|Year : 2020 | Volume
| Issue : 68 | Page : 8-12
A unified method for different placental products species identification
Tien-Lin Chang1, Shang-Fang Wu2, Der-Yuan Wang3, Chih-Hung Huang4
1 Department of Chemical Engineering and Biotechnology, Ph.D. program, Institute of Biochemical and Biomedical Engineering, National Taipei University of Technology, Taipei, Taiwan
2 Department of Chemical Engineering and Biotechnology, Master program of Biochemical and Biomedical Engineering, Institute of Biochemical and Biomedical Engineering, National Taipei University of Technology, Taipei, Taiwan
3 Department of Chemical Engineering and Biotechnology, Institute of Biochemical and Biomedical Engineering, National Taipei University of Technology; Division of Research and Analysis, Taiwan Food and Drug Administration, Taipei, Taiwan
4 Department of Chemical Engineering and Biotechnology, Institute of Biochemical and Biomedical Engineering, National Taipei University of Technology, Taipei, Taiwan
|Date of Submission||23-Jul-2019|
|Date of Decision||11-Oct-2019|
|Date of Web Publication||31-Mar-2020|
Department of Chemical Engineering and Biotechnology, Institute of Biochemical and Biomedical Engineering, National Taipei University of Technology, 1, Section 3, Zhongxiao E. Rd., Taipei 10608
Source of Support: None, Conflict of Interest: None
| Abstract|| |
Background: Placentas are widely used in the production of cosmetics, Chinese medicines, and injections. As the esthetic medicine market grows, challenges such as impure preparations and an increasing number of counterfeit products develop, owing to the lack of standard inspection methods and accurate description of the composition of these products. Objectives: To develop a universal species identification method for placental products. Materials and Methods: Fresh swine, cattle, goat, and frozen human placenta were used as references, while Placenta Hominis and placental extract products from the market as trial samples. Samples were prepared by kit-free DNA isolation, and vertebrate-specific primer sets of mitochondrial 16S and 12S rRNA were used for species-specific region amplification. Direct sequence of amplicons and National Center for Biotechnology Information database comparison was carried for species identification of the testing samples. Results: Trail tests had confirmed the usability of this method. Six of seven Placenta Hominis samples were shown to contain human DNA traces, while the seventh showed no DNA signals belonging to any mammals and a paper-like material underneath the vegetal part, instead of membrane-like placental tissue by visual inspection, suggesting the possibility of counterfeit. Eight injectable and cosmetic placental extract products were tested, and none of them contained analyzable DNA or comparable protein fragments except one, which DNA was from rainbow trout (Oncorhynchus mykiss) rather than sheep, as per the product claim, and was hence misbranded as per the U.S. Food and Drug Administration definition. Conclusion: Our species identification method is easy-to-operate, unified, and resource-saving, which can be applied to different placental products.
Keywords: Mitochondrial DNA, Placenta Hominis, placental extract, species identification, vertebrate-specific primers
|How to cite this article:|
Chang TL, Wu SF, Wang DY, Huang CH. A unified method for different placental products species identification. Phcog Mag 2020;16, Suppl S1:8-12
|How to cite this URL:|
Chang TL, Wu SF, Wang DY, Huang CH. A unified method for different placental products species identification. Phcog Mag [serial online] 2020 [cited 2020 Jul 5];16, Suppl S1:8-12. Available from: http://www.phcog.com/text.asp?2020/16/68/8/281695
In this study, we reported the following achievements:
- Kit-free placental material and placenta extract DNA isolation and a unified polymerase chain reaction condition with minimal laboratory resources requirement
- Great impurity resistance by vertebrate mitochondrial DNA identification in Placenta Hominis
- Species identification capability in injectable placental extract products containing DNA fragments.
