Recent evidence suggests that the epoxyeicosatrienoic acids (EETs), which are products of cytochrome P450 (CYP) epoxygenases, possess mitogenic and angiogenic effects in vascular endothelial cells. However, the mechanisms underlying these effects are not fully elucidated. Because sphingosine kinase (SK) and its product S1P play essential roles in cell growth, survival and migration, we hypothesized that SK activation by EETs may mediate some of its angiogenic effects.
We studied the effects of EETs on SK activity in human umbilical vein endothelial cells (HUVECs). Treatment with EETs, particularly 11,12-EET, markedly augmented SK activity in HUVECs. At the concentration of 1 µmol/L, 11,12-EET increased SK activity by 110% and the maximal effect on SK activation was observed at 20 min after 11,12-EET addition. Furthermore, inhibition of SK by a specific inhibitor, SKI-II, markedly attenuated 11,12-EET-induced EC proliferation. Importantly, 11,12-EET-induced activation of Akt kinase and transactivation of the epidermal growth factor (EGF) receptor was also inhibited by SKI-II. To investigate the isoform-specific role of SK in EET-induced angiogenesis, inhibition of SK1 by expression of dominant-negative SK1(G82D) substantially attenuated 11,12-EET-induced EC proliferation, migration, and tube formation in vitro and Matrigel plug angiogenesis in vivo. Furthermore, knockdown of SK1 expression by specific siRNA also inhibited 11,12-EET-induced EC proliferation and migration, whereas SK2 siRNA knockdown was without effect.
These results suggest that SK1 is an important mediator of the 11,12-EET-induced angiogenic effects in human ECs. Thus, SK1 may represent a novel therapeutic modality for the treatment of angiogenesis-related diseases such as cancer and ischaemia.
The cis-epoxyeicosatrienoic acids (EETs) are vasoactive eicosanoid products of cytochrome P450 (CYP) epoxygenases.1,2 The EETs have properties similar to those of endothelium-derived hyperpolarizing factor (EDHF) because they hyperpolarize and relax vascular smooth muscle cells (SMCs) by activating calcium-sensitive potassium (KCa) channels.3,4 Recently, several CYP epoxygenases, including members of the CYP2B, CYP2C, and CYP2J subfamilies, have been identified in vascular endothelial cells and shown to metabolize arachidonic acid to EETs (5,6-EET, 8,9-EET, 11,12-EET, and 14,15-EET). However, their relative importance in endothelial EET biosynthesis has not yet been determined. Treatment of porcine coronary artery endothelial cells, in vitro, with either β-naphthoflavone or nifedipine induces CYP2C expression, increases 11,12-EET biosynthesis, and enhances bradykinin-induced coronary artery relaxation via SMC membrane hyperpolarization.1,5 Moreover, transfection of endothelial cells with an antisense oligonucleotide to CYP2C8/9, or treatment with the selective CYP2C9 inhibitor sulphaphenazole, attenuates EDHF-mediated vascular responses, thus providing supporting evidence that the EDHF synthase in the porcine vascular bed may be a CYP2C isoform.1,5
In contrast to their vasodilatory effects, EETs have also been shown to possess numerous membrane potential-independent effects on endothelial cell signalling and vascular homeostasis. For example, EETs, which are produced by a member of the CYP2J family, CYP2J2, inhibit cytokine-induced endothelial cell adhesion molecule expression by inhibiting the proinflammatory transcriptional factor, NF-кB.6 Furthermore, addition of physiological relevant concentrations of EETs, or overexpression of CYP2J2, increases both tissue plasminogen activator (tPA) expression and fibrinolytic activity in endothelial cells, and inhibits vascular SMC migration via a cAMP-mediated mechanism.7,8 However, neither the anti-inflammatory, anti-migratory nor the fibrinolytic actions of EETs were blocked by KCa channel inhibitors, indicating that these EET effects were independent of their membrane-hyperpolarizing effects.6–8 Furthermore, accumulating evidence suggests that EETs play a critical role in endothelial proliferation, migration, and angiogenesis.9 Exogenously addition of EETs and the overexpression of CYP epoxygenases, were found to inhibit endothelial cell apoptosis and induce endothelial cell proliferation and angiogenesis, via a mechanism involving both the activation of the ERK and PI3K/Akt pathways and the transactivation of the epidermal growth factor (EGF) receptor.10–12
