|
|
||||||||
on airway epithelial cell chemokine expressionDepartments of Pediatrics and Communicable Diseases, Molecular and Integrative Physiology, University of Michigan, Ann Arbor, Michigan
Submitted 13 February 2007 ; accepted in final form 10 July 2007
| ABSTRACT |
|---|
|
|
|---|
, a cytokine increased in asthma and chronic obstructive pulmonary disease, on airway epithelial cell proinflammatory gene expression. Costimulation of 16HBE14o- human bronchial epithelial cells and primary mucociliary-differentiated tracheal epithelial cells with RV and TNF-
induced synergistic increases in IL-8 and epithelial neutrophil attractant-78 production. Similar synergism was observed for IL-8 promoter activity, demonstrating that the effect is transcriptionally mediated. Whereas increases in ICAM-1 expression and viral load were noted 16–24 h after costimulation, cooperative effects between RV39 and TNF-
were evident 4 h after stimulation and maintained despite incubation with blocking antibody to ICAM-1 given 2 h postinfection or UV irradiation of virus, implying that effects were not solely due to changes in ICAM-1 expression. Furthermore, RV39 infection induced phosphorylation of ERK and transactivation of the IL-8 promoter AP-1 site, which functions as a basal level enhancer, leading to enhanced TNF-
responses. We conclude that RV infection and TNF-
stimulation induce cooperative increases in epithelial cell chemokine expression, providing a cellular mechanism for RV-induced exacerbations of airways disease. asthma; ERK; chronic obstructive pulmonary disease; intercellular adhesion molecule-1; interleukin-8
RV is a positive, single-stranded RNA virus from the Picornaviridae family responsible for the majority of common colds. To date,
100 serotypes of RV have been detected with 90% of serotypes (e.g., RV14, RV16, and RV39) using intercellular adhesion molecule (ICAM)-1 as a receptor. The kinetics of Picornavirus replication are rapid, the cycle being completed in 5–10 h. A small amount of RV2 synthesis occurs as early as 3 h after infection (29).
Clinical studies suggest that RV stimulates exacerbations of asthma by inducing bronchial epithelial cell production of the Glu-Leu-Arg(ELR)+ neutrophil chemoattractant, IL-8/CXCL8, leading to a neutrophilic inflammatory response (18, 20, 25, 44). Two other ELR(+) C-X-C chemokines, epithelial neutrophil attractant (ENA)-78/CXCL5 and growth-related oncogene (GRO)-
/CXCL1, are increased in the airways of patients with COPD exacerbations (1, 21, 46). RV infection has been shown to increase expression of IL-8, ENA-78, and GRO-
in cultured airway epithelial cells (16, 23, 38, 49, 52, 57, 59). After RV16 infection, asthmatic patients show increased levels of IL-8 in their nasal lavage, which correlates with the level of airways responsiveness (25), in contrast to unaffected individuals in whom IL-8 does not increase (13). Together, these data suggest that RV infection of airway epithelial cells may potentiate preexisting proinflammatory pathways, enhancing chemokine production and airway inflammation.
To test the hypothesis that RV infection and cytokines present in inflamed airways combine to induce augmented airway epithelial cell responses, thereby promoting further airway inflammation, we examined the effect of RV39 on airway epithelial cell C-X-C chemokine production in response to TNF-
. TNF-
is produced by macrophages, T cells, mast cells, and epithelial cells, and is increased in asthmatic airways (2, 7, 8). Sputum TNF-
is also increased during acute COPD exacerbations (12). Finally, segmental allergen challenge increases bronchoalveolar TNF-
levels in allergic subjects, and the effect is potentiated by prior RV16 infection (9). We found that RV39 and TNF-
induce synergistic increases in C-X-C chemokine production, thereby providing a cellular mechanism for RV-induced exacerbations of airways disease.
| MATERIALS AND METHODS |
|---|
|
|
|---|
Primary human bronchial epithelial cell cultures obtained from lung transplant donor tracheas (University Health Network, Toronto) were grown under submerged conditions in bronchial epithelial cell growth medium (Cambrex Bio Science, Walkersville, MD) containing epidermal growth factor (25 ng/ml), bovine pituitary extract (65 ng/ml), all-trans retinoic acid (5 x 10–8 M), BSA (1.5 µg/ml), hydrocortisone (0.5 µg/ml), insulin (5 µg/ml), transferrin (10 µg/ml), epinephrine (0.5 µg/ml), triiodothyronine (6.5 ng/ml), gentamycin (50 µg/ml), and amphotericin (50 µg/ml). The Human Ethics Committee at the Toronto General Hospital and the Institutional Review Board at the University of Michigan approved the collection and use of these normally discarded tissue samples. In selected experiments, cells were cultured at air-liquid interface, as described (48).
