From the Departments of Ophthalmology and Visual Sciences (Drs Walker, van Ginkel, Albert, and Polans; Mr Gee; and Mss Ahmadi and Subramanian), Biomolecular Chemistry (Ms Subramanian and Dr Polans), and Population Health Sciences and the State Hygiene Laboratory (Dr Meisner), University of Wisconsin Medical School, Madison; and the Schepens Eye Research Institute and the Department of Ophthalmology, Harvard Medical School, Boston, Mass (Dr Ksander).
To study the expression of angiogenic factors Cyr61 and tissue factor(TF) in uveal melanoma and its correlation with blood vessel density.
Suppression subtractive hybridization was used to identify genes that are differentially expressed between cell lines of uveal melanoma and normal uveal melanocytes. Expression of these genes was subsequently verified in primary uveal melanomas and correlated with the number of blood vessels in archival specimens by immunohistochemical analysis.
Cyr61 and TF are expressed at elevated levels in cell lines of uveal melanoma compared with normal uveal melanocytes. Duplication of a region of chromosome arm 1p, encompassing the genes encoding Cyr61 and TF, was observed in the melanoma cell line used in the initial subtractive hybridization. Both genes are also expressed in primary uveal melanomas, and a correlation was found between expression of TF and the number of blood vessels in archival specimens.
Cyr61 and TF may contribute to the angiogenic phenotype associated with uveal melanoma. A region of chromosome arm 1p also may contain oncogenes or tumor suppressor genes pertinent to the origins of this type of ocular tumor.
New immunotherapies have been devised for the treatment of cancer based on the expression of TF. Similar approaches may be effective in treating uveal melanoma.
APPROXIMATELY 1300 to 2000 new cases of uveal melanoma are diagnosed in the United States each year.1 It is the most common primary malignancy originating in the adult eye. In addition to impairment of vision, loss of life primarily because of metastatic disease remains a serious risk. Once metastatic disease is detected in the liver, for example, life expectancy usually is reduced to a few months.2,3 Some occupations involving exposure to certain chemicals or UV light may increase the risk of uveal melanoma,4,5 but no cause has been delineated for the disease.6 Comparisons between ocular melanoma and cutaneous melanoma have not yet proved informative; they differ in their systemic symptoms, metastatic patterns, and susceptibility to treatments.7
Folberg et al8 initially described a series of microcirculatory patterns associated with malignant melanomas of the uvea. A correlation was found between certain vessel patterns and patient survival from metastatic disease.9- 11 As with other types of cancer, these vessels may be leaky, resulting in breakdown of the normal ocular-blood barrier12,13 and enhanced opportunity for primary tumor cells to enter the general circulation and exit the eye. In other areas of cancer research, exciting new therapies are being implemented that target the tumor vasculature.14,15 Recently, tissue factor (TF) has been used as a target for tumor immunotherapy,16 and the topical application of angiostatic drugs has been shown to retard tumor growth in an animal model of uveal melanoma.17 To enhance the opportunities for clinical application in humans, however, more angiogenic-related targets need to be identified.
We demonstrate in this article the expression of 2 putative angiogenic growth factors, Cyr61 and TF, not previously studied in the eye. Cyr61 and TF are expressed in more malignant epithelioid cell uveal melanoma cell lines compared with either normal uveal melanocytes or less malignant spindle cell uveal melanomas. In addition to these in vitro studies, both Cyr61 and TF were detected in fresh biopsy samples and archival specimens of human uveal melanoma. Their expression may contribute to the growth and dissemination of the tumor, and their further study offers therapeutic opportunities for the treatment of this blinding and life-threatening disease.
