Pharmaceuticals, Houston; and 4Division of Cardiology, Department of Internal Medicine, Medical School, University of Texas Health Science Center at Houston, Houston, Texas
Submitted 27 December 2007 ; accepted in final form 6 March 2008
 |
ABSTRACT |
|---|
 TOP ABSTRACT MATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
Harlan rats; Sasco rats; carotid artery injury; vascular smooth muscle cells; terminal deoxynucleotidyl transferase-deoxyuridine nick-end labeling; 5-bromo-2'-deoxyuridine
-actin-immunoreactive vascular smooth muscle cells (VSMCs) (13, 20, 23). Although the identification of genes that would predispose patients to restenosis is of great use, such studies have not been consistent in their findings (10, 19, 25, 26). In 1925, a colony of Sprague-Dawley rats was initiated by the Sprague-Dawley Company. Fifty years later, it was divided into two colonies owned by different companies. Sasco, which has since been absorbed into Charles River Laboratories (Kingston, NY), established its colony in 1979; Harlan (Indianapolis, IN) established its colony in 1980. Despite a common genetic origin, the two colonies of rats, hereinafter called Sasco and Harlan, have since come to exhibit quite distinct phenotypes (7, 15, 16, 29, 33, 34), suggesting a substantive divergence over a number of generations with the same ancestrous genetic background. In our laboratory, we observed that the carotid arteries of Sasco and Harlan rats responded quite differently to balloon-induced vascular injury (i.e., those of Sasco rats undergo strikingly more intimal hyperplasia). Consequently, we hypothesized that Harlan and Sasco rat carotid arteries represented a restenosis-resistant and a restenosis-prone model, respectively. Morphometric characterization of balloon-injured Harlan and Sasco carotid arteries clearly supported this notion as we will describe in RESULTS for Figs. 1–4.
|
|
-actin-immunoreactive VSMCs (13, 20, 23) and a number of studies have demonstrated the key role of VSMCs in restenosis (8, 18, 21, 27), we further hypothesized that the above phenotypic difference between Harlan and Sasco Sprague-Dawley rats was due to the intrinsic (namely, genetic) differences in responses of their VSMCs to growth and apoptotic stimuli. To test the hypothesis, we isolated carotid artery VSMCs (CAVSMCs) from Harlan and Sasco Sprague-Dawley rats and subjected them to molecular analyses. The data, as described in RESULTS for Fig. 8, suggest the significant correlation between the degree of neointimal proliferation and the propensity of VSMCs to proliferate and die less.
|
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
Animals. Male Sprague-Dawley rats (400–450 g) were purchased from the Sasco Division of Charles River Laboratories or Harlan. All rats were housed individually and cared for in accordance with the National Institutes of Health's (NIH) Guide for the Care and Use of Laboratory Animals (NIH Publication No. 85-23, Revised 1996). All the experiments described in the current study were approved by the Institutional Animal Welfare Committee of the University of Texas Health Science Center (Houston, TX).
Surgical intervention. The original rat carotid artery balloon injury model (6) was modified. Briefly, individual rats (n = 72) were weighed and then anesthetized with 2% halothane (Halocarbon, River Edge, NJ) using 100% oxygen as the carrier gas. A midline neck incision was made to expose the left common carotid artery; the iliac artery was exposed through another incision above its junction with the femoral artery and then ligated at the distal end. A 2-Fr Fogarty arterial embolectomy catheter (Baxter Healthcare, Santa Ana, CA) was inserted into the iliac artery and passed through the aorta to the distal portion of the left common carotid artery. The placement of the catheter was checked via the midline incision in the neck. The balloon catheter was inflated with a manually driven inflator device (Encore; Scimed, Maple, MN) to 2.5 atm and then retracted while still inflated to the origin of the left common carotid artery at the aorta. The balloon catheter was deflated, returned to its original position, inflated, and retracted through the carotid artery twice. The Fogarty catheter was removed via the iliac artery that was ligated proximal to the incision site; the skin in both the neck and hindlimb was closed, and the incision sites were treated with topical antibiotics. The right common carotid artery served as an uninjured, baseline control.
