Article Text

PDF

Helicobacter pylori augments the acid inhibitory effect of omeprazole on parietal cells and gastric H+/K+-ATPase
  1. W Beila,
  2. K-F Sewinga,
  3. R Buschec,
  4. S Wagnerb
  1. aDepartment of General Pharmacology, Hannover Medical School, Hannover, Germany, bDepartment of Gastroenterology and Hepatology, Hannover Medical School, Hannover, Germany, cDepartment of Physiology, School of Veterinary Medicine, Hannover, Germany
  1. Dr W Beil, Department of General Pharmacology, Medizinische Hochschule Hannover, Carl-Neuberg-Straβe 1, D-30625 Hannover, Germany. Beil.Winfried@MH- Hannover.de

Abstract

BACKGROUND In duodenal ulcer patients, intragastric acidity during omeprazole treatment is significantly lower before Helicobacter pylori eradication than after cure.

AIMS To determine ifH pylori enhances the acid inhibitory potency of omeprazole in isolated parietal cells and on H+/K+-ATPase.

METHODS Rat parietal cells and pig gastric membrane vesicles enriched in H+/K+-ATPase activity were incubated withH pylori and the H pylori fatty acid cis9,10-methyleneoctadecanoic acid (MOA), and the inhibitory effects of omeprazole on parietal cell acid production, H+/K+-ATPase enzyme activity, and ATPase mediated proton transport were assessed.

RESULTS In isolated parietal cells, H pylori and MOA increased the acid inhibitory potency of omeprazole 1.8 fold.H pylori did not affect the inhibitory potency of omeprazole on H+/K+-ATPase enzyme activity. In proton transport studies, H pylori (intact bacteria and sonicate) and MOA accelerated the onset of the inhibitory effect of omeprazole and enhanced the proton dissipation rate in response to omeprazole. H pylori itself increased proton permeability at the vesicle membrane.

CONCLUSION Our results show that H pylori augments the acid inhibitory potency of omeprazole in parietal cells and enhances omeprazole induced proton efflux rate from gastric membrane vesicles. We suggest that omeprazole unmasks the permanent effect ofH pylori on proton permeability at the apical parietal cell membrane, which is counteracted in the absence of a proton pump inhibitor by a reserve H+/K+-ATPase capacity.

  • H pylori
  • parietal cells
  • omeprazole
  • H+/K+-ATPase

Statistics from Altmetric.com

Infection with Helicobacter pylorialters the regulatory mechanisms for gastric acid secretion. During initial colonisation, H pylori decreases acid secretion.1 2 This may be due to release of a bacterial acid inhibitory protein3 or the bacterial fatty acid cis 9,10-methyleneoctadecanoic acid (MOA) which block H+/K+-ATPase activity and H+/K+-ATPase mediated proton transport.4 During chronic infection, there appears to be no consistent effect on basal and stimulated acid secretion5 although H pyloriinfected patients have increased serum gastrin levels, enhanced gastrin responses to stimulation with gastrin releasing peptide, and decreased somatostatin levels.6-10

The proton pump inhibitor omeprazole produces higher intragastric pH values in H pylori positive subjects than in non-infected subjects.11 Recently it has been shown thatH pylori eradication in duodenal ulcer patients causes a decrease in the acid inhibitory effect of omeprazole.12 The mechanism by which H pylori augments the inhibitory effect of omeprazole is not well understood. H pylori generates ammonia, and a higher content of buffer substances in the infected stomach may lead to lower luminal acidity during omeprazole treatment. In addition, an additive effect of omeprazole and acid inhibitory substances released from H pylori is possible.

Thus the present study was undertaken to elucidate the effect ofH pylori and H pylori products on the inhibitory effect of omeprazole in parietal cells and on its target enzyme, the gastric H+/K+-ATPase.