SK is a key enzyme catalyzing the phosphorylation of sphingosine to sphingosine-1-phosphate (SPP or S1P). Recent studies have shown that S1P is a potent lipid messenger that plays important and diverse roles in biological processes such as cell growth, proliferation, and survival and calcium homeostasis, by directly acting on G protein-coupled S1P receptors and/or functioning as an intracellular second messenger.13 To date, five receptors (EDG-1/S1P1, EDG-5/S1P2, EDG-3/S1P3, EDG-6/S1P4, and EDG-8/S1P5) have been shown to bind to both S1P and sphinganine-1-phosphate (dihydro-S1P) with high specificity.13 These receptors, via coupling to different G proteins, exert a wide array of cellular responses, including vascular maturation, angiogenesis, and migration and cardiac development.13 In particular, accumulating evidence has suggested that SK, and its derived product S1P, are key mediators in angiogenesis.13,14 Thus, the mechanisms mediating EET-induced angiogenesis are not well understood. Considering the significant importance of the SK/S1P signalling pathway in regulating cell survival, proliferation, and migration, we hypothesized that SK is an important modulator of angiogenesis induced by EETs. In the present study, we examined whether SK is activated in ECs in response to EET stimulation, and whether the activation of SK is involved in EET stimulated angiogenesis.
EETs were obtained from Cayman Chemical. VPC23019 was purchased from Avanti Polar Lipids, Inc. SKI-II and pertussis toxin were obtained from Sigma. Calcein AM was purchased from Molecular Probes, Inc. [3H]Thymidine was purchased from PerkinElmer Life and Analytical Sciences.
Human umbilical vein endothelial cells (HUVECs) were purchased from Clonetics and cultured in medium199 (M-199; BioWhittaker) containing endothelial cell growth supplement, 10% foetal bovine serum, and antibiotics (100 U/mL penicillin, 10 µg/mL streptomycin, and 20 µg/mL neomycin).
Western blotting and immunoprecipitation
Proteins were prepared and separated on SDS–PAGE as described15,16 Immunoblotting was performed using monoclonal antibodies to phospho-Akt (Ser473), and Akt1 (1:1000 dilution; Cell Signaling Tech), to anti-EGFR and anti-phospho-EGFR (1:1000 dilution, Transduction Laboratories). Immunodetection was accomplished using a sheep anti-mouse secondary antibody (1:2000 dilution) or donkey anti-rabbit secondary antibody (1:2000 dilution) and the enhanced chemiluminescence (ECL) kit (Amersham Corp.). Immunoprecipitation was performed essentially as previously described.15
Gene silencing with siRNA
Chemically synthesized siRNA duplexes were purchased from QIAGEN. The siRNA targeted sequences were as follows: GGGCAGGCATATGGAGTAT (SK1), CCTCATCCAGACAGAACGA (SK2), and TTCTCCGAACGTGTCACGT (scrambled control siRNA). siRNA duplexes were transfected to the cells with Oligofectamine (Invitrogen). Forty eight hours after transfection, the targeted gene expression levels were detected by RT–PCR and immunoblot analysis.17
Thymidine incorporation assay
HUVECs plated in 24-well plates were quiesced for 24 h in M-199 containing 1% FBS for 24 h. 11,12-EET was added for 24 h in the presence or absence of inhibitors. Cells were then labelled with 1 µCi of [3H]thymidine/well for 12 h, washed with phosphate-buffered saline (PBS), and fixed with cold methanol. Proteins were precipitated with a series of trichloroacetic acid washes. Following the addition of 1 N NaOH for 30 min and then 1 N HCl, lysates were collected and [3H]thymidine incorporation was assessed by scintillation quantification.
Cell migration assay
Endothelial cell migration was conducted as previously described.16 Briefly, quiescent HUVECs were washed with Hanks’ balanced salt solution, trypsinized, and neutralized with trypsin neutralizing solution. 2×104 cells in suspension were then added to the upper chamber. The bottom chamber was filled with 500 µL medium containing vehicle or 11,12-EET at the indicated concentrations. The assembly was then incubated at 37°C to allow cell migration. After 8 h, the cells were fixed and stained by Diff-Quik solutions (VWR Scientific). Cells that did not migrate through the membrane were gently removed from the upper surface. Cells were counted in five randomly selected squares/well under a light microscope and presented as the number of migrated cells/field.