Tissue processing. Rats were intraperitoneally injected with 150 mg/kg 5-bromo-2'-deoxyuridine (BrdU; Sigma) 18 and 6 h before euthanasia. Twelve rats (6 from each substrain) were euthanized at 14 days after balloon injury for morphometric analysis. Another 24 rats (12 from each substrain) were euthanized at 1, 3, 7, or 14 days after balloon injury (3 from each substrain at each time point) for time-course analysis of cell number. An additional 36 rats were euthanized at 7, 14, and 32 days after balloon injury (6 from each substrain at each time point) for the evaluation of delayed neointimal response of Sasco and Harlan rats. After perfusion fixation, the middle third of each carotid artery was placed in buffered formalin solution, embedded in paraffin, sectioned, and stained with hematoxylin and eosin.
Morphometric Analysis
Sections were analyzed by using a BIOMAX-BX40 triocular microscope (Olympus, Tokyo, Japan) connected to a CCD-IRIS video camera (Sony, Tokyo, Japan), which allowed the projection of the entire observed field onto a television monitor. Luminal, intimal, medial, and adventitial areas (in mm2) were measured with the aid of a JAVA computerized image analysis system (Jandel Scientific, Corte Madera, CA). The intima was defined as the layer between the endothelial lining of the lumen and the internal elastic lamina; the media, as the layer between the internal and external elastic laminae; and the adventitia, as the layer between the external elastic lamina and the edge of the loose fibroadipose tissue surrounding the carotid arteries.
Determination of Cell Number
The cell number was determined by examining hematoxylin and eosin-stained tissue sections under a light microscope and by counting all cell nuclei in each compartment of the vessel (intima, media, and adventitia).
Terminal deoxynucleotidyl transferase-deoxyuridine nick-end labeling staining. Terminal deoxynucleotidyl transferase-deoxyuridine nick-end labeling (TUNEL) staining (30) was performed using a FragEL DNA fragmentation detection kit (Oncogene Research Products, Boston, MA) according to the manufacturer's instructions, with a following modification: proteinase K+ treatment was shortened to 10 min. The apoptotic index, defined as the number of cells with 3,3'-diaminobenzidine-positive nuclei divided by the total number of cells counted and expressed as a percentage, was then calculated. All cells within the intima were counted.
BrdU staining. Tissue was incubated with a anti-BrdU antibody solution (Clone B44, Becton and Dickinson Immunocytometry Systems, San Jose, CA), supplemented by 3% normal horse sera. After counting all the nuclei, a BrdU index was determined as the number of cells with 3,3'-diaminobenzidine-positive nuclei divided by the total number of cells counted and expressed as a percentage.
Tissue Culture
Isolation of CAVSMCs from Harlan and Sasco rats.
VSMCs were isolated from the common carotid arteries of Harlan and Sasco rats, as previously described (17) but with the following modifications: the harvested carotid arteries were placed under a dissecting microscope; trimmed free of adventitial, neural, and fatty tissue; and cut into 
1-mm-long pieces using ophthalmologic scissors. Tissues were washed with PBS and then incubated for 30–60 min at 37°C in digestion buffer containing 340 U/ml collagenase (Sigma), 15 U/ml elastase (Sigma), 1 mg/ml soybean trypsin inhibitor (Sigma), and 1 mg/ml dithiothreitol in Hanks’ balanced salt solution (GIBCO-Invitrogen, Carlsbad, CA) (pH 7.4). Released CAVSMCs were suspended in 20% fetal calf serum (FCS) in Dulbecco's modified Eagle's medium (DMEM), plated in six-well plates, incubated at 37°C with 5% CO2-95% room air in a tissue culture incubator (Sanyo Scientific, Bensenville, IL) for 7–10 days, and finally characterized as described in Characterization of CAVSMCs.
Characterization of CAVSMCs.