Generation of RV stocks. RV39 was obtained from American Type Culture Collection (Manassas, VA), and viral stocks were generated as previously described (18). Briefly, HeLa cells were infected with RV until 80% of the cells were cytopathic. HeLa cell lysates were harvested, and cellular debris were pelleted by centrifugation (10,000 g for 30 min at 4°C). RV was concentrated and purified by centrifugation with a 100,000-molecular weight cutoff Centricon filter (2,000 rpm at 4°C for 8 h; Millipore, Billerica, MA) (42). For experiments measuring phosphorylation of ERK, we further decreased the possibility of contamination by serum or HeLa cellular proteins by infecting the HeLa cells in serum-free medium and harvesting the cell-conditioned media rather than cell lysates. Virus was titered by infecting confluent HeLa monolayers with serially diluted RV (range: undiluted to 10–9) and assessing cytopathic effect 5 days after infection. TCID50 values were determined by the Spearman-Karber method (31). In some experiments, RV39 was UV irradiated on ice for 30 min using a UVB CL-1000 Crosslinker at 1,200 µJ/cm2 (42). The effect of UV irradiation was confirmed by the absence of a cytopathic effect on HeLa cell monolayers.
RV infection. 16HBE14o- cells were infected with RV39 for 1 h at 33°C at multiplicity of infection of 1.0. Mucociliary-differentiated cultures were infected at a tissue culture infectivity dose 50% (TCID50)/ml of 3 x 106. In selected experiments, the cells were incubated with mouse anti-human ICAM-1 antibody (20 µg/ml; Serotec, Raleigh, NC) either 1 h before RV infection or 2 h post-RV infection. As control, equal concentrations of irrelevant mouse anti-human IgG (Sigma) were used. Selected cultures were pretreated for 1 h with U0126 (3–10 µM) in DMSO (from Sigma, St. Louis, MO) or carrier alone.
Measurement of chemokine and cytokine protein levels.
Cells were grown to 80% confluence, serum starved for 24 h, and then infected with RV39 for 1 h. Inoculum was replaced with serum-free medium containing TNF-
when appropriate. Conditioned medium was collected 4–48 h postinfection, centrifuged to remove cell debris, and frozen at –80°C. For mucociliary cultures, RV and TNF-
were applied to the apical surface for 1 h. TNF-
was also placed in the basolateral media. Forty-eight hours postinfection, supernatants were collected from the basolateral compartment. IL-8, ENA-78, and IFN-
protein levels were measured by ELISA (R&D Systems, Minneapolis, MN). Granulocyte/macrophage colony-stimulating factor (GM-CSF), TNF-
, IL-1
, and IL-1
levels were measured by multiplex immunoassay (BioRad, Hercules, CA).
Transient transfection of 16HBE14o- airway epithelial cells.
16HBE14o- cells were grown to 50% confluence, washed in Optimem (Invitrogen, Carlsbad, CA), and transfected with plasmid DNA, Optimem, and Lipofectamine (Invitrogen). After 4 h, the solution was replaced with MEM supplemented with 10% FBS. For all transfection experiments, cells were serum-starved for 16 h and infected with RV39 for 1 h. Inoculum was replaced with fresh serum-free medium, containing TNF-
when appropriate, and cells were harvested for analysis 24 h postinfection.
The –162/+44 fragment of the human IL-8 promoter subcloned into luciferase (–162/+44 hIL8/Luc) was obtained from Dr. Allan Brasier (14, 19). NF-
B and AP-1 reporter plasmids were purchased from Stratagene (La Jolla, CA). Dominant-negative MEK-2A was obtained from Dr. Dennis Templeton (58). pRL family Renilla luciferase plasmid was purchased from Promega (Madison, WI). Luciferase activity was measured using a luminometer. Changes in promoter activity were normalized for transfection efficiency by dividing luciferase light units by Renilla luciferase light units. Results were then reported as fold increase over the empty vector/untreated control.