Tube formation assay
The 96-well culture plates were coated with 105 µL of growth factor-reduced Matrigel per well and then allowed to polymerize for 30 min at 37°C. HUVECs cultured for 24 h in M-199 with 1% foetal bovine serum were seeded on coated plates at a density of 2×104 cells per well in M-199 supplemented with 1% foetal bovine serum and the agents as indicated in the Figure legend and then incubated for 18 h at 37°C. Cells were washed with Hank balanced salt solution and stained with Calcein AM (8 µg/mL in Hank balanced salt solution). Pictures were taken at ×40 magnification with a digital output camera (Olympus DP11) attached to an inverted phase-contrast microscope (Olympus IX70); total tube length was measured by using the NIH Image program (National Institutes of Health, Bethesda, MD, USA) as described.16
Matrigel plug angiogenesis assay
The Matrigel plug assay was performed as previously described.18 The investigation conforms to the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No. 85-23, revised 1996). Briefly, C57BL/6 mice (8 weeks old) were lightly anesthetized with sodium pentobarbital (50 mg/kg intraperitoneally) and injected subcutaneously with 0.5 mL of Matrigel premixed with vehicle or 100 µM 11,12-EET. To test the effects of Ad-LacZ and Ad-DNSK1 on 11,12-EET-induced angiogenesis, the adenoviruses (5×109 pfu/mL) were added to the Matrigel prior to injection into the mice. One week or two weeks later, the animals were killed by inhalation of CO2, and the Matrigel plugs were retrieved from underneath the skin and analyzed for haemoglobin with Drabkin’s reagent (Sigma).
Sphingosine kinase activity assay
SK activity was determined as previously described.19 Briefly, after incubation with or without EETs, cells were washed with ice-cold PBS and homogenized in kinase buffer.19 The SK assay was performed with cell extracts supplemented with 50 µM D-erythro-sphingosine dissolved in 0.1% Triton X-100 and γ-[32P]ATP for 30 min at 37°C. The [32P]ATP-labelled sphingosine 1-phosphate (S1P) was extracted and separated by thin layer chromatography on Silica Gel G60 (Whatman) using chloroform/methanol/acetic acid/water (90:90:15:6). The radioactive spots corresponding to sphingosine phosphate were scraped and counted in a scintillation counter.
Sphingosine-1-phosphate level determination
HUVECs were seeded at 1×106 cells/well in six-well plates and incubated for 4 h with phosphate-free DMEM and metabolically labelled with [32P]orthophosphate (20 µCi/mL) for 4 h. Cells were washed with phosphate-free DMEM and treated with 1 µmol/L EETs for 20 min. Cells were then scraped in 400 µL methanol and 25 µL HCl was added. Lipids were extracted by the addition of 400 µL chloroform, 400 µL KCl, and 40 µL 3 M NaCl. The aqueous phase was mixed with 50 µL HCl and then with 400 µL chloroform, and lipid phase was extracted. The lipid phase extracts were combined, dried under vacuum, and resolved by TLC as above.
All data are expressed as mean ± SE and analyzed by means of one- or two-way ANOVA and Fisher exact test for post hoc analyses. A value of P < 0.05 was considered statistically significant.
3.1 Epoxyeicosatrienoic acids stimulate sphingosine kinase activity
HUVECs were treated with EET regioisomers for 30 min and cell lysates were then subjected to SK activity assay. As shown in Figure 1A, at the concentration of 1 µmol/L, the 8,9-EET, 11,12-EET and 14-15-EET regioisomers significantly stimulated SK activity by over 100%, whereas the 5,6-EET exhibited a modest, but still significant stimulatory effect on SK activity. Consistent with the increase in SK1 activity in EET treated cells, the 8,9-EET, 11,12-EET and 14-15-EET regioisomers significantly increased intracellular S1P levels, whereas the 5,6-EET regioisomer did not significantly increase the intracellular S1P levels (Figure 1B). These results suggest that there is some degree of regioselectivity in the stimulatory effects of the EETs on SK activity in HUVECs. In a concentration-dependent manner, 11,12-EET activated SK with an EC50 of 0.1 µmol/L (Figure 1C). The maximal stimulatory effect of 11,12-EET on SK activation was observed at 1 µmol/L, with increased SK activity by 105% (Figure 1C). The time course of SK activation, when incubated with 1 µmol/L 11,12-EET, showed a maximum value at 20 min after 11,12-EET addition (Figure 1D).