Abbreviations used: BLAST: Basic Local Alignment Search Tool; Bp: Base pair; COI: Cytochrome c oxidase subunit I; DNA: Deoxyribonucleic acid; EDTA: Ethylene diamine tetraacetic acid; FDA: U.S. Food and Drug Administration; G: Gram; G: Gravitational force unit; H: Hour; Kb: Kilo base pair; M: Molar; Mg: Milligram; Min: Minute; Ml: Milliliter; mtDNA: mitochondrial DNA; NCBI: National Center for Biotechnology Information; rRNA: Ribosomal ribonucleic acid; PCR: Polymerase chain reaction; S: Second; Seq: Sequence; SDS: Sodium dodecyl sulfate; TE: Tris-EDTA; μl: Microliter; %: Percent.
| Introduction|| |
Placenta is a temporary supportive organ that connects a developing fetus to the maternal uterine wall and is involved in nutrient uptake, waste elimination, and gas exchange through the mother's bloodstream. After birth, the placenta is expelled from the mother's body. Considerable amounts of nutrients and hormones are retained in the placenta during pregnancy. A wide range of mammals eat their placenta after childbirth that is a phenomenon known as placentophagy., Placenta has been used in medicine throughout human history. Placenta Hominis is a traditional Chinese herbal material made using human placenta. The production method involves deep-fried placenta mixed with other herbs, corn, or bean powders and baked at low heat until dry. The placental tissue shrinks into a membrane-like structure covered with vegetable powders. Placenta Hominis has been used both as a lactagogue , as an energy supplement and for treating night sweats, impotence, and infertility. In other countries, placentophagy also exists as a folk remedy., However, due to limited supply and lack of regulation, extensive use of placental sources from other species prompts the need for better regulation and product identification.
In modern studies, placental extract products have shown beneficial effects as antioxidants,, antidepressants, in wound healing, in the treatment of rheumatoid arthritis, in osteoarthritis  and arthritis, as pain relievers, biologic fillers, anti-inflammatory substances, and hair growth promoters, among other effects. A wide range of nucleic acids, hormones, and proteins within placenta play important roles in these functions.,, Placental extract products have also been widely used in cosmetic products since the early twentieth century., Placental extract is produced by homogenization, protein hydrolysis, or supercritical fluid extraction  followed by further filtration and purification steps. Different extraction methods yield different compositions of the final products; however, all these products are sold under the same name, “placental extract,” confusing the public and posing challenges for the administration. According to the U.S. Food and Drug Administration (FDA) cosmetic handbook, products should not be identified by the name “placental extract” but by a name which describes their composition more accurately. Many of the current products with placental extract contain both purified and unpurified components. Due to limited knowledge of the composition of placental extract and lack of standard inspection protocols, several challenges, such as impurities in placental ingredients and production of counterfeits, exist, leading to limited effectiveness and causing irritation or allergies.
Earlier studies on Placenta Hominis identification were focused on cytochrome c oxidase subunit I (COI) region by barcode  or species-specific primer sets. Without specific primer sets separating animal and vegetal species, these approaches require either more delicate sample preparation or separate amplification for every target species, to reduce signal noise due to contamination.
In this study, we developed a kit-free DNA isolation method adopted for raw placenta and placental products followed by a unified polymerase chain reaction (PCR) condition using vertebrate-specific primer sets to amplify part of mitochondrial 12S and 16S rRNA for species identification. This species identification method has high accuracy, is time-consuming and cost-saving, and is a unified procedure. It has the potential to identify components from raw animal material through to end products, thereby providing a powerful tool for setting certification test standards, commodity inspection, and customs examination of animal materials that are imported or exported.
| Materials and Methods|| |
All fresh placentas except human placenta were obtained from a local farm and slaughterhouse. Human placental samples were purchased from ILSbio (sample number 1-410-810-7506). Placenta Hominis samples were purchased from local Chinese medicine shops. Placental extract product samples (seven injectable and one cosmetic) were purchased directly from the suppliers.
Reagents, bacterial strain, and plasmids
All chemicals were from Merck except agarose from AMRESCO, antibiotics from Sigma, and enzymes from Roche, New England Biolabs, and TAKARA. DNA extraction and PCR clean-up kits were from Viogene. Cloning vector pBluescript II KS(-) from Stratagene was used for PCR product cloning. Escherichia coli strain DH5α was used, and standard culture conditions for molecular cloning were followed.