Measurement of surface ICAM-1 expression. 16HBE14o- cells were harvested using cell dissociating buffer (Invitrogen). Cells were washed with PBS-BSA (PBS containing 1% BSA) and incubated with anti-ICAM-1 (1 µg/ml) or matched isotype IgG control for 1 h on ice. Cells were washed and incubated for 30 min on ice with Alexa Fluor 488-conjugated rabbit anti-mouse IgG (1:1,000; Molecular Probes, Portland, OR). Cells were washed, suspended in PBS-BSA, and analyzed with a BD Biosciences FACSCalibur using CellQuest software (San Jose, CA).
Semiquantitative real-time PCR. cDNA was generated using 0.2–0.5 µg of RNA, Taqman reverse transcriptase reagent kit (Applied Biosystems, Foster City, CA), and 2.5 µM of either an RV-specific primer (5' GAA ACA CGG ACA CCC AAA GTA GT 3') or oligo(dT). Real-time PCR was performed using an ABI Prism 7000 thermocycler with Sybr green PCR master mix (all from Applied Biosystems). We used 0.4 nM each primer and 80–100 ng of cDNA template in a reaction volume of 25 µl for a total of 40 cycles. Specific RV primer sequences were forward primer 5' TGG CAG ATG AGG CTA GAA ATA CCC CAC TG 3' and nested RV reverse primer 5' CAT CCC GCA ATT GCT CGT TAC 3'. The primer sequences for GAPDH, used as an internal control, were forward primer 5' CTT CAC CAC CAT GGA GAA GGC 3' and reverse primer 5' GGC ATG GAC TGT GGT CAT GAG 3'. For each time point, samples were run in triplicate, and the cycle threshold (CT) was determined. Relative gene expression was calculated as previously described (17).
Immunoblotting. 16HBE14o- cells were lysed, cellular proteins were resolved by 10% SDS-PAGE, and proteins were transferred to a nitrocellulose membrane. Membranes were probed with antibodies against Thr202/Tyr204 phospho-ERK1/2 and total ERK1/2 (both from Cell Signaling Technology, Beverly, MA). Signals were amplified and visualized with horseradish peroxidase-conjugated secondary antibody (BioRad) and chemiluminescence solution (Pierce, Rockford, IL).
Data analysis. Statistical significance was assessed by ANOVA. Differences identified by ANOVA were pinpointed by the Student-Newman-Keuls multiple range test.
| RESULTS |
|---|
|
|
|---|
have cooperative effects on C-X-C chemokine expression in human airway epithelial cells.
16HBE14o- human bronchial epithelial cells were infected with RV39 (1 h at 33°C) and stimulated with TNF-
(0.3–10 ng/ml). Forty-eight hours after infection, cell supernatants were collected and examined for the C-X-C chemokines IL-8 and ENA-78. RV39 and TNF-
each independently increased IL-8 and ENA-78 production, but RV39 infection increased TNF-
-induced chemokine production significantly above that induced by TNF-
alone (Fig. 1, A and B). In some cases, the effects were additive, but in other instances synergistic (greater than additive) increases in protein expression were noted. For example, RV39 and TNF-
(3 ng/ml) each stimulated 16HBE14o- cells to produce
1,000 pg/ml IL-8 over the 48-h treatment period, but the combination increased mean IL-8 to nearly 4,000 pg/ml. On the basis of the synergistic effects of RV39 and 3 ng/ml TNF-
, this concentration was used for subsequent experiments.
|
-induced IL-8 protein levels relative to that found after TNF-
alone (Fig. 1, C and D). These data demonstrate that the synergistic effects observed in the 16HBE14o- cells also occur in a more physiological cell culture model.
We examined the potential cooperative effects of RV and TNF-
on the protein abundance of GM-CSF. GM-CSF, a stimulus of granulocyte (neutrophil, eosinophil) differentiation and degranulation, is increased in the airways of patients with asthma (28, 51, 55) and COPD (3). RV and TNF-
had synergistic effects on GM-CSF protein abundance (Fig. 1E). At the concentrations of RV and TNF-
employed, we did not observe cooperative increases in IFN-
inducible protein-10/CXCL10, eotaxin/CCL11, or regulated upon activation, normal T cell expressed and secreted (RANTES)/CCL5 (data not shown).