Sphingosine kinase mediates epoxyeicosatrienoic acid-stimulated endothelial cell proliferation
EET has been shown to stimulate EC proliferation.11,20 Since SK plays a critical role in the regulation of cell proliferation, we investigated whether SK can mediate some of the stimulatory effects of 11,12-EET on EC proliferation. We found that treatment of HUVECs, with 1 µmol/L of 11,12-EET, significantly stimulated endothelial cell proliferation, as determined by [3H]-thymidine incorporation (Figure 2A). However, in the presence of SK specific inhibitor SKI-II (3 µmol/L),21 11,12-EET-induced EC proliferation was markedly attenuated (Figure 2A). Furthermore, recombinant adenovirus mediated overexpression of dominant negative DNSK1 (G82D) (Ad-DNSK1)22 substantially inhibited 11,12-EET (1 µmol/L) induced SK activity in HUVECs (Figure 2B), suggesting that SK1 is the major isoform of SK selectively activated by 11,12-EET. Accordingly, 11,12-EET stimulated cell proliferation was also robustly attenuated, whereas overexpression of LacZ was without effect (Figure 2C).
To verify whether S1P receptors are involved in the 11,12-EET-induced EC proliferation, we treated HUVECs with pertussis toxin (PTX), an inhibitor of Gi proteins. As shown in Figure 2D, treatment of HUVECs with PTX (100 ng/mL) markedly inhibited 11,12-EET (1 µmol/L) induced cell proliferation. Moreover, treatment of HUVECs with VPC23019 (10 µmol/L), a selective S1P1/S1P3 receptor antagonist,23 also substantially attenuated 11,12-EET-induced EC proliferation. Taken together, these results suggest that the Gi protein coupled S1P1/S1P3 receptor may account for the SK1-mediated stimulatory effects of 11,12-EET on EC proliferation.
Sphingosine kinase-1 mediates epoxyeicosatrienoic acid induced angiogenesis
Thus so far, two isoforms of SK have been identified, namely SK1 and SK2.24,25 To further substantiate the isoform-specific role of SK in the EET-induced angiogenesis, we first investigated the effect of overexpression of dominant negative SK1 (DNSK1) on 11,12-EET-induced tube formation in HUVECs. Quiescent HUVECs transduced with either Ad-LacZ or Ad-DNSK1, for 48 h, were seeded onto 96-well plates coated with growth factor-reduced Matrigel and then treated with 11,12-EET for 18 h, and tube formation was measured.16 At the concentration of 1 µmol/L, 11,12 EET stimulated Ad-LacZ transduced HUVEC tube formation by 95% (Figure 3A and B). Transduction of Ad-DNSK1 substantially inhibited 11,12-EET-induced tube formation. We next determined the effect of DNSK1 on HUVEC migration using a modified Boyden chamber method.16 Similar to the effect of DNSK1 on tube formation, overexpression of DNSK1, but not LacZ, substantially inhibited 11,12-EET-induced HUVEC migration (Figure 3C). Together, these results further suggest that SK1, but not SK2, is an important mediator for 11,12-EET-induced angiogenesis.
To examine whether SK1 mediates 11,12-EET mediated angiogenesis in vivo, we performed Matrigel plug angiogenesis in mice. SK inhibitor SKI-II (10 µmol/L) with and without 100 µmol/L 11,12-EET was mixed with growth factor-reduced Matrigel and injected into mice underneath the skin. Seven days later, the Matrigel plugs were retrieved and haemoglobin content was analyzed using Drabkin’s reagent. As shown in Figure 4A, 11,12-EET-induced angiogenesis by 1.8-fold, which was substantially inhibited in the presence of SK inhibitor SKI-II. Consistent with previous studies, PI3K inhibitor LY294002 (250 µM) also significantly inhibited 11,12-EET-induced in vivo angiogenesis (Figure 4A). Furthermore, adenovirus-mediated expression of DNSK1, but not LacZ, substantially blocked 11,12-EET-induced angiogenesis (Figure 4B). Taken together, these results further indicate that SK1 is an important mediator in 11,12-EET-induced angiogenesis both in vitro and in vivo.