Sample DNA isolation
Total DNA of the raw placenta and Placenta Hominis samples were isolated using the following protocol: 0.1 g of tissue was taken from the test sample. It is important that samples are cut vertically if using Placenta Hominis so that they contain both vegetal and animal parts. Samples were washed and soaked till softened in 10 times volume of the Tris-ethylene diamine tetra acetic acid (TE) buffer containing 0.1% sodium dodecyl sulfate (SDS), crushed into small pieces, and rinsed with TE buffer containing 0.1% SDS at 37°C for 5 min. Samples were vortexed and the supernatant was discarded. The rinsing procedure was repeated four times. The sample was treated with 0.4 mg proteinase E, in 700 ml TE buffer containing 0.1% SDS at 50°C for 8 h. After addition of another 0.4 mg of proteinase E, samples were incubated for another 4 h. For the removal of proteins and digested peptides, 100–200 μl of phenol/chloroform was added and mixed completely, the sample centrifuged at 15,000 g for 10 min, and the upper layer collected into a new Eppendorf tube. This step was repeated until the white interlayer disappeared. DNA was precipitated by adding 0.1 volume of 3 M sodium acetate and 0.65 volume isopropanol. The sample was mixed by inverting the tube 2–3 times and then allowing to stand for 10 min at room temperature (20°C–30°C). The sample was spun at 15,000 g for 10 min, the supernatant was discarded, and the pellet was washed with 70% ethanol (−20°C), air-dried, and then resuspended in sterile water.
For placental extract products, 300 μl of the sample was taken and followed the same protocol starting from phenol/chloroform treatment to the end.
Primers were synthesized by a local biotech company (Genomics). Primer sequences were identical to those in a published study  and are listed in [Table 1].
|Table 1: Universal vertebrate primer sequences of 12S and 16S mitochondrial deoxyribonucleic acid|
Click here to view
Polymerase chain reaction amplification
PCR amplification was carried out using high-fidelity thermostable DNA polymerase (TAKARA) with the following protocol: 98°C initialization for 30 s, 35 amplification cycles of 98°C denaturation for 10 s, 57°C annealing for 15 s, 72°C elongation for 30 s, and 72°C final elongation for 5 min and then held at 4°C. Positive control containing cloned placental 12S and 16S partial region as templates, while negative controls replacing DNA template with ddH2O were included in each test batch to eliminate contamination due to human handling. Amplified PCR products were checked by gel electrophoresis.
Cloning and sequencing
After PCR amplification, PCR products were sent for sequencing. For positive control preparation, PCR products of partial 12S and 16S regions from placental tissue of reference animal species were cloned into pBluescript II KS(-) vector and transformed into DH5α competent cells. IPTG/X-gal selection single colonies were picked from Luria-Bertani (LB) agar plates with ampicillin and amplified in LB broth. Amplified PCR product-containing plasmids were sequenced, ensuring correct preparation of the positive controls.
| Results|| |
Establishment of a method for accurate identification of distinct animal raw placentas
The DNA isolation and unified PCR protocol were tested with fresh placentas from cattle, goat, swine, and frozen human tissue as the standard materials. Three experimental repetitions were done for each species, and the isolated DNA was checked by gel electrophoresis [Figure 1]a. Clear trace of isolated DNA was observed from all samples. PCR was done with two sets of vertebrate-specific mitochondrial DNA (mtDNA) primers, resulted in a clear 200–300 bp amplified band in all samples [Figure 1]b. Each amplicon was sequenced directly, BLAST with National Center for Biotechnology Information database and confirmed to be 100% identical to its related species. This results showing that our unified procedure works with the raw placentas, providing a trustworthy base before we examing placental products.
|Figure 1: DNA isolation of fresh placenta and polymerase chain reaction amplification, (a) DNA extraction of fresh placenta. (b) Polymerase chain reaction amplification results of mitochondrial DNA 12S and 16S fragments. All samples were running in 0.8% agarose gels except cattle sample in (a) was running in 1.2% agarose gel. At column width. M: 100 bp ladder marker; M1: 1 kb ladder marker; H: Human; C: Cattle; G: Goat; S: Swine|
Click here to view
Applicability of this method to Placenta Hominis
Seven samples of Placenta Hominis were collected from different local Chinese medicine shops. We successfully isolated DNA from all the samples; six out of seven DNA samples were amplified by PCR and sequenced. All of the sequenced samples showed the highest similarity to the human reference sequence. BLAST results of the six samples showed 99.59%–100% identity to human mtDNA with the expect value between 2e −121 and 4e −123 for the 16S mtDNA and 99.53%–100% identity with the expect value between 2e −105 and 5e −107 for the 12S mtDNA. We inspected the particular sample with no PCR amplification; it is very similar to other Placenta Hominis other than a paper-like material underneath the vegetal part with cellulose-like fibers at the edges, instead of membrane-like placental tissue  [Figure 2]. This paper-like membrane layer was not found in other Placenta Hominis samples. We sent this sample for proteomic analysis by Proteomics Research Center of National Yang-Ming University. The peptide signals failed to relate to any particular mammal (data not shown). These findings suggest that this sample might not contain the critical component of Placenta Hominis. These results demonstrated that our method can also be used to examine Placenta Hominis, even in the presence of vegetal materials.
