European Journal of Endocrinology (2004) 151 U7–U12 ISSN 0804-4643 Stem cell research: immortality or a healthy old age? Christine Mummery Hubrecht Laboratory, Netherlands Institute for Developmental Biology and the Interuniversity Cardiology Institute of the Netherlands, Uppsalalaan 8, 3584 CT Utrecht, The Netherlands (Correspondence should be addressed to C Mummery; Email: [email protected])

Abstract Stem cell research holds the promise of treatments for many disorders resulting from disease or trauma where one or at most a few cell types have been lost or do not function. In combination with tissue engineering, stem cells may represent the greatest contribution to contemporary medicine of the present century. Progress is however being hampered by the debate on the origin of stem cells, which can be derived from human embryos and some adult tissues. Politics, religious beliefs and the media have determined society’s current perception of their relative value while the ethical antipathy towards embryonic stem cells, which require destruction of a human embryo for their derivation, has in many countries biased research towards adult stem cells. Many scientists believe this bias may be premature and basic research on both cell types is still required. The media has created confusion about the purpose of stem cell research: treating chronic ailments or striving for immortality. Here, the scientific state of the art on adult and embryonic stem cells is reviewed as a basis for a debate on whether research on embryonic stem cells is ethically acceptable.

European Journal of Endocrinology 151 U7–U12

Introduction was referred to as ‘transdifferentiation’ or adult stem cell ‘plasticity’. Stem cells are primitive cells with the capacity to self- In 1998, the first stem cell lines were derived from renew (divide and produce more of themselves) or to human blastocyst stage embryos (Fig. 1) by Thomson differentiate to specialized cells such as bone, brain et al. (1). This milestone in cell biological research and heart cells. Stem cells are often placed in one of had been preceded by more than 30 years of research two categories: adult stem cells derived from adult or on another very similar cell type, an embryonal carci- fetal tissues and embryonic stem (ES) cells derived noma or teratocarcinoma stem cell, derived from a from very early (mouse or human) embryos at the blas- spontaneous testis tumour found in mice and men tocyst stage of development (Fig. 1) prior to implan- (Fig. 1). In mice, teratocarcinomas can be formed tation in the uterine wall. experimentally by transferring young embryos to extra- Stem cells of bone marrow have been used for dec- uterine sites, for example, under the skin or kidney cap- ades in the successful treatment of blood disorders sule. Stem cells were derived directly from early mouse such as leukemia. Bone marrow transplantation to a embryos as mouse ES cells in 1981, circumventing the patient whose own bone marrow has been destroyed necessity of generating a tumour as an intermediate as part of treatment will allow the patient to start (2, 3). The way was then already open to deriving regenerating blood cell lineages and forming blood human ES cell lines as assisted reproduction by in with a normal cellular composition. Likewise, a skin vitro fertilization was becoming clinically routine. wound will result in stem cells in the basal layers Embryos were being discarded even after three to four resuming proliferation and differentiating to all cell had been simultaneously transferred to a patient types present in normal skin in a manner that recon- because no methods had been developed for successful structs the dermal and epidermal layers. Convention- freezing. These first attempts were not successful and ally, adult or tissue-specific stem cells in adults were took place under circumstances where there was no regarded as having the ability to proliferate at times of legislation governing human embryo research. tissue damage and to differentiate to the cell types in In the mid 1980s it became possible to freeze the tissue in which they were found. Exceptionally, in embryos and therefore sources dried up; legislation sex non-matched organ and bone marrow transplants, was developed in several countries that would even- where, for example, tissue from a male donor was given tually govern human embryo research. Thomson’s to a female recipient, there were occasional reports of article clearly demonstrated that the human ES cell some cells being found in multiple organs. They had lines that he had derived were immortal and pluripo- apparently adopted the phenotype of the local tissue. tent, i.e. they could form derivatives of the three pri- These observations provided the first evidence of what mary germ layers and in principle all tissues of the

q 2004 Society of the European Journal of Endocrinology Online version via http://www.eje.org

Downloaded from Bioscientifica.com at 09/26/2021 09:09:37AM via free access U8 C Mummery EUROPEAN JOURNAL OF ENDOCRINOLOGY (2004) 151