Effects of RV39 and TNF-
on transcription from the IL-8 promoter.
We examined the effects of RV39 and TNF-
on transcription from the IL-8 promoter. 16HBE14o- cells were transfected with –162/+44 fragment of the full-length human IL-8 promoter subcloned into a luciferase reporter. Cells were infected with RV39 (1 h at 33°C) and stimulated with TNF-
(3 ng/ml for an additional 24 h). RV39 and TNF-
induced synergistic increases in reporter activity relative to RV39 or TNF-
alone (Fig. 1F), suggesting that the observed cooperative increases in IL-8 production are transcriptionally regulated.
Potential contribution of changes in ICAM-1 expression to the observed cooperative effects of RV39 and TNF-
stimulation on IL-8 production.
As noted above, RV39 utilizes ICAM-1 as a cellular receptor. As we (38) and others (54) have demonstrated previously, incubation of cells with monoclonal antibody to ICAM-1 blocks RV39-induced responses. TNF-
has been shown to significantly increase the surface expression of ICAM-1 after 12 h of treatment (6, 34). We therefore designed a series of experiments to test whether TNF-
increases RV39-induced IL-8 production by increasing epithelial cell ICAM-1 expression, which could in turn increase RV39 binding and infectivity. First, we examined the effects of TNF-
on surface ICAM-1 expression. ICAM-1 expression increased as early as 16 h after stimulation (Fig. 2A). Second, we reasoned that, if the observed cooperative effects between RV39 and TNF-
on chemokine expression were due to increased surface ICAM-1 expression, then TNF-
treatment should increase viral load in 16HBE14o- cells. To test this proposition, total RNA was collected at 4, 8, and 24 h postinfection and examined for viral RNA by real-time PCR. We found that TNF-
increased the number of RV39 transcripts at 24 h after stimulation, with a trend towards significance at 16 h (Fig. 2B). Collectively, these data are consistent with the notion that TNF-
-induced ICAM-1 expression, with subsequent increased binding of newly formed RV39, is responsible for the observed synergistic effects between RV and TNF-
.
|
-induced ICAM-1 expression to cooperative responses, we examined the time course of IL-8 production in response to RV39 and TNF-
. Cell supernatants were collected at 4, 8, 16, 24, and 48 h postinfection and examined for IL-8 protein (Fig. 2C). Additive effects between RV39 and TNF-
were evident 4 h after stimulation, before surface ICAM-1 expression was increased in response to TNF-
. Maximal, synergistic effects in IL-8 expression occurred at 16 and 48 h postinfection. A biphasic IL-8 protein response to RV infection has been observed previously (23).
Next, to specifically determine the requirement of ICAM-1 expression for observed synergistic response, 16HBE14o- cells were infected with RV39 (1 h at 33°C), stimulated with TNF-
(3 ng/ml), and incubated with ICAM-1-blocking antibody 2 h postinfection, thereby preventing newly replicated RV from binding to newly expressed ICAM-1. As a control, ICAM-1 antibody was added to cells 1 h before infection. Supernatants were collected 48 h postinfection. While there was a significant reduction in the combined IL-8 response to RV and TNF-
, synergy between RV and TNF-
was maintained (Fig. 2D), suggesting ICAM-1-independent as well as -dependent effects. On the other hand, incubation of ICAM-1 antibody before infection abolished RV responsiveness.
Finally, we reasoned that, if the observed cooperative effects between RV39 and TNF-
on chemokine expression were due to increased surface ICAM-1 expression, then there should be no synergistic effect of TNF-
and UV-irradiated virus, since only newly replicated virus is capable of binding to new surface ICAM-1. As has been shown previously, replication-deficient, UV-irradiated virus was still capable of inducing significant IL-8 responses (Fig. 2E). Furthermore, there was still significant synergy between UV-irradiated virus and TNF-
. Together, these data suggest that, although TNF-
increases airway epithelial cell ICAM-1 expression, this increase does not fully account for the cooperative effects between TNF-
and RV on IL-8 expression.