|Figure 2: Morphology comparison of normal Placenta Hominis (left) and the one with no DNA trace and amplicon by vertebrate-specific primers (right). A paper-like membrane layer (red arrow) is underneath the yellowish vegetal coat. At column width|
Click here to view
Placental extract containing nucleotide fragments can be identified by this method
Many injectable placental extract products in the market are used for antiaging and wound healing in clinics. Placental extract has also been introduced into cosmetics products based on the antiaging function. Earlier studies showed that polydeoxyribonucleotide (PDRN) is capable of stimulating vascular endothelial growth factor to trigger wound healing ,, and skin regeneration. PDRN is a mixture of deoxyribonucleotide polymers with chain lengths ranging between 50 and 2000 bp and is very abundant in the placental extract. Generally, placentas are available from livestock farms as by-products. However, the current process of producing placental extract means that proteins and nucleotides are hydrolyzed into small fragments, which makes it difficult to use them for species identification. Few products claim to contain DNA, but we were interested to see whether our method could be applied in this field. Eight injectable and cosmetics placental extract product samples were collected from different manufacturers made by different animal sources. Results showed that of the eight samples, the only sample that produced a DNA signal and generated a PCR-amplified product was an injectable placental extract product claimed to contain sheep extract. We could detect neither clear DNA signals by electrophoresis nor the PCR-amplified products from the rest of the samples. However, sequence results from the only sample containing DNA were unexpected. BLAST result showed 99% identity to salmon trout (Oncorhynchus mykiss) mtDNA with only a single base pair mismatch (16S: Seq ID-MF621750.1, identity: 99.59%, expect value: 2e −122; 12S: Seq ID-KP013084.1, identity: 97.64%, expect value: 5e −107) while the product claimed to be made from sheep placenta. Although sperm from salmon and trout is a good source of PDRN,, this should not be labeled as a placental extract and is a clear example of misbranded cosmetics, according to the FDA definition. With the unexpected result, we sent all these samples for proteomic analysis. However, due to protein hydrolysis and removal of larger residues during placental extract production, the peptide fragments from the samples were unable to make trustworthy species comparison (data not shown).
| Discussion|| |
This method produced high copy numbers of mtDNA segments by PCR, increasing the rate of successful species identification. The mtDNA 12S and 16S sequences are composed of highly conserved regions and species-specific regions across different species, making it easier to separate different species using short sequences. The similarity between human and swine, cattle, and goat 16S fragments is 88%, 87%, and 87% while the similarity between 12S fragments of the same species above is 86%, 88%, and 85%, respectively [Figure 3]. Although being a powerful tool for protein identification, proteomics by mass spectrometry-driven sequence similarity searches  has its limitations. Especially with placental extract products, random digestion by hydrolysis, over-fragmentation during production treatments, and removal of larger peptide fragments increase the difficulty of applying this analysis. Compared with the use of protein identification, mtDNA identification is faster and cheaper and can be carried out by any molecular biology laboratory with standard equipment. This protocol does not need any commercial kit to isolate DNA from samples, and the DNA is amplified by the unified PCR condition with vertebrate universal primer sets. These processes require fewer laboratory resources, and DNA sequencing can be finished by self or sequencing companies. Compared with the existing COI-based sequence method, this method benefits from the following aspects: first, vertebrate universal primers are inactive with vegetal components or parasites, making sampling easier; second, a single tube reaction instead of multiple tubes using different species-specific primers, reducing the chance of contamination or handling error; and third, a kit-free sample preparation requires fewer resources. This method depends heavily on sequencing of the PCR products and is very sensitive to sample impurity/contamination because of specific amplification regions across species that could not be identified by size. Thus, positive/negative controls were introduced to prevent contamination. Theoretically, if there were multiple placental sources from different species in one sample, it can be distinguished from reading the sequencing peaks overlapped at species variation regions by directly sequencing the amplicons. In our study, no cross-species mixture in testing samples had been detected. This work is the first report identifying the vertebrate source of injectable placental extract products.