Knockdown of sphingosine kinase-1 abolishes epoxyeicosatrienoic acid-induced endothelial cell proliferation and migration
To further investigate whether activation of SK1 is indeed responsible for the 11,12-EET-induced angiogenesis, we employed an siRNA approach to knockdown the expression of SK1 and SK2 in HUVECs. Transfection of SK1 siRNA substantially reduced SK1 expression by ∼80%, as detected by both RT–PCR and immunoblot analysis (Figure 5A). Likewise, SK2 siRNA also significantly inhibited SK2 expression by 70%, as detected by RT–PCR (Figure 5A). In addition, 11,12-EET-induced SK activation was markedly blocked by SK1 specific siRNA transfection (Figure 5B). In contrast, transfection of SK2 siRNA resulted in a reduction of total SK activity, but had no effect on 11,12-EET-induced SK activation (Figure 5B), suggesting that 11,12-EET mainly activates SK1, but not SK2, in HUVECs. Likewise, the stimulatory effects of 11,12-EET on endothelial cell proliferation and migration were also markedly attenuated in the SK1 specific siRNA transfected HUVECs. In contrast, siRNA knockdown of SK2 expression was without effect (Figure 5C and D). These results further suggest that SK1 activation is mainly responsible for the 11,12-EET-induced angiogenic effects in HUVECs.
Sphingosine kinase-1 mediates epoxyeicosatrienoic acid-induced phosphorylation of Akt and transactivation of epidermal growth factor receptor
It has previously been shown that EET stimulates endothelial DNA synthesis and angiogenesis via activation of PI3K/Akt signalling pathway.9 Furthermore, the activation of SK has also been implicated in the regulation of PI3K/Akt pathway in endothelial cells.26 Therefore, to determine whether SK1 plays a role in EET-induced Akt activation, we studied the effect of 11,12-EET on Akt activation in HUVECs in the presence of SK inhibitor SKI-II. Treatment of quiescent HUVECs with 11,12-EET (1 µmol/L, 30 min) resulted in an increased phosphorylation of Akt, which was markedly attenuated in the presence of SKI-II (3 µmol/L) (Figure 6A). Interestingly, the 11,12-EET-induced transactivation of EGFR was also substantially inhibited by SKI-II (Figure 6B). Together, these results suggest that SK1 is an important upstream target in EET-induced activation of Akt kinase and EGFR cascades.
In the present study, we demonstrated the role of SK1 in EET-induced EC proliferation and angiogenesis. We showed that 11,12-EET activates SK1 to generate S1P in endothelial cells. Generation of S1P, by mainly acting on Gi protein-coupled S1P1 and S1P3 receptors in HUVECs, leads to the activation of Akt kinase and the transactivation of EGF receptor, which results in angiogenesis.
Recently, it has been demonstrated that CYP-derived EETs may function as intracellular signal transduction molecules to exert diverse biological activities, via the activation of kinase cascades and modulation of gene expression in a variety of cell types.27 In ECs, overexpression of CYP 2C9, and exogenous application of 11,12-EET, induce cell proliferation and angiogenesis.11,12 The mechanisms mediating EET-induced angiogenic effects, however, are not fully elucidated. Several signalling pathways, including the PI3K/Akt pathway, the MAP kinase pathway and the cAMP/PKA pathway, have been proposed, depending on the species, type of endothelial cells, and the EET regioisomer that initiates the process.9 In HUVECs, the PI3K/Akt pathway seems to be a major signalling pathway mediating the 11,12-EET-induced cell proliferation and angiogenesis.11 Treatment of HUVECs with 11,12-EET, or overexpression of CYP2C9, has been reported to result in a decreased expression of the cyclin D1 inhibitor, p27Kip1, and an increased expression of cyclin D1, both of which are dependent on the activation of the PI3K/Akt pathway.11 Moreover, the PI3K/Akt pathway has also been implicated in the EET initiated angiogenic response in bovine aortic endothelial cells and murine pulmonary endothelial cells.10,28 However, at present, little is known about how 11,12-EET activates the PI3K/Akt pathway in endothelial cells. Although some evidence indicates that an EET receptor may exist,29,30 it has not been purified or cloned thus far. Accordingly, the presence of an EET receptor is questionable. The present study reveals a new mechanism, namely, that 11,12-EET activates the PI3K/Akt pathway by means of activation of SK1 in endothelial cells, albeit the mechanism by which 11,12-EET activates SK1 in endothelial cells remains to be defined. In addition to the PI3K/Akt pathway, the transactivation of EGFR has also been implicated in EETs-induced angiogenesis.12,28 The mechanism underlying the EETs-induced transactivation of EGFR in endothelial cells, however, remains completely unknown. In the present study, our data suggest that 11,12-EET transactivates EGFR, via the activation of SK1, and also probably acts via Gi protein-coupled S1P1/S1P3 receptors in ECs. In this regard, our results are consistent with the previous notion that SK1 activation mediates oestrogen-induced transactivation of EGFR in human breast cancer cells.31 Given that SK1 contributes to both the PI3K/Akt and the EGFR signalling cascades, activation of SK1 by 11,12-EET may represent an important mechanism through which EETs exert mitogenic and angiogenic effects in ECs.