Isolated CAVSMCs were seeded on a chamber slide (Nalge Nunc International), allowed to differentiate in differentiation media (0.4% FCS and 50 µg/ml heparin in DMEM) for 10 days, and stained with anti--actin antibody (Chemicon, Temecula, CA). Nuclei were counterstained with 4,6-diamidino-2-phenylindole (Sigma). Only those batches of CAVSMCs that showed a 90% or higher -actin positivity were used for further analyses.
Maintenance of CAVSMCs. Isolated CAVSMCs were maintained in Media 231 with smooth muscle cell growth supplements (Cascade Biologics, Portland, OR). All experiments were performed using cells of the seventh passage or less. At least two independent isolates were examined, and representative data from one independent experiment were presented.
Migration assay. Migration assays were performed as previously described (32) with the following modifications. Seventy microliters of serum at various concentrations (0%, 0.05%, 0.5%, 5%, and 100%) were added to the bottom chambers of the ChemoTx System plates (5 µm diameter, Neuro Probe, Gaithersburg, MD). CAVSMCs (2 x 104) suspended in 25 µl of serum-free medium were seeded in the upper chambers and incubated at 37°C for 6 h.
Growth assay. Growth assays were performed as previously described (32). Briefly, CAVSMCs (1 x 105) were seeded in duplicate in six-well plates, synchronized by serum starvation (24 h), and then serum stimulated (5% FCS) for various periods of time. After growth stimulation, cells were harvested by trypsinization and counted using a hemocytometer.
Thymidine incorporation assay. Thymidine incorporation assays were performed as previously described (32). The thymidine index was calculated as follows: [total counts (in dpm)]/[total protein amount (in µg)].
Cell death assay.
The viability of Harlan and Sasco CAVSMCs exposed to TNF- was determined by measuring their capacity to exclude the vital dye trypan blue (Sigma) as previously described (1) with the following modifications. For dose-response assays, CAVSMCs were cultured in six-well plates and incubated for 48 h at 37°C in the presence of various concentrations of TNF- (0, 1, 10, and 100 nM; R&D Systems, Minneapolis, MN). For time-course assays, CAVSMCs were cultured in six-well plates and incubated for 0, 48, and 96 h at 37°C in the presence or absence of 10 nM TNF- (R&D Systems). After the incubation, both floating and adherent cells were harvested, pelleted by centrifugation, resuspended in 500 µl of culture medium containing 0.1% trypan blue, loaded into a hemocytometer, and examined by light microscopy. Viable (white) and nonviable (blue) cells were then counted, and cell death rate was expressed as the percentage of dead cells.
Statistical Analysis
All data were expressed as means ± SD. Two-sample t-test (Figs. 2, 4, 5, 6, and 8B) and ANOVA (Figs. 3, 7C, and 8, A, C, and D) were used to detect differences between two and more groups, respectively. All statistical comparisons were performed with Minitab software (Minitab, State College, PA). P values < 0.05 were considered significant.
|
|
|
|
|
|
RESULTS |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
Light microscopic examination of balloon-injured left carotid arteries from Harlan and Sasco rats 14 days after injury showed drastically different vascular responses. Despite their identical baseline morphology (Fig. 1, baseline), injured Harlan carotid arteries showed a greater luminal area, whereas injured Sasco carotid arteries showed a greater intimal area (Fig. 1, injury and injury-high magnification). Computer-assisted morphometric analyses confirmed these findings, showing that injured Sasco carotid arteries had a significantly smaller luminal area (Fig. 2A, P < 0.01), larger intimal area (Fig. 2B, P < 0.01), thicker intimal wall (Fig. 2C, P < 0.005), and higher intima-to-media ratio (IMR; Fig. 2D, P < 0.01). Sasco carotid arteries also contained significantly more intimal cells than did those of Harlan rats (Fig. 2E, P <0.005). The intimal area per cell after injury (in mm2/cell number x 106) did not differ in carotid arteries from both substrains (Fig. 2F, P = not significant). This was also true at day 7 (252.4 ± 18.9 vs. 309.5 ± 50.6 mm2/cell number x 106; P = not significant). The areas of the external elastic lamina (whole vessel area) were the same for both substrains (Fig. 2G), whereas the adventitial areas were significantly larger in Harlan than in Sasco rats (Fig. 2H).
Sasco Intimal Cells Proliferate More Rapidly than Harlan Intimal Cells After Carotid Artery Injury
To further characterize the vascular responses to injury of Sasco and Harlan carotid arteries, intimal, medial, and adventitial cell numbers were determined on days 1, 3, 7, and 14 after balloon injury (Fig. 3). Strikingly, the intimal cells of Sasco carotid arteries proliferated significantly more rapidly than did those of Harlan carotid arteries (P < 0.001, two-way ANOVA) (Fig. 3A). On the contrary, at no observed time point did medial and adventitial cell numbers differ significantly between Sasco and Harlan carotid arteries (Fig. 3, B and C, media and adventitia, two-way ANOVA).
Sasco Neointima Remains Significantly Larger than Those of Harlan 32 Days After Balloon Injury
To determine whether the robust neointimal formation seen in Sasco carotid arteries would be caught up by Harlan in a later time point, we examined the intima of injured carotid arteries of these substrains 32 days after balloon-injury. At day 32, the uninjured carotid arteries of Sasco and Harlan rats exhibited the similar number of cells (Fig. 4A) and IMR (Fig. 4B). Strikingly, the neointima of the injured carotid arteries of Sasco rats continued to exhibit significantly higher numbers of cells (Fig. 4A, P < 0.005) and IMRs (Fig. 4B, P < 0.05) than did Harlan rats at day 32.
Harlan Intimal Cells Exhibit Higher TUNEL Indexes than Sasco Intimal Cells at Day 14
To test whether Harlan neointimal cells underwent more apoptotic death, TUNEL staining was carried out (Fig. 5A). As is shown in Fig. 5B, Harlan neointima exhibited more TUNEL-positive cells (P < 0.005) than did Sasco neointima at 14 days, whereas Sasco neointima exhibited a trend toward more TUNEL-positive cells at 32 days (P = 0.082).
Sasco Intimal Cells Exhibit Higher BrdU Indexes than Harlan Intimal Cells at Day 14
To test whether Sasco neointimal cells underwent more proliferation, BrdU staining was carried out (Fig. 6A). As is shown in Fig. 6B, Sasco neointima exhibited more BrdU-positive cells (P < 0.05) than did Harlan neointima at 14 days, whereas BrdU indexes did not differ statistically at 32 days.
Sasco and Harlan CAVSMCs Are Equal in Their Morphology and Propensity to Migrate
To determine whether Sasco and Harlan CAVSMCs possess the distinct properties that would explain the difference in the degree of neointima formation in response to vascular injury in vivo, CAVSMCs were isolated from normal Sasco and Harlan carotid arteries and cultured under the same growth conditions. Morphologically, Sasco and Harlan CAVSMCs were indistinguishable in both proliferative state (Fig. 7A) and -smooth-muscle-actin-positive state (Fig. 7B). In addition, Sasco and Harlan CAVSMCs migrated equally in a standard migration assay (Fig. 7C).
Sasco CAVSMCs Proliferate More Rapidly and Incorporate More Thymidine than Harlan CAVSMCs Under the Same Growth-Stimulating Conditions
Although Sasco and Harlan CAVSMCs were identical in their morphology and migration capabilities, their growth rates were drastically different (Fig. 8A). Although their numbers were similar at baseline (Fig. 8A, day 0, P = 0.37), Sasco CAVSMCs began to proliferate more rapidly than did Harlan as early as day 1 of serum stimulation (P < 0.05, ANOVA-General Linear Model, Turkey comparison methods). Their growth curves then rapidly diverged so that, by day 5 of serum stimulation, the Sasco CAVSMC population far outnumbered the Harlan CAVSMC population (P < 0.001) (Fig. 8A, day 5). In thymidine incorporation assay, Sasco CAVSMCs showed a higher rate of thymidine incorporation than did Harlan CAVSMCs under the exact same growth conditions (P < 0.05, 2-sample t-test) (Fig. 8B).
Sasco CAVSMCs Are Less Sensitive to TNF--Induced Cell Death than Harlan CAVSMCs
Sasco CAVSMCs were more resistant to TNF--induced cell death: at all concentrations of TNF- ranging from 1–100 nM, a significantly lesser percentage of Sasco CAVSMCs were dead than Harlan CAVSMCs for 1–100 nM TNF- concentrations (P < 0.001 by ANOVA-General Linear Model) (Fig. 8C). Consistently, at 10 nM TNF- concentration, a lesser percentage of Sasco CAVSMCs were dead over 96 h (P < 0.001) (Fig. 8D).
|
DISCUSSION |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
65% greater in the Harlan substrain (Figs. 1 and 2A). Since the areas per cell were the same in both substrains (Fig. 2F), the lesser degree of neointima formation in the Harlan substrain was due to the significantly lower number of intimal cells in the substrain (Figs. 2E and 3A). TUNEL and BrdU analyses (Figs. 5 and 6) suggest that these phenotypes came about, at least partly, from more apoptosis and less proliferation of Harlan intimal cells. In addition, Sasco's neointimal formation was still substantially more than Harlan's on the 32nd day (Fig. 4), suggesting that the Harlan's represents a true restenosis resistant substrain rather than a delayed responder to vascular injury. The lack of significant difference in TUNEL and BrdU indexes of Harlan and Sasco on the 32nd day is most likely due to the fact that vascular remodeling after injury has been mostly completed by that day (Figs. 5 and 6). It has been shown that VSMC growth plays an important role in the formation of neointima after vascular injury (13, 20, 23, 24). Although we have shown that Sasco rats exhibit far greater neointimal formation after vascular injury than do Harlan rats (Figs. 1–4), left unanswered was whether restenosis-prone (Sasco) and -resistant (Harlan) phenotypes could be explained by the intrinsic, genetic difference in Sasco and Harlan VSMCs. In other words, it was not known whether VSMCs from restenosis-prone (Sasco) and -resistant (Harlan) rats behave differently when they were separated from the external influence of platelets, coagulation cascades, leukocytes, and growth-promoting factors. It was possible that VSMCs of Sasco and Harlan were identical in every way and that Sasco's greater neointimal formation results solely from greater growth stimuli to VSMCs occurring after vascular injury in Sasco rats. To answer this, we isolated CAVSMCs from Harlan and Sasco rats (Fig. 7) and subjected both populations of cells to the exact same growth-promoting or apoptosis-inducing conditions in vitro (Fig. 8). Sasco CAVSMCs looked identical to Harlan CAVSMCs in their morphology (Fig. 7, A and B) and propensity to migrate (Fig. 7C). Strikingly, however, Sasco CAVSMCs grew more rapidly (Fig. 8A) and took up more thymidine (Fig. 8B) under the same growth-stimulating conditions, whereas Harlan CAVSMCs were more sensitive to apoptotic stimuli under the same apoptosis-inducing conditions (Fig. 8, C and D). Taken together, it is likely that these intrinsic properties of Harlan VSMCs in vitro, namely, highly sensitive to cell death yet slow to respond to growth stimuli, manifested themselves in the strikingly less neointimal proliferation after balloon injury in vivo.
Although we have shown here that the in vitro responses of Harlan and Sasco CAVSMCs to growth and cytotoxic stimuli are entirely concordant with in vivo phenotypes of Harlan and Sasco rats in response to balloon vascular injury, the current data do not completely exclude the possibility that the phenotypic differences between Harlan and Sasco carotid arteries were also contributed by factors other than VSMCs, such as responses of platelets, leukocyte, coagulation cascades, and growth-factor production in the setting of vascular injury, all of which have been shown to play critical roles in the pathogenesis of restenosis (5, 11, 14, 28, 31). Nevertheless, the current work, to our knowledge, is the first to report restenosis-prone and -resistant rat substrains that originated from the same ancestrous strain (i.e., Sasco and Harlan Sprague-Dawley rats) and to suggest that a difference in the genetic (intrinsic) program on proliferation and cell death (Fig. 8), but not migration or differentiation (Fig. 7), of VSMCs can manifest itself in a difference in the degree of neointimal formation upon vascular injury (Figs. 1–4).