|Figure 3: Alignment of the polymerase chain reaction target sequences of 12S and 16S of four species. Using the human sequence as the reference, the identical and gap bases of other species are abbreviated by dot and dash, respectively. At full page width|
Click here to view
| Conclusion|| |
With the increase of placental products being adopted by the global market, it is essential to have better regulation rules and methods towards raw material inspection, effective ingredients analysis, detection of counterfeits and many more. In this work, we described a species identification method of placental products composed of sample preparation, mtDNA amplification by a unified PCR condition, and DNA sequencing. This method had been tested on the fresh placenta and Placenta Hominis, and the results were convincing and reproducible. For injectable and cosmetic products containing placental extract, only part of the samples containing nucleotides above a certain length can be detected with this method and one misbranded product was detected in this work. It is essential to identify the species of placental products because the limited legal source of human placenta is the main reason for counterfeit Placenta Hominis. Our results also showed the importance of precise regulation and standardized inspection methods for placenta-related products with massive market adoption by placental extract products, as a matter of public interest. Benefit from the nature of mtDNA, species identification can be carried with DNA fragments contained in products after different treatments through the production process while tissue/organ-specific identification could not be resolved by DNA-based inspection. While this testing procedure can improve safety for the end consumers by preventing counterfeits from entering the market and providing better quality control for retail products, this method still needs to pair with other inspection tools for better identification when DNA trace was removed from many placental extract products.
The authors wish to thank Dr. Yeou-Guang Tsay and the Proteomics Research Center of National Yang-Ming University to complete this work.
Financial support and sponsorship
This study was funded by the Research Support Scheme of the Taiwan Food and Drug Administration, grant no. 102TFDA-A-521E.
Conflicts of interest
There are no conflicts of interest.
| References|| |
Phuapradit W, Chanrachakul B, Thuvasethakul P, Leelaphiwat S, Sassanarakkit S, Chanworachaikul S. Nutrients and hormones in heat-dried human placenta. J Med Assoc Thai 2000;83:690-4.
Young SM, Benyshek DC. In search of human placentophagy: A cross-cultural survey of human placenta consumption, disposal practices, and cultural beliefs. Ecol Food Nutr 2010;49:467-84.
Ober WB. Notes on placentophagy. Bull N
Y Acad Med 1979;55:591-9.
Ou Yang CG. The study on quality control of placenta hominis and its capsule. Chin J Inf Trad Chin Med 2004;10:888-9.
Keldenich H, Bizzi G, Longhi O. Treatment of hypogalactia with a placental extract. Clinical study. Minerva Ginecol 1976;28:1015-20.
Choi HY, Kim SW, Kim B, Lee HN, Kim SJ, Song M, et al
. Alpha-fetoprotein, identified as a novel marker for the antioxidant effect of placental extract, exhibits synergistic antioxidant activity in the presence of estradiol. PLoS One 2014;9:e99421.
Togashi S, Takahashi N, Iwama M, Watanabe S, Tamagawa K, Fukui T. Antioxidative collagen-derived peptides in human-placenta extract. Placenta 2002;23:497-502.
Apari P, Rózsa L. Deal in the womb: Fetal opiates, parent-offspring conflict, and the future of midwifery. Med Hypotheses 2006;67:1189-94.
Hong JW, Lee WJ, Hahn SB, Kim BJ, Lew DH. The effect of human placenta extract in a wound healing model. Ann Plast Surg 2010;65:96-100.
Rosenthal M. The application of an extract of human placenta in the treatment of rheumatic affections. Int J Tissue React 1982;4:147-51.
Kim JK, Kim TH, Park SW, Kim HY, Kim Sh, Lee Sy, et al
. Protective effects of human placenta extract on cartilage degradation in experimental osteoarthritis. Biol Pharm Bull 2010;33:1004-10.
Yeom MJ, Lee HC, Kim GH, Shim I, Lee HJ, Hahm DH. Therapeutic effects of hominis placenta injection into an acupuncture point on the inflammatory responses in subchondral bone region of adjuvant-induced polyarthritic rat. Biol Pharm Bull 2003;26:1472-7.