Figure 1 Pluripotent human stem cells are found in teratocarcinomas and in the inner cell mass (ICM) of blastocyst stage embryos. human body. The implications for cell-based therapies ethics believe that life begins at fertilization, other were immediately obvious and a list of diseases or ail- beliefs consider 40 days after conception as the crucial ments for which human ES cells might be useful time point in determining the moral status of the because only one or, at most, a few cell types were embryo. affected was easy to draw up (Table 1). The ‘price’ of In the following sections, the current status of this research, however, is destruction of human research on adult and ES cells (Table 2) is reviewed embryos that have the potential to become new (4). The conclusion is that research on neither adult human beings. The fact that these embryos were des- nor ES cells is sufficiently advanced for a definitive and tined to be discarded and, by implication, destroyed exclusive choice on whether one is better or worse because the gamete donors who had lead to their for- than the other as a basis for developing a broad range mation no longer wished to continue with their of stem cell therapies. Each is most likely to have its parent programme, was for many not an argument to own niche in therapy. In some cases the best option redirect their fate towards stem cell derivation however may be combined adult and ES cell therapy. If it were noble the cause of curing chronic disability. The simul- already proven that human ES cells are useful in taneous revisiting by a number of scientists of the abil- curing patients it would be much easier to justify sacri- ity of adult stem cells to transdifferentiate may have, in ficing embryos for those cures. However, it is not proven part, derived from the aversion to embryo destruction. and it is under those circumstances that we have to It is of note, however, that not all societies or religions decide whether derivation of new human ES cell lines adopt the same point of view: while Christian-based is justified for the purposes of research in the light of

Table 1 Examples of ailments for which solutions in stem cell therapy are being sought. Table 2 Stem cell origins.

Stroke Stem cells are found in: Parkinson’s disease 1. Adult tissues: all tissues and organs that can repair themselves Diabetes after damage or regenerate have adult stem cells. Examples Heart failure are: skin, bone marrow/cord blood, intestine, brain, liver, fat. Vascular disease Not ethically sensitive but there are very few cells present in Arthritis normal tissue and only a few different cell types can be formed. Multiple sclerosis 2. Embryos: embryonic stem cells. Spinal cord lesions Ethically very sensitive as destruction of the embryo is necessary but many cells available in the laboratory and all In general diseases associated with increasing age (in italics) where one or cells of the human body can be formed. at most a few cells types have been destroyed or malfunction.