Materials and methods

H PYLORI CULTURE

H pylori 87, a cytotoxin producingH pylori strain derived from a patient with a duodenal ulcer, was used in all study protocols.13Comparative experiments were also conducted with the non-cytotoxin producing strain H pylori 83. The presence of the vacuolating toxin was determined by incubation of the gastric cell line HM02 with concentrated supernatants fromH pylori broth filtrates, as described previously.13 Bacteria were grown in brain heart infusion broth supplemented with 10% fetal calf serum for 36 hours at 37°C under microaerophilic conditions. Bacteria were harvested by centrifugation (10 000 g, 10 minutes) and resuspended in phosphate buffered saline (parietal cell experiments) or in buffer A (2 mmol/l MgCl2, 150 mmol/l KCl, 10 mmol/l piperazine-N,N′-bis [2-ethane-sulphonic acid]/Tris, pH 7.2; H+/K+-ATPase experiments and proton transport studies). For estimation of the desired final bacterial number, optical density measurements were performed at 600 nm. For sonicates, bacterial suspensions (in buffer A) were sonicated on ice (power 50 W) in six consecutive treatments lasting 30 seconds.

PREPARATION OF PARIETAL CELLS

Two male Wistar rats were used per experiment. After removal, the stomachs were everted and filled with 1.5 ml of buffer B (70 mmol/l NaCl, 5 mmol/l KCl, 1 mmol/l Na2HPO4, 0.5 mmol/l NaH2PO4, 50 mmol/lN-[2-hydroxyethyl]piperazine-N′-[2-ethane-sulphonic acid], 20 mmol/l NaHCO3, 10 mmol/l glucose, 2 mmol/l ethylenediaminetetraacetic acid (EDTA), bovine serum albumin 20 mg/ml; pH 7.4) containing 6 mg/ml of pronase E. The everted stomachs were incubated for 2×30 minutes at 37°C in carbogen gas bubbled buffer B and 3×30 minutes in buffer C (same as buffer B except omission of EDTA, and addition of 1.2 mmol/l MgCl2 and 1.5 mmol/l CaCl2). Cells recovered from the third to fifth incubation procedures were harvested by centrifugation and stored in buffer D (same as buffer C, bovine serum concentration 1 mg/ml). The viability of isolated cells determined by trypan blue was 90–95%. Parietal cell content in the four cell preparations ranged from 28% to 35%, as identified by light microscope examination.

PREPARATION OF GASTRIC MEMBRANE VESICLES

Gastric apical membrane vesicles containing high concentrations of H+/K+-ATPase were prepared from pig stomachs, as described previously.14 In brief, scraped fundic mucosa was minced in buffer E (0.25 mol/l sucrose, 2 mmol/l EDTA, 20 mmol/l Tris HCl; pH 7.0) with a tissue chopper. The minced material was homogenised using a Potter-Elvehjem homogeniser. The 20 000g supernatant of the homogenate was centrifuged for 60 minutes at 100 000 g to obtain membrane vesicles. The vesicles were stored in buffer E at −80°C. In these “inside out” vesicles, protons are secreted by the H+/K+-ATPase into the hollow centre where H+ ions are sequestered.

MEASUREMENT OF PARIETAL CELL ACID PRODUCTION

Acid production in parietal cells was monitored by the [14C]aminopyrine uptake technique. Gastric cells (2×106 cells) were incubated with or withoutH pylori or MOA in 1 ml of buffer D at 22°C for 30 minutes. To remove H pylori, cells were spun down (300 g, five minutes) and washed three times with buffer D. Aminopyrine (8.3 μmol/l) and omeprazole (dissolved in polyethylene glycol) were added to the cells and acid production was initiated by dibuturyl cyclic adenosine monophosphate (dbcAMP 0.1 mmol/l). After 40 minutes of incubation at 37°C, cells were separated from the medium by centrifugation through silicon oil and radioactivity of intracellularly trapped aminopyrine was counted.

H+/K+-ATPase ENZYME ACTIVITY MEASUREMENTS

H+/K+-ATPase activity in gastric membrane vesicles was measured as inorganic phosphate (Pi) release from adenosine 5′-triphosphate (ATP). Pi released was analysed according to Carter and Karl.15 Vesicles (20 μg protein/ml) were incubated with or without indicated amounts ofH pylori or MOA in buffer A at 22°C for 30 minutes. The bacteria were then separated from the medium by centrifugation (10 000 g, five minutes). Subsequently, ATPase activity was initiated in 1 ml samples by adding ATP (final concentration 2 mmol/l) and the K+ ionophore valinomycin (25 μg/ml), and was stopped after 30 minutes of incubation at 37°C. Omeprazole (stock solutions in methanol) was added to the vesicles immediately before starting the enzyme reaction. Specific K+ and valinomycin stimulated ATPase activity—that is, the fraction of enzyme activity that correlates with acid accumulation—was determined by addition of 10 μmol/l SCH28080 (stock solutions in dimethylsulphoxide), a selective K+competitive inhibitor of H+/K+-ATPase.16 K+ and valinomycin stimulated ATPase activity was 35±1.6 μmol Pi/mg protein×h.