Cooperative effects of TNF-
and RV on airway epithelial cell cytokine expression.
It is conceivable that the cooperative effects of TNF-
and RV on IL-8 expression could be due to the synergistic induction of a second IL-8-inducing cytokine. We performed focused gene arrays examining the mRNA expression of various cytokines in response to RV infection. These pilot studies were performed in two isolates of primary mucociliary-differentiated human airway epithelial cells. Consistent with previous results (10, 36), there was surprisingly little cytokine expression in response to RV infection. Cells showed a small but significant (2-fold) increase in TNF-
(1/2 isolates), IL-1
(2/2), IL-1
(1/2), and IFN-
(1/2). On this basis, we measured RV- and TNF-
-induced expression of these proteins in 16HBE14o- cells. Conditioned media was collected at 4, 8, 16, 24, and 48 h after costimulation. We found no protein expression of TNF-
, IL-1
, or IL-1
. There was a synergistic increase in IFN-
protein expression at 24 h (Fig. 3). These data suggest that, for the most part, synergistic effects between RV and TNF-
are not due to autocrine production of IL-8-inducing cytokines.
|
costimulation, and the failure of changes in ICAM-1 and cytokine expression to fully explain the observed synergistic effects, we hypothesized that RV increases the IL-8 response to TNF-
by activating the IL-8 promoter AP-1 site, which functions as a basal level enhancer (35). Cells were transfected with cDNA encoding AP-1 responsive promoter elements subcloned into luciferase (AP-1-TATA-Luc) and examined for AP-1 transactivation 24 h after RV39 infection. RV39 increased AP-1 transactivation more than sevenfold (Fig. 4A). TNF-
had relatively modest effects on AP-1 transactivation, and costimulation had additive effects. In contrast, TNF-
significantly increased NF-
B transactivation relative to RV39 (Fig. 4B).
|
stimulation on ERK phosphorylation. RV39 infection, but not sham infection (lysates from uninfected HeLa cells) or 3 ng/ml TNF-
, induced phosphorylation of ERK (Fig. 5A). Phosphorylation of ERK was noted 2 min after infection but returned to near-basal levels by 120 min. These data suggest that RV39 enhances TNF-
-induced IL-8 expression by activating the ERK/AP-1 signaling pathway.
|
on IL-8 expression, we employed two inhibitors of MAP kinase/ERK kinase (MEK) 1/2, U0126 and MEK-2A. U0126 at 3 µM inhibited IL-8 expression induced by the combination of RV39 and TNF-
, but not either stimulus alone (Fig. 5B). U0126 at 10 µM attenuated RV39-, TNF-
-, and RV39- and TNF-
-induced responses. Expression of a dominant-negative MEK (MEK-2A) abolished the cooperative effects of RV39 and TNF-
on transcription from the IL-8 promoter (Fig. 5C). Together, these data suggest that ERK is required for RV39-mediated augmentation of TNF-
-induced IL-8 expression. | DISCUSSION |
|---|
|
|
|---|
, which is increased in asthmatic airways (2, 7, 8) and during acute COPD exacerbations (1, 12, 21). The level of TNF-
in diseased airways has varied widely, but concentrations in the ng/ml range have been found (1, 8, 9). We found that RV39 infection significantly increases TNF-
-induced IL-8 and ENA-78 expression and that the cooperative effects on IL-8 expression were mediated in a transcriptional manner. We also found that the combination of RV and TNF-
induced synergistic increases in airway epithelial cell GM-CSF expression. GM-CSF has also been implicated in the pathogenesis of asthma and COPD exacerbations (3). Together, these data prove the principle that RV infection and proasthmatic cytokines can produce cooperative effects on airway epithelial cell chemokine expression and provide a cellular mechanism for RV-induced exacerbations of asthma and COPD.