www.eje.org

Downloaded from Bioscientifica.com at 09/26/2021 09:09:37AM via free access EUROPEAN JOURNAL OF ENDOCRINOLOGY (2004) 151 Stem cell research U9 the current availability of a number of human ES cell derived from his own bone marrow and implanted; lines already established. bone fide bone then developed and functioned as a normal ribcage. A second example concerned a young woman who underwent surgery for removal of a Adult stem cells in clinical practice or large tumour in her jaw. This left her with a serious clinical trials facial disfigurement. Again, tissue engineers recon- There are already a number of clinical applications for structed a bone matrix but now in the form of a jaw. several types of adult stem cells derived from different Her stem cells were seeded to the matrix and bone tissue sources (see Table 3). Some of these are autolo- developed in situ of such quality that artificial teeth gous, a patient receiving some of his own stem cells could be implanted and the jaw functioned essentially to repair a particular organ, others are heterologous as normal. This is an area being very actively pursued. and the stem cells are provided by a donor. The main drawback is an age limitation of the suit- ability and numbers of stem cells. Above the age of Pancreas duct stem cells These stem cells are cur- 50 years, the method generally does not work. Another rently being tested in clinical trials for the treatment drawback is the size of the engineered implant; the of ‘brittle’ diabetes. This form of diabetes is very seeded cells require adequate nutrients and blood severe, difficult to control with insulin and patients supply, and therefore only tend to survive well in the are often not aware of an imminent hyperglycemic inci- outermost regions of the bone matrix. Strategies are dent. The pancreatic duct contains a population of stem being developed to encourage the matrices to vascular- cells which can be expanded to a limited extent in cul- ize by also seeding with vascular endothelial cells. ture using a protocol developed in Edmonton, Canada which has as a result become known as the Edmonton Skeletal muscle stem cells for treatment following protocol. This is an example of heterologous transplan- heart attack Skeletal muscle has an undifferentiated tation. When first developed, the protocol required the population of stem cells known as satellite cells that ducts from five deceased donors for isolation, growth retain the capacity to differentiate to skeletal muscle. and differentiation of cells to b-islet-like colonies for In the case of cardiac patients, satellite cells are isolated the treatment of just one brittle diabetes patient. This from their muscle, expanded in culture and trans- has now improved to two donors for one patient and planted into the ischemic regions of the heart. While the current goal is to improve this to one donor for the transplanted cells do beat in situ, they fail to five patients. It will be clear, however, in the light of couple to the host myocardium and have been known the high incidence of diabetes worldwide and the scar- to induce fatal arrythmias. The arrythmias could some- city of donors, that this is by definition at present a limi- times be treated by drugs but the effects were severe ted therapy for only the very worst cases. enough for one clinical trial to be terminated prema- turely. Retrospectively, experiments in rats showed Mesenchymal stem cells ( from bone marrow or that there was no functional coupling between trans- fat) for bone and cartilage repair This is in principle planted and host muscle (5; references in review 4). an autologous treatment, for which there are several spectacular anecdotal clinical examples, several of ‘Fresh’ bone marrow injected directly into heart or which have been used in combination with tissue intravenously for treating heart attack The mech- engineering to create three-dimensional structures. anism of how this might work is unknown and the One example is of a boy born lacking the left side of effects are essentially unproven in humans because his ribcage so that there was little structure to protect no large-scale randomized trial has yet been carried his heart from injury. Tissue engineers constructed a out. Most results are anecdotal. The original experi- ribcage using an artificial bone matrix composed of ment upon which these trials were based was carried hydroxy-appetite. This was seeded with stem cells out in mice several years ago (6) but in more than 80 independent attempts, others have not been able to reproduce the finding. Two very recent publications Table 3 Clinical applications or ongoing clinical trials with adult stem cells. clearly demonstrated that the original results are likely to have been an artifact of the experimental Pancreas duct stem cells for ‘brittle’ diabetes. method (7, 8). There is no evidence of transdifferentia- Mesenchymal stem cells (from bone marrow or fat) for bone and tion of bone marrow stem cells to heart cells in animals cartilage repair. but it has clearly been shown that bone marrow stem Skeletal muscle stem cells for heart attack (but drugs are necessary to control heart rate). cells can fuse with host cells in heart, liver and brain ‘Fresh’ bone marrow injected directly into heart or intravenously and adopt those phenotypes with low frequency. This for treating heart attack (mechanism unknown and unproven in is probably the explanation of the data described humans but no evidence of transdifferentiation to heart cells in above concerning sex non-matched donors; trans- animals). Skin stem cells for skin transplantation. planted stem cells fuse locally at low incidence giving the impression that transdifferentiation has taken

www.eje.org

Downloaded from Bioscientifica.com at 09/26/2021 09:09:37AM via free access U10 C Mummery EUROPEAN JOURNAL OF ENDOCRINOLOGY (2004) 151 place. Such a mechanism is unlikely to be of benefit embryo, ES cells can grow indefinitely in culture in a since there will be no net increase in cardiac cell num- ‘primitive’ or undifferentiated state provided they are bers to replace those cells lost as a result of ischemic supplied with a so-called feeder cells. These cells provide damage. More importantly if nuclei also fuse during them with as yet unknown factors that encourage self- cell fusion, tetraploid nuclei may arise with an renewal without differentiation. Most of the human ES increased tumorigenic potential. It is of note, however, cells to date and all of those on a list approved by Presi- that in experiments in pigs with a myocardial infarction dent Bush for government (NIH) funding, have been into which genetically marked bone marrow was derived and cultured using feeder cells from mouse injected, the marked cells contributed to neovasculari- embryos. They are therefore perceived as presenting a zation in the infarct region; no marked cardiomyocytes xenorisk (such as cross-species recombination of were found but there were no significant numbers of viruses) to patients and are unlikely to be approved for vascular endothelial cells. Such a mechanism could clinical use. This is currently one of the driving forces be beneficial post-infarction but more experiments in behind the wish to derive new cell lines from human animals are necessary to investigate and optimize this embryos surplus to in vitro fertilization requirements. possibility. The bone marrow contains several different There are estimated to be hundreds of thousands of subpopulations of cells defined on the basis of how these worldwide; 400 000 in the US alone. ES cells they are selected. It is not clear which population show a strong resemblance to teratocarcinoma (EC) might be most useful if neovascularization were the stem cells. Many markers have been developed by path- essential mechanism behind any effect. The endothelial ologists to determine whether there was a residual stem precursor cell population may well be the one of choice. cell population in teratocarcinomas (which in turn In this area there is particularly sensitive discord determines whether it is malignant or not, and thus between scientists and physicians. While scientists the patient’s prognosis and treatment). The cells are prefer to optimize and understand the mechanism immortal and, just as cancer cells, express telomerase before considering application, physicians – faced activity that controls the telomere length protecting sometimes with terminal patients – do not necessarily the ends of chromosomes at every cell division. regard understanding how something works as a Human ES cells form derivatives of all three embryonic high priority. That it might work and is not dangerous germ layers in vitro and in teratocarcinomas in mice is enough. For review see Hassink et al. (9). in vivo. This means that it is most likely that they can form all ,200 cell types of the human body. It also Skin stem cells for skin transplantation This tech- means, however, that they may form teratomas post- nology is making a significant contribution, particu- transplantation in patients. Rigorous animal exper- larly to the treatment of burns patients and patients imentation will be required to establish their safety; with varicose ulcers. Stem cells are taken from skin i.e. to determine whether they are likely to form biopsies, expanded in culture and grown on ‘rafts’ of tumors or not in the event that undifferentiated cells extracellular matrix proteins at an air–water interface. contaminate the differentiated cell populations to be In the presence of a growth factor cocktail and ascorbic transplanted. acid, they will differentiate into several layers of skin A variety of protocols have been useful in driving the types, which can be transferred as sheets to large differentiation of mouse ES cells to particular lineages, areas devoid of skin. the most useful of which has been the growth of the Together, these form the major areas in which adult cells as aggregates in suspension as so-called embryoid stem cells have proven or are proving their worth. bodies. Addition of various trophic or growth factors Research on neural stem cells derived from adult or has resulted in cells becoming biased towards particu- fetal brain is also actively being carried out. However, lar lineages. It has thus been possible to derive cell the results of clinical trials on the usefulness, for populations enriched with neural cell precursors or car- example, of human fetal brain stem cells for the treat- diac cell precursors. Pure populations of differentiated ment of Parkinson’s disease, have been somewhat con- cells are, however, never obtained without some tradictory with positive and no effect outcomes being methods for selection of the required phenotype. reported. These may include using the promoter of a lineage or There is little evidence as yet that cord blood will be cell type specific gene coupled to an antibiotic resist- useful for more than the treatment of blood disorders, ance gene for example. Addition of an antibiotic then including leukemia, as it is now. results in the death of all cells not expressing the gene since they are not protected. The required cell population becomes selected in culture. Human ES cells Copying the signals used by an embryo to create different cell types from common precursors could Human ES cells are derived from a small group of cells lead to new methods for deriving cells suitable for present in the early ‘pre-implantation’ embryo at the transplantation. We have used this approach to derive ‘blastocyst’ stage (Fig. 1). After isolation from the cardiomyocytes (heart cells) from human ES cells.

www.eje.org

Downloaded from Bioscientifica.com at 09/26/2021 09:09:37AM via free access EUROPEAN JOURNAL OF ENDOCRINOLOGY (2004) 151 Stem cell research U11

Reasoning that the heart develops adjacent to the inva- cells would probably require replacement. Here, bio- ginating gut and using evidence from species such as reactors in combination with (genetic) selection are chick and frog, where defects in endoderm development likely to be useful. How can requirements of ‘good medi- result in cardiac defects, we co-cultured human ES cells cal practice’ (GMP) be achieved (i.e. no mouse feeder with visceral endoderm-like cells and obtained beating cells, no fetal calf serum, etc.)? Some human ES cell muscle cells with a phenotype resembling human lines have already been derived on human feeder cells fetal ventricle (10, 11). We have recently transplanted but it is hoped that, in the near future, the relevant these to the hearts of immuno-compromised but other- factors for self-renewal will have been identified. Most wise healthy mice and shown long-term survival and of the necessary ‘ingredients’ are available now at functional integration of the human ES derivatives. GMP grade. We do not yet know whether these cells would ‘cure’ How can potential problems of tissue rejection be a mouse with a myocardial infarction or ischemic resolved? This should be no problem for transplantation damage. These experiments thus join several others to the brain/central nervous system, but what about that have shown successful transplantation to mice or other organs? Immunosuppressive drugs and tissue rats. Human ES cell derived neurons survive transplan- ‘matching’ as carried out for whole organ transplants tation to the brains of young mice; endothelial cells are likely to be the most useful option. With this in from human ES cells have now been transplanted mind a stem cell bank has already been established into three-dimensional matrices then into immunodefi- by the Medical Research Council in the UK and cient mice where they form capillaries through which Sweden. Immunotolerance is another option. It has mouse blood circulates (reviewed in reference 1). been known for many years that if a patient receives Otherwise, transplantation studies with human ES bone marrow from a kidney donor prior to transplan- cells to date are still fairly limited. Exceptionally, in the tation, chimeric bone marrow can develop in vivo and US, neural cells derived from teratocarcinoma stem the patient is later immunotolerant for the transplanted cells were transplanted to the brains of 11 stroke patients kidney. Finally therapeutic cloning has been discussed in a phase I trial (safety). In 1999, the transplanted cells as an option. Proof of principle was recently reported were still detectable by positron emission tomography by a group in Korea (13). They showed that they (PET) scan in all patients and in 2000, tumour develop- could remove a nucleus from a human oocyte (egg ment was not detected in any of the 11 patients. In 6 cell), replace it by a somatic cell and allow the recon- patients, symptoms ‘improved’; in 2, this improvement structed embryo to grow to the blastocyst stage. From was reported as significant (12). Nevertheless, such this blastocyst they could isolate human ES cells con- ‘experiments’ on humans are regrettable. If only a few taining the derivatives of the somatic nucleus. If the undifferentiated cells had been present the patients somatic nucleus were from a patient, for example may have developed teratomas or, worse, teratocarcino- from the skin, then the human ES cell derivatives mas. The whole field would have been brought into would be recognized as self and would not be rejected discredit before it had even begun. at transplantation. Insulin-producing cells, bone cells, blood cells and In summary, human ES cells are at present an excit- liver cells have all been reported to develop from ing research challenge but we have now entered a human ES cells but have not yet transplanted to phase that should include quiet, in-depth research to animal models. The majority of transplantation studies solve the many scientific problems that still bar the and the most spectacular evidence of control of self- way to clinical application. This is the light in which renewal and differentiation have been carried out only their current value should be debated. on mouse ES cells. This is largely where over interpret- ation by the media originates since the extrapolation from mouse to human is (too) easily made. A simple case in point; the self-renewal factor from feeder cells Conclusions for mouse ES cells is leukemia inhibitory factor (LIF) There are at present three main conclusions: but this does not work on human ES and the ‘feeder factor’ remains unknown to date. The indications are, (i) Rapid advances are being made in both adult and however, that the same transcription factors in both ES cell research. However, a great deal more species are involved in control of self-renewal. research is necessary before either of these While it is unquestionable that human ES cells have becomes part of therapy in mainstream medicine, a much wider differentiation potential than adult stem apart for the treatment of blood disorders. cells, it is still fairly difficult to grow them and to control (ii) It is a pity that interest in stem cell research has their differentiation. Questions that need to be become ‘hyped’ in the media. The practical and addressed before human ES cells can be used clinically ethical problems are far from solved (14). Never- include the following. How can we derive pure theless, the potential of stem cells in the develop- populations of cells in sufficient numbers for transplan- ment of therapies for chronic disorders remains tation? For example, for myocardial infarction 108-109 and they could represent a major contribution

www.eje.org

Downloaded from Bioscientifica.com at 09/26/2021 09:09:37AM via free access U12 C Mummery EUROPEAN JOURNAL OF ENDOCRINOLOGY (2004) 151

to treatment of diseases likely to increase in the 3 Evans MJ & Kaufmann MH. Establishment in culture of coming decennia as the general population ages. pluripotential cells from mouse embryos. Nature 1981 292 154–156. (iii) Adult stem cells are preferable for ethical reasons 4 Passier R & Mummery CL. The origin and use of embryonic and and because they can, in principle, be autologous. adult stem cells in differentiation and tissue repair. Cardiovascular However, present scientific evidence is that they Research 2003 58 324–335. are unlikely to be broadly applicable because 5 Menasche P,Hagege AA, Scorsin M, Pouzet B, Desnos M, Duboc D, Schwartz K, Vilguin JT & Mardlean JP. Myoblast transplantation their developmental potential is limited, their for heart failure. Lancet 2001 357 279–280. numbers are extremely low and only few have 6 Orlic D, Kajstura J, Chimenti S et al. Bone marrow cells repair the proven amenable to ex vivo expansion. Commer- infarcted myocardium. Nature 2001 410 701–705. cial companies offering services to freeze cord 7 Murry CE, Soonpaa MH, Reinecke H, Nakajima H, Nakajima HO, blood or bone marrow for the donors own Rubart M, Pasumarthi KB, Virag JI, Bartelmez SH, Poppa V, Bradford G, Dowell JD, Williams DA & Field LJ. Haematopoietic future use should thus be viewed with caution. stem cells do not transdifferentiate into cardiac myocytes in myo- Most important is that open dialog, faithfully reflecting cardial infarcts. Nature 2004 428 664–668. 8 Balsam LB, Wagers AJ, Christensen JL, Kofidis T, Weissman I & the scientific facts and as far as possible including con- Robbins RC. Haematopoietic stem cells adopt mature haemato- sistent ethical values without political bias, is possible poietic fates in ischemic myocardium. Nature 2004 428 in this highly sensitive field. Elizabeth Blackburn, a 668–670. scientist highly respected for her work on telomerase, 9 Hassink RJ, Brutel de la Riviere A, Mummery CL & Doevendans PA. Transplantation of cells for cardiac repair. Journal of the was sacked from President Bush’s Council for Bioethics American College of Cardiologists 2003 41 711–717. for criticizing a report (Report on Monitoring Stem Cell 10 Mummery CL, Ward D, van den Brink CE, Bird SD, Research (15)): ‘the scientific results have been manipu- Doevendans PA, Opthof T, Brutel de la Riviere A, Tertoolen L, lated to bias away from ES cells towards adult stem cell van der Heyden M & Pera M. Cardiomyocyte differentiation of research’. This is counterproductive for scientists, mouse and human embryonic stem cells. Journal of Anatomy 2002 2000 233–242. society and politicians alike. 11 Mummery CL, Ward-van Oostwaard D, Doevendans PA, Spijker R, The future? Hopefully healthy old age and not van den Brink S, Hassink R, van der Heyden M, Opthof T, Pera M, immortality. Brutel de la Riviere A, Passier R & Tertoolen L. Differentiation of human embryonic stem cells to cardiomyocytes: the role of co- culture with visceral endoderm-like cells. Circulation 2003 107 2733–2740. Acknowledgements 12 Kondziolka D, Weschler L, Goldstein S, Meltzer C, Thulborn KR, Gebel J, Jannetta P, DeCesare S, Elder EM, McGrogan M, I thank Dorien Ward, Robert Passier, Stieneke van den Reitman MA & Bynum L. Transplantation of cultured human Brink, Leon Tertoolen and Rutger Hassink for their con- neuronal cells for patients with stroke. Neurology 2000 55 tributions to the work on cardiomyocytes from human 565–569. ES cells, Alan Trounson and Martin Pera for providing 13 Hwang WS, Ryu YJ, Park JH, Park ES, Lee EG, Koo JA, Chun HY, Lee BC, Kang SK, Kim SJ, Ahn C, Hwang JH, Park KY, Cibelli JB & the human ES cell lines used and Embryonic Stem Cell Moon SY. Evidence of a pluripotent human embryonic stem cell International for financial support. line derived from a cloned blastocyst. Science 2004 303 1669–1674. 14 de Wert G & Mummery CL. Human embryonic stem cells: References research ethics and policy. Human Reproduction 2003 58 324–335. 15 President’s Council on Bioethics (2004): Monitoring stem cell 1 Thomson JA, Itskovitz-Eldor J, Shapiro SS, Waknitz MA, research. http://bioethics.gov/reports/stemcells/index Swiergiel JJ, Marshall VS & Jones JM. Embryonic stem cell lines derived from human blastocysts. Science 1998 282 1145–1147. 2 Martin GR. Isolation of a pluripotent cell line from early mouse embryos cultured in medium conditioned by teratocarcinoma Received 18 May 2004 stem cells. PNAS 1981 78 7634–7638. Accepted 10 August 2004

www.eje.org

Downloaded from Bioscientifica.com at 09/26/2021 09:09:37AM via free access