Cyr61 is a member of an emerging family of immediate-early genes that includes transcription factors, proto-oncogenes, and cytokines.18 Cyr61 can be induced by growth factors (platelet-derived growth factor and fibroblast growth factor) or oncogenes (v-src), and, once induced in growing cell cultures, Cyr61 is secreted as a cysteine-rich, 40-kd protein that associates with the extracellular matrix.19 Cyr61 secretion increases the growth factor–induced synthesis of endothelial cell DNA and endothelial cell migration through the integrin αvβ3- and αvβ5-dependent pathways, respectively.20 Cyr61 also induces neovascularization when implanted in rat corneas, and its expression can increase the tumorigenicity of cells.19
Tissue factor is also an immediate-early gene product induced by growth factors and cytokines. Although best known for its role as the primary initiator of blood coagulation, TF is also involved in vasculogenesis, probably through intracellular signaling via its cytoplasmic domain, whereas activation of the coagulation pathway requires only the membrane anchoring of the extracellular portion of the molecule.21
Inactivation of the TF gene in knockout mice leads to embryonic lethality owing to impaired vascular development.22 These animals lack a developed periendothelial muscular wall, as seen in animals missing other angiogenic modulators, such as angiopoietin 1. This phenotype is distinct from animals deficient in the coagulation pathway,21 which die of hemostatic bleeding shortly after birth.23
Tissue factor is expressed in a variety of cancer cells that are more malignant and metastatic, and its expression correlates with the angiogenic phenotype and poorer prognosis.24- 26 Tissue factor promotes metastasis in transfected human cutaneous melanoma cells,27 and animals expressing TF have larger and more vascularized tumors.28
The finding of TF and Cyr61 in uveal melanoma is a unique observation; expression of these angiogenic factors along with vascular endothelial growth factor (VEGF) may indicate that multiple growth factors contribute to the malignant and metastatic phenotype of ocular melanoma.
Normal uveal melanocytes were obtained from human donor eyes and kept in culture according to published procedures.29 Mel 270 and Mel 290 were established from biopsy samples of human uveal melanoma following published procedures30; Mel 270 is composed primarily of spindle cells, and Mel 290 is composed primarily of epithelioid cells. Tumor cell cultures were maintained at 37°C with 5% carbon dioxide in complete RPMI-1640 medium supplemented with 10% fetal bovine serum, 0.01M HEPES, penicillin (100 U/mL), streptomycin (100 µg/mL), and 0.1% amphotericin B.
Methods essentially followed the procedures of Diatchenko et al31 using 2 µg of poly A+ RNA derived from normal uveal melanocytes and from Mel 290 to amplify differentially expressed sequences.
Cyr61 and TF transcript levels were compared in uveal melanocytes Mel 270 and Mel 290 by RNase protection assay using a 249-base antisense Cyr61 riboprobe (nt559-808 of the human Cyr61 complementary DNA [cDNA] sequence), a 254-base TF riboprobe (nt765-1019 of the human TF cDNA sequence), and a 177-base antisense glyceraldehyde-3-phosphate dehydrogenase (GAPDH) riboprobe(nt634-811 of the human GAPDH cDNA sequence) as an internal standard. Labeled riboprobes were generated with either T3 or T7 RNA polymerase, and RNase protection assays were performed as described elsewhere,32 with some modifications. [α-33P]uridine triphosphate–radiolabeled riboprobes were hybridized to 20 µg of total RNA in solution, followed by RNase digestion with RNase A (50 µg/mL) (Sigma-Aldrich Corp, St Louis, Mo) and RNase T1 (750 U/mL) (Boehringer Mannheim Biochemicals, Indianapolis, Ind). Transcript levels were compared between samples as a ratio to GAPDH.
Normal uveal melanocytes and melanoma cells were lysed in 10mM Tris buffer, pH 6.8, in the presence of protease inhibitors, sonicated, and then boiled in a solubilization buffer containing sodium dodecyl sulfate and β-mercaptoethanol. Proteins were resolved using a sodium dodecyl sulfate–15% polyacrylamide gel. (Protein derived from 100 000 cells was loaded per lane.) Nonspecific sites on blots were saturated by sequential incubation with 5% wt/vol nonfat dry milk and 5% wt/vol bovine serum albumin in TBST (10mM Tris-Cl, pH 8.0; 150mM sodium chloride; and 0.05% vol/vol Tween-20). Blots were incubated with primary antibodies (mouse anti–TF IgG [No. 4509; American Diagnostica, Greenwich, Conn], rabbit anti–Cyr61 antiserum [Munin Corp, Chicago, Ill]; and mouse anti–GAPDH IgG [Biogenesis Inc, Kingston, NH]) for 1 hour. After washing, blots were incubated for 1 hour with a horseradish peroxidase–conjugated goat anti-IgG (Jackson Laboratories, West Grove, Pa), diluted 1:10 000. Antibody binding was detected by chemiluminescence according to the manufacturer's procedures (Amersham Pharmacia Biotech, Piscataway, NJ). Glyceraldehyde-3-phosphate dehydrogenase immunostaining served as an internal control to verify uniform sample loading.
Chromosome studies were performed by exposing the proliferating cells for 30 minutes to colcemide (0.1 µg/mL final concentration). The cultures were then treated with a hypotonic solution of 0.075M potassium chloride for 20 minutes, followed by fixation in 3:1 methanol glacial acetic acid. The slides were banded by immersion in a 1% trypsin solution (Sigma-Aldrich Corp) for 25 seconds at room temperature, rinsed in tap water, air dried, and immersed in Leishman stain for 2 minutes, rinsed, and air dried. Twenty well-banded metaphase cells were examined under a microscope, photographed, and karyotyped.
Fluorescent in situ hybridization analysis, including controls, was performed essentially following published procedures33 using paraffin sections of the original Mel 290 tumor specimen and a digoxigenin-labeled 1.8-kilobase TF cDNA probe. Binding was detected using a fluorescein isothiocyanate conjugated–labeled antidigoxigenin antibody (Ventana Medical Systems, Tuscon, Ariz).
RNA was isolated from biopsy samples of uveal melanoma obtained through the oculoplastics service of the Department of Ophthalmology and Visual Sciences, University of Wisconsin Hospital. (Portions of the tumor also were processed for routine pathological examination.) The RNA was treated with DNase I and reverse transcribed with oligo(dT) primers using MMLV reverse transcription followed by polymerase chain reaction (RT-PCR) amplification using specific primers for TF (forward, 5′-TAA CCG GAA GAG TAC AGA CAG-3′; reverse, 5′-AAG TTC TCG GTC ACA GTG CA-3′), Cyr61 (forward, 5′-GTT TCC AGC CCA ACT GTA AAC ATC-3′; reverse, 5′-TTT CTC GTC AAC TCC ACC TCG GAG-3′), VEGF (forward, 5′-TCA CGA AGT GGT GAA GTT CAT GG-3′; reverse, 5′-AAG CTC ATC TCT CCT ATG TGC TGG-3′), and GAPDH (forward, 5′-TCC GGG AAA CTG TGG CGT GAT-3′; reverse, 5′-TTT CTA GAC GGC AGG TCA GGT-3′). Aliquots were removed at 15, 20, 25, and 30 cycles and subjected to electrophoresis. The relative amount of each cDNA was compared between cell lines in a linear range using GAPDH to standardize.
Recent studies demonstrate that RNA can be extracted from paraffin-embedded tissue samples for subsequent PCR amplification.34- 37 Briefly, uveal tumor tissue was excised from a paraffin section using an alternate hematoxylin-eosin–stained section for orientation. Procedures essentially followed those of Stanta and Schneider.35 Primers specific for Cyr61 (forward, same as for fresh tissue; reverse, 5′-TAA CTT TGA CCA GCC GAG GGT T-3′), TF (forward, 5′-GAA CCC AAA CCC GTC AAT CAA-3′; reverse, 5′-GGT GAG GTC ACA CTC TGT GT-3′), VEGF (same as for fresh tissue), and GAPDH (forward, 5′-ACC ATG GGG AAG GTG AAG G-3′; reverse, 5′-CAT TGA TGG CAA CAA TAT CCA C-3′) were chosen that spanned a region of each gene separated by an intron.
Sections (5 µm) of uveal melanoma were deparaffinized with xylene, rehydrated in a graded series of ethanol, and endogenous peroxidase activity reduced by washing with 0.3% methanolic hydrogen peroxide. In some experiments, duplicate slides were incubated overnight with primary antibody, 1 µg/mL(either mouse anti-TF [American Diagnostica] or mouse anti-CD34 [NovoCastra, Newcastle, England]) diluted with phosphate-buffered saline containing 1% wt/vol bovine serum albumin and 5% vol/vol normal goat serum. Sections then were incubated sequentially with a biotinyl anti-IgG, 1 µg/mL (Vector Laboratories, Burlingame, Calif), and streptavidin-HRP complex, 1 µg/mL(Amersham Pharmacia Biotech). Typically, Vector VIP (Vector Laboratories) was used as the substrate to optimize differences between the reaction product(blue/purple) and background melanin. Tissue factor immunostaining was scored as high or low by 2 observers (R.L.G., H.A.), and the number of blood vessels in alternate sections was counted in 3 separate fields (original magnification ×20) per duplicate tumor section. The evaluation was masked, whereby specimens were coded separately for TF immunostaining and blood vessel counting. Duplicate sections from 9 tumor specimens were evaluated in this manner. The protocol for accessing tumor specimens was approved by the institutional review board at the University of Wisconsin.
To identify differentially expressed genes that may contribute to the malignant and metastatic phenotypes associated with human uveal melanoma, the method of suppression subtractive hybridization was initially used to compare messenger RNA patterns between normal uveal melanocytes and Mel 290, an established epithelioid cell line derived from a biopsy specimen of human uveal melanoma. Two genes identified by this approach, Cyr61 and TF, are known to be expressed in a variety of cancer cells, and their expression correlates with the angiogenic phenotype and poorer prognosis.19,24,25
The differential expression of Cyr61 and TF was confirmed initially by Western blot analysis (Figure 1). An immunoreactive band corresponding to TF was detected at the appropriate relative molecular mass (47 kd) in Mel 290 but not in lysates of normal uveal melanocytes (UM3 and UM4) or the spindle cell line (Mel 270). Similar results were obtained with antibodies to Cyr61 (relative molecular mass, 41 kd [Figure 1]). GAPDH immunostaining served as an internal control to verify uniform sample loading in the same series of immunoblots.
Lysates of normal uveal melanocytes(UM3 and UM4), Mel 270 spindle cells, and Mel 290 epithelioid cells were transferred to Immobilon (Millipore Corp, Bedford, Mass), a polyvinylidene fluoride transfer membrane for proteins, and stained with antibodies for tissue factor (TF) and Cyr61. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) immunostaining served as an internal control to verify uniform sample loading.
To determine whether the differential expression of Cyr61 and TF could be explained by a difference in the relative amounts of transcript, an RNase protection assay was performed using RNA obtained from normal uveal melanocytes, Mel 270 spindle cells, and Mel 290 epithelioid cells (Figure 2). The Cyr61 and TF transcripts were not detected in the normal uveal melanocytes and were only minimally detected in the spindle cells of Mel 270. In contrast, both Cyr61 and TF were detected in the Mel 290 cell line, thereby supporting the outcome of the immunoblotting analysis and the suppression subtractive hybridization experiment, which selected for elevated gene expression in Mel 290.
Transcript levels for Cyr61 and tissue factor (TF) were measured by RNase protection assay in normal uveal melanocytes (UMs), Mel 270 spindle cells, and Mel 290 epithelioid cells. Transcript levels were compared between samples as a ratio to glyceraldehyde-3-phosphate dehydrogenase (GAPDH).
The cytogenetic analysis of Mel 290 revealed an altered female karyotype(Figure 3): 46, XX, del(1)(p22.3p34.3), der(8)t(1;8)(p13;q24.3). Mel 290 was initially thought to display a terminal deletion of chromosome 1 at p32 and the translocation of the entire short arm of chromosome 1 to the terminal long arm of chromosome 8. However, use of fluorescent in situ hybridization with a probe for 1p36 demonstrated that the terminal region was present and, therefore, that the chromosome 1 deletion was interstitial, involving loss of 1p22.3→p34.3. Therefore, the translocation of the entire chromosome 1 short arm to chromosome 8 resulted in duplication of 1p13.1→1p22.3 and 1p35→pter. Other than this region, the karyotype was balanced, all of the cells demonstrated the same karyotype, and the cell line remained relatively stable.
Cytogenetic analysis of Mel 290. Note the duplication of a portion of chromosome arm 1p and its translocation to 8q (arrows). The inset shows a section of the original Mel 290 specimen demonstrating, by fluorescent in situ hybridization, 3 copies of the tissue factor gene in a population of tumor cells.
Tissue factor maps to band 1p21-1p22, and Cyr61 similarly maps to 1p22.3.38 Duplication of these genes due to a chromosomal imbalance, therefore, may explain the elevated levels of Cyr61 and TF measured in the epithelioid cells of Mel 290. Alternatively, genetic events not directly arising in this region of chromosome 1 may nonetheless induce expression of Cyr61 and TF. These structural changes in the Mel 290 cell line reflect alterations in the original patient tumor from which the cell line was derived. Fluorescent in situ hybridization analysis demonstrated 3 copies of the TF gene in a population of tumor cells (Figure 3, inset), confirming the duplication of this region of chromosome arm 1p in the original patient specimen. Some cells in the tumor specimen, however, had fewer TF fluorescent signals, indicating either heterogeneity in the tumor or, more likely, that chromosomes in these cells were confined to a different plane of section.
The independent cytogenetic findings from Mel 290 are consistent with the outcome of the suppression subtractive hybridization experiment and the data obtained from the RNase protection assays, RT-PCR, and Western blot analysis that expression of Cyr61 and TF is enhanced in Mel 290.
To ensure that expression of Cyr61 and TF is not the result of in vitro culture conditions but rather reflects actual occurrences in primary tumor tissues, fresh biopsy samples and archival specimens of uveal melanoma were analyzed for expression of Cyr61 and TF. Figure 4A illustrates the results from RT-PCR experiments using 9 different archival specimens of uveal melanoma that had been fixed in formalin and embedded in paraffin. Cyr61 and TF were detected in all 9 samples; thus, expression of these 2 factors occurs in primary tumors and is not simply the result of culture conditions.
A, RNA was extracted from paraffin sections of human uveal melanoma and used for reverse transcription–polymerase chain reaction (RT-PCR) to detect transcripts corresponding to tissue factor(TF), Cyr61, vascular endothelial growth factor (VEGF), and glyceraldehyde-3-phosphate dehydrogenase (GAPDH). Routine pathological examination of parallel sections confirmed that samples 1 through 4 consisted primarily of spindle cells, whereas samples 5 through 9 contained primarily epithelioid cells. B, RT-PCR was performed using RNA extracted from a uveal melanoma within 30 minutes of enucleation.
Expression of VEGF was measured in each specimen as well, but the detection was less pronounced. In addition to archival specimens, expression of each of these angiogenic factors was demonstrated by RT-PCR using RNA obtained from fresh biopsy samples of uveal melanoma (Figure 4B). Glyceraldehyde-3-phosphate dehydrogenase amplification from each sample served as an internal control to verify uniform sample loading.
In further immunohistochemical experiments, duplicate sections from 9 tumor specimens were stained with antibodies specific for TF, and alternate sections from the same specimens were stained with CD34 antibodies to help delineate blood vessels. First, the immunostaining associated with TF was uniformly associated with tumor cells and not with a subpopulation of other cell types (Figure 5A, purple stain). Next, TF immunostaining was rated as high or low by 2 independent observers(R.L.G., H.A.), the number of blood vessels was counted, and the 2 variables were compared (Figure 5). Five specimens were rated high for TF immunoreactivity and had a mean ± SD blood vessel count of 35.0 ± 16.8, and 4 specimens with low TF immunoreactivity had a mean ± SD blood vessel count of 1.5 ± 1.4. These data demonstrate a correlation between TF and the number of blood vessels. A larger sampling of specimens is now being studied to corroborate these initial findings, and similar studies with Cyr61 are also being conducted.
Alternate sections of human uveal melanoma were stained with antibodies specific for tissue factor (TF) (A and C) or CD34 to delineate blood vessels (B and D). A and B illustrate sections stained with TF and scored as high and containing numerous blood vessels (enhanced purple reaction product denoted in some cases with arrows), whereas C and D illustrate low-scoring TF and fewer blood vessels.
Like most solid tumors, uveal melanomas may require a new blood supply to grow beyond 1 to 2 mm. Because the eye lacks lymphatics, permeable blood vessels also form the primary route by which tumor cells enter the circulation, escape the eye, and disseminate to distant regions of the body. Most research has focused on VEGF as the principal angiogenic factor contributing to tumor neovascularization in the eye. More extensive investigations have been conducted on the role of VEGF in blinding diseases with associated neovascularization, as in diabetic retinopathy, retinopathy of prematurity, and age-related macular degeneration (for a review, see Aiello39).However, it is unclear whether uveal melanoma tumor cells express VEGF in a manner that can fully explain the proliferation and morphogenesis of vascular endothelial cells requisite for new blood vessel formation extending throughout the tumor mass.40,41 Instead, these tumor cells may express multiple angiogenic factors that independently or together with VEGF regulate new vessel formation. In addition to conventional angiogenesis, these same factors may support vascular mimicry, an alternative source of tumor circulation.42
In this article we demonstrate expression of Cyr61 and TF, 2 angiogenic factors that contribute to tumor growth and patient prognosis in other types of cancer. These factors have not been studied extensively in the eye. Neither factor is expressed in normal uveal melanocytes, but several independent measurements reported herein demonstrate their simultaneous expression in the more malignant epithelioid cell line used in the initial subtractive hybridization experiments that revealed their identity. In addition, genes encoding these 2 factors map to a region of chromosome arm 1p that is duplicated in the same epithelioid cell line, possibly explaining their elevated expression. This region of 1p may also harbor genes related to the initial events of transformation, since no other structural mutations thus far have been associated with Mel 290. Newer techniques, such as comparative genomic hybridization, however, have further resolving power and may detect additional mutations. Of further significance, Cyr61 and TF can be detected in fresh biopsy samples and archival specimens of uveal melanoma by molecular methods and by immunohistochemical techniques, reducing concerns about artifacts from culture conditions. Although RNase protection assays demonstrated higher levels of expression of Cyr61 and TF in epithelioid cell lines than did spindle cells, more extensive amplification by RT-PCR revealed expression of both factors in archival specimens of spindle cell and epithelioid cell uveal melanomas. More quantitative methods need to be applied to archival specimens to determine whether there is a precise correlation between the level of these specific transcripts and the tumor cell type.
Interest recently has been raised in the selective expression of TF in tumor blood vessels and not mature vessels, since TF can be used as a target for immunotherapy.16 A similar approach could be applied to the treatment of uveal melanoma, since we demonstrated its expression in tumor cells but not in melanocytes. Immunotherapies directed toward the tumor circulation also could augment other types of treatment.
Ultimately, to determine whether TF, Cyr61, and VEGF contribute to new vessel formation, it will be necessary to alter the expression of each of these factors alone or in combination with one another in cell lines of uveal melanoma and then to evaluate vascular density and tumor growth in animals inoculated with these transfected cell lines. Elucidation of the pathways involving Cyr61, TF, and other angiogenic factors will provide new opportunities for intervening in the growth and progression of uveal melanoma.
Submitted for publication April 4, 2002; final revision received July 15, 2002; accepted August 6, 2002.
This research was funded by grants EY12768 (Dr Polans) and EY09294 (Dr Albert) from the National Institutes of Health, Bethesda, Md; a grant from the University of Wisconsin Comprehensive Cancer Center; an unrestricted grant from Research to Prevent Blindness Inc, New York, NY; and a grant from the Retina Research Foundation, Houston, Tex. Dr Polans is a Research to Prevent Blindness Inc Jules and Doris Stein Professor.
The authors have no affiliations with or financial interest in any organization or entity that conflicts with or has financial interest in the subject matter or materials discussed in the article.
We thank Hans Prydz, MD, University of Oslo, Oslo, Norway, for the gift of the 1.8-kilobase tissue factor complementary DNA clone.
Corresponding author and reprints: Arthur S. Polans, PhD, Department of Ophthalmology and Visual Sciences, University of Wisconsin Medical School, Room K6/466 Clinical Sciences Center, Madison, WI 53792 (e-mail: firstname.lastname@example.org).
Thank you for submitting a comment on this article. It will be reviewed by JAMA Ophthalmology editors. You will be notified when your comment has been published. Comments should not exceed 500 words of text and 10 references.
Do not submit personal medical questions or information that could identify a specific patient, questions about a particular case, or general inquiries to an author. Only content that has not been published, posted, or submitted elsewhere should be submitted. By submitting this Comment, you and any coauthors transfer copyright to the journal if your Comment is posted.
* = Required Field
Disclosure of Any Conflicts of Interest*
Indicate all relevant conflicts of interest of each author below, including all relevant financial interests, activities, and relationships within the past 3 years including, but not limited to, employment, affiliation, grants or funding, consultancies, honoraria or payment, speakers’ bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued. If all authors have none, check "No potential conflicts or relevant financial interests" in the box below. Please also indicate any funding received in support of this work. The information will be posted with your response.
Some tools below are only available to our subscribers or users with an online account.
Download citation file:
Web of Science® Times Cited: 17
Customize your page view by dragging & repositioning the boxes below.
Enter your username and email address. We'll send you a link to reset your password.
Enter your username and email address. We'll send instructions on how to reset your password to the email address we have on record.
Athens and Shibboleth are access management services that provide single sign-on to protected resources. They replace the multiple user names and passwords necessary to access subscription-based content with a single user name and password that can be entered once per session. It operates independently of a user's location or IP address. If your institution uses Athens or Shibboleth authentication, please contact your site administrator to receive your user name and password.