Gurgel LA, Santos FA, Rao VS. Effects of human placental extract on chemical and thermal nociception in mice. Eur J Pain 2000;4:403-8.
Oh EJ, Kim TK, Shin JH, Choi JH, Chung HY. Biologic filler using human fibroblasts and placenta extracts. J Craniofac Surg 2011;22:1557-60.
Banerjee KK, Bishayee A, Chatterjee M. Anti-inflammatory effect of human placental extract: A biochemical mechanistic approach. Riv Eur Sci Med Farmacol 1992;14:361-6.
Zhang D, Lijuan G, Jingjie L, Zheng L, Wang C, Wang Z, et al
. Cow placenta extract promotes murine hair growth through enhancing the insulin-like growth factor-1. Indian J Dermatol 2011;56:14-8.
] [Full text]
Tonello G, Daglio M, Zaccarelli N, Sottofattori E, Mazzei M, Balbi A. Characterization and quantitation of the active polynucleotide fraction (PDRN) from human placenta, a tissue repair stimulating agent. J Pharm Biomed Anal 1996;14:1555-60.
Lubowe II. Topical use of placenta-extract gel (non-estrogenic) in the treatment of aging skin. J Am Geriatr Soc 1963;11:914-7.
Ito K, Yamada R, Matsumoto N, Imamura T. Evaluation of fibroblast growth factor activity exerted by placental extract used as a cosmetic ingredient. J Cosmet Dermatol 2018;17:821-9.
Liu G, Zheng Y, Xu X, Liu X, Yuan F, Gao Y. Removal of lipid and cholesterol from goat placenta by supercritical carbon dioxide extraction. J Food Process Eng 2011;34:657-70.
Food and Drug Administration. FDA's Cosmetic Handbook. Washington, D.C.: U.S. Government Printing Office; 1993.
Chen J, Jia J, Xu XL, Xin TY, Zhang HY, Shi LC, et al
. Identification of Placenta hominis and its adulterants using COI barcode. Zhongguo Zhong Yao Za Zhi 2014;39:2204-7.
Lo YT, Yik MH, Shaw PC. Effective authentication of Placenta Hominis. Chin Med 2018;13:32.
Kitano T, Umetsu K, Tian W, Osawa M. Two universal primer sets for species identification among vertebrates. Int J Legal Med 2007;121:423-7.
Sambrook J. Molecular Cloning: A Laboratory Manual. 3rd
ed. New York: Cold Spring Harbor Laboratory Press; 2001. p. 1.110.
Galeano M, Bitto A, Altavilla D, Minutoli L, Polito F, Calò M, et al
. Polydeoxyribonucleotide stimulates angiogenesis and wound healing in the genetically diabetic mouse. Wound Repair Regen 2008;16:208-17.
Valdatta L, Thione A, Mortarino C, Buoro M, Tuinder S. Evaluation of the efficacy of polydeoxyribonucleotides in the healing process of autologous skin graft donor sites: A pilot study. Curr Med Res Opin 2004;20:403-8.
Rubegni P, De Aloe G, Mazzatenta C, Cattarini L, Fimiani M. Clinical evaluation of the trophic effect of polydeoxyribonucleotide (PDRN) in patients undergoing skin explants. A Pilot Study. Curr Med Res Opin 2001;17:128-31.
Veronesi F, Dallari D, Sabbioni G, Carubbi C, Martini L, Fini M. Polydeoxyribonucleotides (PDRNs) From Skin to musculoskeletal tissue regeneration via adenosine A2A receptor involvement. J Cell Physiol 2017;232:2299-307.
Lee JH, Han JW, Byun JH, Lee WM, Kim MH, Wu WH. Comparison of wound healing effects between Oncorhynchus
keta-derived polydeoxyribonucleotide (PDRN) and Oncorhynchus mykiss
-derived PDRN. Arch Craniofac Surg 2018;19:20-34.
Squadrito F, Bitto A, Irrera N, Pizzino G, Pallio G, Minutoli L, et al
. Pharmacological activity and clinical use of PDRN. Front Pharmacol 2017;8:224.
Habermann B, Oegema J, Sunyaev S, Shevchenko A. The power and the limitations of cross-species protein identification by mass spectrometry-driven sequence similarity searches. Mol Cell Proteomics 2004;3:238-49.
[Figure 1], [Figure 2], [Figure 3]