Other investigators have also explored the animal models of restenosis. Assadnia and others balloon-injured the iliac arteries of 11 rat strains and found that the IMR varied from 0.24 to 0.58 at 8 wk after balloon injury (3). Harmon and others (9) ligated the left common carotid arteries of 11 mouse strains and found that they exhibited highly variable remodeling patterns with IMR ranging from 0 (SM/J mice) to 1.025 (C3H/HeJ mice) to 1.826 (FVB/NJ mice) at 4 wk. In vitro, aortic VSMCs from FVB/NJ mice proliferated more rapidly than those from C3H/HeJ mice, suggesting that the propensity of FVB/NJ VSMCs to proliferate contributed to more extensive inward remodeling and neointimal proliferation in response to ligation (9). Inoue and others (12) cather-injured the left carotid arteries of C57BL/6 and FVB mice and found that FVB mice developed much larger neointimas after the injury. BrdU indexes were significantly higher in the neointimal cells of FVB mice than those of C57BL/6 mice. VSMCs from FVB mice expressed higher levels of sphingosine-1-phosphate (S1P) receptor-1 (S1P-1) and migrated more rapidly in response to S1P stimulation, suggesting that the S1P-1-mediated migration of FVB-VSMCs from media to intima and the continued proliferation of VSMCs in intima contributed to the development of larger neointimas in FVB mice (12). In contrast to studies (4, 9, 12) where different inbred strains with drastically different genetic backgrounds were used, we observed strikingly different neointimal responses to vascular injury between Sasco and Harlan substrains that originated from the same ancestral strain of Sprague-Dawley rats and that would have less genetic variability.
It is difficult to determine whether 1) Sasco rats (IMR = 1.02) are a restenosis-prone substrain and Harlan rats a normal responder or 2) Harlan rats (IMR = 0.53) are a restenosis-resistant substrain and Sasco rats a normal responder. The nomenclatures used here, calling Harlan a restenosis-resistant substrain and Sasco a restenosis-prone substrain, are totally relative. Although the observation by Assadnia and others (3) of the IMR ranging from 0.24 to 0.58 suggests that the Sasco substrain is a restenosis prone substrain, their experimental conditions are different from ours: they injured iliac arteries (as opposed to carotid arteries in our case) and harvested the arteries at 8 wk (as opposed to 2–4 wk in our case) after balloon-injury (3).
The roles of genetic factors in postangioplasty and in-stent restenosis have been studied, showing the role in restenosis of polymorphisms of β2-adrenergic receptor gene (ADRB2) (19), angiotensin-converting enzyme (26), and others (10, 25). Despite these human studies on preselected sets of gene polymorphisms, there have been no consistent genetic markers identified. A different approach, such as one starting from identifying animal models of restenosis proneness and resistance and characterizing such animals in a systematic fashion, may be needed to identify critical genes for the pathogenesis of restenosis.
In conclusion, we report that Sasco and Harlan substrains of Sprague-Dawley rats represent restenosis-prone and -resistant substrains, respectively. The mechanistic analyses described in RESULTS for Figs. 5–8 suggest that the restenosis-proneness of Sasco Sprague-Dawley rats is due to the ability of Sasco VSMCs to more rapidly proliferate and to be less sensitive to noxious stimuli, at both whole animal and cellular levels. Sasco and Harlan Sprague-Dawley rats and VSMCs from these rats will prove to be powerful tools to study genes involved in the pathogenesis of restenosis.
|
GRANTS |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
|
REFERENCES |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