H+/K+-ATPase MEDIATED PROTON TRANSPORT MEASUREMENTS

H+/K+-ATPase mediated proton transport in membrane vesicles was monitored by fluorescence quenching of the weak base acridine orange, as described previously.17 Vesicles (60 μg protein/ml) were incubated with or without indicated amounts of H pylori, H pylori sonicate, and MOA in buffer A at 22°C for 30 minutes. After centrifugation, to remove H pylori,acridine orange (final concentration 5 μmol/l) was added to the supernatant and proton transport was initiated by adding ATP and valinomycin. Omeprazole (10 μmol/l) was added to the vesicles before starting the pump reaction or after intravesicular acidification had reached maximal levels. The maximal generated pH gradient at the vesicle membrane is up to 4 pH units—that is, intravesicular pH is approximately 3.0.17

COMPOUNDS

Omeprazole was obtained from AB Hässle (Mölndal, Sweden). SCH 28080 was supplied by Schering-Plough Corporation (Bloomfiels, New Jersey, USA). Cis9,10-methyleneoctaecanoic acid was synthesised by J Holzkamp, Department of Organic Chemistry, University of Hannover, Hannover, Germany. All other chemicals were purchased from Sigma (München, Germany). [14C]Aminopyrine was purchased from Amersham (Braunschweig, Germany).

DATA ANALYSIS

All experiments were performed with four different cell and three different vesicle preparations. Results (parietal cell and ATPase measurements) are expressed as mean (SEM). For statistical analysis, the Student's t test for paired comparisons was used. Probability values <0.05 were considered significant. The effects of H pylori and omeprazole on proton transport did not differ in the three preparations tested. The results shown (figs 3-6) are representative recordings.

Results

H PYLORI ENHANCES THE ANTISECRETORY POTENCY OF OMEPRAZOLE IN PARIETAL CELLS

H pylori(108–109 bacteria/ml=50–500 cfu/gastric cell) progressively inhibited dbcAMP stimulated acid production in parietal cells. An inhibitory effect was detected at bacteria/cell ratios ⩾150, being half maximal at a bacteria/cell ratio of 250. Omeprazole inhibited dbcAMP stimulated aminopyrine uptake with an IC50 value of 0.074 (0.01) μmol/l. The inhibitory action of omeprazole at low concentrations (0.05 and 0.1 μmol/l) was enhanced in cells exposed to H pylori (50 cfu/gastric cell); the IC50 value was shifted to 0.042 (0.009) μmol/l (p<0.05 v omeprazole alone) (fig 1). In common with H pylori, theH pylori fatty acid MOA (5–50 μmol/l) inhibited dbcAMP stimulated acid production at concentrations ⩾25 μmol/l; the IC50 value was 46 (8) μmol/l. The antisecretory effect of omeprazole was significantly enhanced in cells treated with 5 μmol/l MOA; the IC50 value for omeprazole was shifted to 0.044 (0.01) μmol/l (fig 1).

Figure 1

Effect of H pylori and cis 9,10-methyleneoctadecanoic acid (MOA) on the inhibitory action of omeprazole on dibuturyl cyclic adenosine monophosphate (dbcAMP) stimulated acid production in rat parietal cells. Gastric cells (2×106 cells, parietal cell content 28–35%) were incubated without (untreated), or with H pylori (50 cfu/gastric cell) or MOA (5 μmol/l). Thereafter, H pylori was removed from the cells. Omeprazole was added to the cells and acid production was initiated by dbcAMP (0.1 mmol/l). After 40 minutes of incubation at 37°C, intracellularly trapped aminopyrine was counted. dbcAMP stimulated aminopyrine uptake in untreated, H pylori, and MOA treated cells was 115 (37), 117 (40), and 124 (12) pmol aminopyrine/105parietal cells, respectively. These values were set to 100%. Values are mean (SEM) of four cell preparations. C, control. *p<0.05 v omeprazole alone.

H PYLORI DOES NOT ALTER THE INHIBITORY ACTION OF OMEPRAZOLE ON H+/K+-ATPase ENZYME ACTIVITY

Omeprazole inhibited K+ and valinomycin stimulated ATPase activity—that is, the fraction of enzyme activity that correlates with acid accumulation in the vesicle interior—in a concentration dependent manner with an IC50 value of 0.7 (0.04) μmol/l (fig 2). H pylori(1.6×105 to 8×106 bacteria/μg membrane protein) slightly reduced ATPase activity; at bacterial loads of 1.6×106 and 8×106 bacteria/μg membrane protein, enzyme activity was reduced by 10% and 37%, respectively. The inhibitory effect of omeprazole was not altered in vesicles exposed to H pylori (5×105 and 2.5×106 bacteria/μg membrane protein) (fig2).

Figure 2

Effect of H pylori on the inhibitory effect of omeprazole on gastric H+/K+-ATPase activity. Gastric membrane vesicles were incubated without (untreated) or with H pylori 5×105 or 2.5×106 bacteria/μg membrane protein. Bacteria were then separated from the medium. Omeprazole was added to the vesicles and ATPase activity was initiated by adding ATP and the K+ ionophore valinomycin. Values are mean (SEM). C, control.

H PYLORI AUGMENTS THE INHIBITORY ACTION OF OMEPRAZOLE ON H+/K+-ATPase MEDIATED PROTON UPTAKE (PUMP ACTIVITY)

Addition of ATP and valinomycin to gastric membrane vesicles induced quenching of acridine orange fluorescence, demonstrating proton transport into vesicles and formation of an acidic interior pH gradient. The maximal pH gradient obtained represents an equilibrium between proton uptake into vesicles and leakage of protons outwards (fig 3, curve A). Omeprazole (10 μmol/l), added to the vesicles before starting the pump reaction, had no effect on the initial rate of acidification. As the vesicles were acidified, omeprazole produced a time dependent dissipation of the pH gradient (fig 3, curve C). The inhibitory characteristic of omeprazole was altered in vesicles exposed to H pylori (8×105bacteria/μg membrane protein). The proton efflux rate in response to omeprazole was enhanced and the time lag for the onset of efflux was shortened (fig 3, curve D). It should be noted that at this bacterial load, H pylori alone had no effect on proton transport into vesicles or the maximal acid interior pH gradient (fig3, curve B). In fig 4, omeprazole was added under equilibrium conditions after intravesicular acidification had been established. Under these conditions omeprazole induced dissipation of the pH gradient after a short time lag of approximately one minute (fig 4, curve C). The time lag for the response represents the time necessary for the transformation of omeprazole into the cyclic sulphenamide and the interaction of this compound with the SH groups of the H+/K+-ATPase. In vesicles which had been pretreated with H pylori(1.6×105 bacteria/μg membrane protein), omeprazole immediately induced dissipation of the pH gradient and the proton dissipation rate was enhanced (fig 4, curve D). At higher bacterial loads (1.6×106 bacteria/μg membrane protein),H pylori reduced the initial acid interior pH gradient and gradually dissipated the established pH gradient (fig4, curve E). Even with reduced intravesicular proton concentration, omeprazole immediately induced dissipation of the established pH gradient (fig 4, curve F).

Figure 3

Effect of omeprazole on H+/K+-ATPase mediated H+ transport in gastric membrane vesicles exposed to H pylori cells. Gastric vesicles (60 μg protein/ml) were incubated with and without H pylori (8×105 bacteria/μg membrane protein) in 2 ml of 10 mmol/l piperazine-N,N′-bis [2-ethane-sulphonic acid]/Tris buffer containing 2 mmol/l MgCl2 and 150 mmol/l KCl. After 30 minutes of incubation, H pylori was separated by centrifugation. Acridine orange and omeprazole (10 μmol/l) were added and proton transport was initiated by adding ATP and valinomycin (arrow).

Figure 4

Effect of omeprazole on H+/K+-ATPase mediated H+ transport in gastric membrane vesicles exposed to H pylori cells. Gastric membrane vesicles were incubated with H pylori (1.6 and 8×105 bacteria/μg membrane protein) as described in fig2. Proton transport was initiated by addition of ATP and valinomycin (arrow) and omeprazole was added (↑) when the pH gradient had reached its maximum value.

EFFECTS OF H PYLORI PRODUCTS AND MECHANISMS OF THE EFFECTS OF H PYLORI

To investigate how H pyloriregulates the inhibitory action of omeprazole on H+/K+-ATPase mediated proton uptake, the influence of H pylori sonicate and the majorH pylori fatty acid MOA18 on the inhibitory action of omeprazole was studied. Similar results as those obtained with intact bacteria were achieved.H pylori sonicate (1.6×105sonicated bacteria/μg membrane protein) and MOA (1 and 5 μmol/l), which alone had no effect on the vesicular pH gradient, enhanced proton efflux rate in response to omeprazole (fig 5). Apart from MOA,H pylori produces tetradecanoic acid and octadecanoic acid.18 In common with MOA, 10 μmol/l tetradecanoic acid (fig 5) and octadecanoic acid (data not shown) enhanced the proton dissipation rate in response to omeprazole. The possible role of the vacuolating toxin was investigated by incubation of vesicles with the cytotoxin negative strain H pylori 83. Similar to the cytotoxin positive strainH pylori 87, H pylori 83 also enhanced the effect of omeprazole to a similar degree (data not shown). At high bacterial loads (1.6×106bacteria/μg membrane protein), H pyloriitself reduced the intravesicular proton concentration (fig 4, curve E) while at bacterial loads of 1.6×105 and 8×105 bacteria/μg membrane protein,H pylori did not influence vesicular proton concentrations (figs 3, 4). H pyloriincreases proton permeability at the vesicular membrane4and it seems likely that this effect is fully compensated for by proton pump activity at bacterial loads of 1.6×105 and 8×105 bacteria/μg membrane protein. To address this issue in more detail, we studied the effect of H pylori on membrane permeability by inhibition of the ATPase reaction using the Mg2+ chelating agent EDTA. Figure 6shows that in vesicles treated with H pylori(8×105 and 1.6×106 bacteria/μg membrane protein), release of protons from the vesicle interior into the buffer solution was increased. An identical effect was seen after treating vesicles with H pylori sonicate (1.6×105 sonicated bacteria/μg membrane protein), MOA (1 and 5 μmol/l), and tetradecanoic acid (10 μmol/l) (data not shown).

Figure 5

Effect of omeprazole on H+/K+-ATPase mediated H+ transport in gastric membrane vesicles exposed to a H pylori sonicate (1.6×105 sonicated bacteria/μg membrane protein), cis 9,10-methyleneoctadecanoic acid (MOA) (1 and 5 μmol/l), and tetradecanoic acid (10 μmol/l). Experimental conditions were as described in fig 2. Proton transport was initiated by addition of ATP and valinomycin (arrow) and omeprazole was added (↑) when the pH gradient had reached its maximum value.

Figure 6

Effect of the magnesium chelating agent ethylenediaminetetraacetic acid (EDTA) on intravesicular acidity in gastric membrane vesicles treated with and without H pylori cells (8×105 and 1.6×106 bacteria/μg membrane protein). Experimental conditions were as described in fig 2. Proton transport was initiated by addition of ATP and valinomycin (arrow) and EDTA was added (↑) to stop the ATPase reaction when the pH gradient had reached its maximum value.

Discussion

Suggestions for the mechanism by which H pylori augments the pH increasing effect of omeprazole include: (i) presence of a higher content of acid neutralising substances (ammonia) in the infected gastric mucosa; (ii) increased levels of cytokines such as interleukin 1β which is known to inhibit gastric acid secretion19; and (iii) exposure of parietal cells to acid inhibitory substances released by H pylori.

We have demonstrated that H pylori increases the antisecretory effect of low concentrations of omeprazole (0.05 and 0.1 μmol/l) in parietal cells, as measured by uptake and retention of [14C]aminopyrine within the secretory canaliculus of parietal cells (fig 1). H pylori and omeprazole interfere with parietal cell H+ transport by different mechanisms. Omeprazole is an acid activated pro-drug. In line with this mechanism of action, we found that the inhibitory effect of omeprazole on H+/K+-ATPase mediated H+ transport in gastric membrane vesicles was pH and time dependent (fig 3). Inhibition of H+/K+-ATPase is correlated with the amount of omeprazole converted to the active inhibitor, the cyclic sulphenamide. At neutral pH the conversion is slow but with lowering of the pH, the rate of transformation is increased.20 21 H pyloriinterferes with parietal cell acid production by two mechanisms: (i) the bacterium increases proton permeability at the secretory membrane of the parietal cell—that is, the bacterium causes back diffusion of protons from the secretory canaliculus into the cytosol of the parietal cell—and (ii) in addition inhibits H+/K+-ATPase enzyme activity.4 The fact that H pylori (1.6×106bacteria/μg membrane protein) causes marked release of H+from intact H+/K+-ATPase vesicles (see fig 4, curve E) but had only a weak inhibitory effect on H+/K+-ATPase enzymatic activity (approximately 10% enzyme inhibition) suggests that the antisecretory effect in parietal cells is predominantly a result of increasing proton permeability at the secretory parietal cell membrane.

H pylori did not alter the inhibitory potency of omeprazole on H+/K+-ATPase enzyme activity (fig 2), indicating that H pyloridoes not interfere with the transformation of omeprazole into the active inhibitor, the cyclic sulphenamide, or binding of sulphenamide to the enzyme. In contrast, in the proton transport studies,H pylori (1.6 and 8×105bacteria/μg membrane protein) enhanced the onset and extent of the proton dissipation rate in response to omeprazole (figs 3, 4). As outlined above, omeprazole inhibits H+/K+-ATPase mediated H+ uptake into gastric vesicles only when acid is accumulated in the vesicle lumen. When vesicle acidification is reduced the inhibitory effect of omeprazole is decreased.14 Interestingly, the extent of the proton efflux rate in response to omeprazole was also increased when the intravesicular proton concentration was reduced by a high bacterial load (1.6×106 H pylori/μg membrane protein) (fig 4 F). H pylori increases proton permeability at the vesicular membrane. This effect was detected when the pump reaction was arrested by EDTA (fig 6). Increased proton permeability could not be detected in pumping vesicles, indicating that there is a reserve capacity of H+/K+-ATPase which compensates forH pylori induced proton efflux. When H+/K+-ATPase mediated proton transport is blocked by omeprazole (or EDTA), this effect of H pylori on proton permeability is unmasked—that is, omeprazole induced dissipation of the pH gradient due to inhibition of ATPase is accelerated by the permanent effect of H pylori on proton permeability.

All effects of H pylori were mimicked by the most abundant H pylori fatty acids MOA and tetradecanoic acid.18 22 The absolute amount of MOA and tetradecanoic acid in 107 H pylori is 201 ng and 172 ng, respectively.22 On the basis of these data we assume that the vesicles were exposed to about 0.7 μmol/l MOA and 0.8 μmol/l tetradecanoic acid during the experiments with the H pylori sonicate (1.6×105 sonicated bacteria/μg membrane protein=107 bacteria/ml). As the H pylori sonicate induced a greater effect than MOA (1 μmol/l), we suggest that there are additive effects between MOA and tetradecanoic acid. However, we cannot exclude the fact that the effect of H pylori may not be solely mediated by its fatty acids.

In conclusion, we have shown that H pylorienhances the inhibitory potency of omeprazole in parietal cells.H pylori increases proton permeability at the secretory parietal cell membrane and this effect is balanced at low bacterial load by compensatory transport of protons by the H+/K+-ATPase into the vesicle. When H+/K+-ATPase is inhibited by omeprazole (or other ATPase inhibitors), this effect of H pylori on proton permeability is unmasked. Proton efflux is enhanced—that is, there are additive effects between omeprazole andH pylori. Recognising that the results of our in vitro study cannot be extrapolated directly to the interaction of H pylori with omeprazole in vivo, the findings suggest that H pylori may interact with proton pump inhibitors to augment the antisecretory effect of these drugs at the secretory parietal cell membrane.

Acknowledgments

This work was supported by a grant from the Deutsche Forschungsgemeinschaft (Sonderforschungsbereich 280, TP A7). The skilful technical assistance of Ute Staar and Helga Hannemann is gratefully acknowledged.

References

View Abstract

Footnotes

  • Abbreviations used in this paper:
    ATP
    adenosine 5′-triphosphate
    EDTA
    ethylenediaminetetraacetic acid
    MOA
    cis9,10-methyleneoctadecanoic acid
    dbcAMP
    dibuturyl cyclic adenosine monophosphate

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.