Tumour angiogenesis is characterized by the formation of new irregular blood vessels from a pre-existing vascular network. The abnormal angiogenesis is required for the growth, survival, and metastasis of most solid tumours.32 Interestingly, cytochrome P450 2J2 (CYP2J2), a predominant P450 arachidonic acid epoxygenase that generates all four EETs, has recently been shown to be highly expressed in certain cancer cell lines and cancer tissues, and contribute to the promotion of in vivo tumour formation.33 Since angiogenesis plays a critical role in the process of tumour growth and metastasis,32 it is plausible that an increased generation of EETs, by CYP2J2, in cancer tissues, may act on endothelial cells through a paracrine pathway to promote angiogenesis, thereby leading to the tumour growth and metastasis. Indeed, SK1 is emerging as a promising target for cancer therapy. SK1 activity was found to be significantly higher in various tumour tissues than in normal tissues.34–36 Specific inhibition of SK1, by an inhibitor SKI-II, significantly inhibited tumour growth in mice.21,37 It would be interesting to determine whether or not the anti-angiogenic property of SKI-II contributes to its anti-tumour effects.
SK1 activation results in an increased production of its product, S1P, which functions by directly acting on G protein-coupled S1P receptors, and/or functioning as an intracellular second messenger.13 S1P1 and S1P3 have been shown to be abundantly expressed in HUVECs.38 S1P1 mainly couples to Gi, whereas S1P3 activates Gi, Gq, and G12/13.39 In the present study, we employed two inhibitors, an inhibitor of Gi proteins, PTX, and a selective S1P1/S1P3 antagonist, VPC23019, to initially evaluate the role of S1P receptor subtypes in 11,12-EET-induced angiogenic effects in ECs. In the presence of PTX, 11,12-EET-induced cell proliferation was substantially abolished, suggesting that the SK1 mediated proliferation in HUVECs is dependent on Gi protein coupled S1P receptors. In addition, VPC23019 also markedly attenuated 11,12-EET-induced EC proliferation. Taken together, these results suggest that Gi protein coupled S1P1/S1P3 receptors are primarily responsible for the SK1-mediated mitogenic and angiogenic effects in HUVECs. In this regard, our results are consistent with previous notions demonstrating that activation of S1P1 and S1P3 receptors enhances endothelial cell proliferation and migration, playing a key role in developmental and pathological angiogenesis.40 In addition, S1P also protects endothelial cells from apoptosis through activation of phosphatidylinositol 3-kinase/Akt/eNOS via S1P1 and S1P3 receptors.41 Nevertheless, further studies are needed to shed light on the mechanism by which 11,12-EET induces SK1 activation in ECs.
In conclusion, our results demonstrate that SK1, but not SK2, is an important mediator of the 11,12-EET-induced proliferation and angiogenesis in human ECs. Targeting SK1 for inhibition by either SK1 specific inhibitors, expression of dominant negative SK1 mutant, or small interference RNA strongly attenuates the 11,12-EET-induced endothelial cell proliferation and angiogenesis. Thus, SK1 may represent a novel therapeutic modality for treatment of angiogenesis related diseases, such as cancer and ischaemia.
This work was supported by American Heart Association Scientist Development Grant (0630047N).
Conflict of interest: None declared.
We thank Lauren Danridge for advice during the preparation of the manuscript.