We considered the possibility that TNF-
-induced alterations in RV receptor expression were responsible for the observed cooperative effects. TNF-
stimulation increases expression of ICAM-1 in airway epithelial cells (6, 34). In the present study, we confirmed that TNF-
increases surface ICAM-1 expression in 16HBE14o- human bronchial epithelial cells. We also found that TNF-
treatment increased the number of RV39 transcripts at 24 h after stimulation, consistent with the notion that TNF-
-induced ICAM-1 expression, with subsequent increased binding of newly formed RV39, is responsible for the observed synergistic effects between RV and TNF-
. However, RV39- and TNF-
-induced cooperative increases in IL-8 protein expression occurred as early as 4 h following stimulation, before increases in ICAM-1 expression, and before TNF-
-induced increases in viral load. Also, synergistic increases in IL-8 expression were observed 48 h after RV39 and TNF-
stimulation despite the presence of ICAM-1-neutralizing antibody, which was added 2 h after initial infection to avoid cellular binding of newly replicated RV. Finally, there was still significant synergy between replication-deficient, UV-irradiated virus and TNF-
. Since only newly replicated virus is capable of binding to TNF-
-induced ICAM-1, these data suggest that, although TNF-
increases ICAM-1 expression, this increase does not fully account for the cooperative effects between TNF-
and RV on IL-8 expression. Furthermore, these data are consistent with previous studies showing that airway epithelial cell IL-8 expression is at least partially independent of viral replication (23, 26, 32, 38, 53).
It is also conceivable that the cooperative effects of RV39 and TNF-
are due to synergistic expression of an IL-8-inducing chemokine. However, we were only able to detect a synergistic increase in IFN-
protein expression at 24 h after costimulation, suggesting that, for the most part, synergistic effects between RV and TNF-
are not due to autocrine production of IL-8-inducing cytokines.
We have previously shown that, in 16HBE14o- human bronchial epithelial cells, maximal TNF-
-induced transcription from the IL-8 promoter requires transactivation of both the NF-
B and AP-1 sites; the NF-
B site is required for TNF-
responsiveness, whereas the AP-1 site functions as a basal level enhancer. Based in part on the rapid response in IL-8 protein expression following RV and TNF-
stimulation, we hypothesized that RV augments TNF-
-induced IL-8 expression by activation of ERK/AP-1 signaling. First, we tested whether the augmented responses were transcriptionally mediated. Consistent with this mechanism, RV39 and TNF-
induced synergistic increases in IL-8 promoter activity relative to RV39 or TNF-
alone. Second, we examined the effect of RV and TNF-
on airway epithelial cell AP-1 transactivation. We found that RV39 significantly increased activity of an AP-1 reporter plasmid. Third, we found that RV39 infection, but not TNF-
, induced significant ERK phosphorylation. Finally, inhibition of ERK signaling by both chemical and genetic inhibitors blocked the cooperative effects of RV39 and TNF-
in our system. Together, these data suggest that RV augments TNF-
-induced IL-8 expression by activation of ERK/AP-1 signaling, thereby "turning up the gain" on NF-
B-dependent IL-8 expression. At later time points, increased ICAM-1 expression and viral attachment, as well as secondary cytokine expression, likely play a role.
Finally, since the cooperative effects of RV and TNF-
appeared to be greater for IL-8 protein than promoter activity, it is conceivable that RV infection increases IL-8 expression via posttranscriptional mechanisms. RV39 has been shown to induce phosphorylation of p38 MAP kinase (23), which in turn has been linked to increased IL-8 mRNA stability (35).
The observed synergy between asthma and COPD triggers is consistent with previous data from both in vitro and in vivo models. For example, previous studies from our laboratory have shown that German cockroach proteases and TNF-
have synergistic effects on human bronchial epithelial cell IL-8 expression (5, 41). RV16 has been shown to synergistically increase NO2 and O3-induced IL-8 expression in BEAS-2B cells (40). Respiratory syncytial virus increases airways hyperresponsiveness and inflammation in allergen-sensitized animals (37, 43, 47). Segmental allergen challenge increases bronchoalveolar TNF-
levels in allergic subjects, and this effect is potentiated by prior RV16 infection (9).
We conclude that RV infection enhances cytokine-induced IL-8 expression by activation of the ERK/AP-1 signaling pathway. These data provide a cellular mechanism for RV-induced exacerbations of airways disease and identify avenues for future study.
| GRANTS |
|---|
|
|
|---|
| 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 |
|---|
|
|
|---|
. Am J Physiol Lung Cell Mol Physiol 273: L749–L759, 1997.
B. Am J Physiol Lung Cell Mol Physiol 273: L814–L824, 1997.
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |