Brain Research Bulletin Vol 55, Issue 2, P 319-325, May 15, 2001 http://www.sciencedirect.com/science/journal/03619230 Serum concentrations of some metals and steroids in patients with chronic fatigue syndrome with reference to neurological and cognitive abnormalities S.J. van Rensburg(*,1), F.C.V. Potocnik(2), T. Kiss(3), F. Hugo(2), P. van Zijl(4), E. Mansvelt(5), M.E. Carstens(6), P. Theodorou(7), P.R. Hurly(8), R.A. Emsley(2) and J.J.F. Taljaard(1) 1 Department of Chemical Pathology, University of Stellenbosch Medical School, Tygerberg Hospital, Tygerberg, South Africa 2 Department of Psychiatry, University of Stellenbosch Medical School, Tygerberg Hospital, Tygerberg, South Africa 3 Department of Inorganic and Analytical Chemistry, University of Szeged, Szeged, Hungary 4 Department of Internal Medicine, University of Stellenbosch Medical School, Tygerberg Hospital, Tygerberg, South Africa 5 Department of Haematology, University of Stellenbosch Medical School, Tygerberg Hospital, Tygerberg, South Africa 6 Department of Cardiology, University of Stellenbosch Medical School, Tygerberg Hospital, Tygerberg, South Africa 7 National Centre for Occupational Health, Johannesburg, South Africa 8 C51 Edingight, Rondebosch, South Africa * Address for correspondence: Dr. S. J. van Rensburg, Department of Chemical Pathology, Tygerberg Hospital, P.O. Box 19113, 7505 Tygerberg, South Africa. Fax: +27-21-938-4640; email: sjvr@gerga.sun.ac.za Available online 17 July 2001. Abstract Chronic fatigue syndrome is defined by the Atlanta Centers for Disease Control (Atlanta, GA, USA) as debilitating fatigue lasting for longer than 6 months. Symptoms include disturbances of cognition. Certain factors have in the past been shown to influence cognition, including metals such as aluminum, iron, and zinc; and steroids such as dehydroepiandrosterone. In the present study, concentrations of these factors were determined in the serum and plasma of patients and their age- and gender-matched healthy controls (10 women and 5 men in each group). In addition, copper, dehydroepiandrosterone sulphate, cortisol, cholesterol, hemoglobin, ferritin and transferrin concentrations, as well as transferrin genetic subtypes were determined in both groups. The results indicate that patients had significantly increased serum aluminum and decreased iron compared to controls. In the females, serum iron and dehydroepiandrosterone sulphate were significantly decreased and correlated. Total cholesterol was significantly increased, and significantly negatively correlated with dehydroepiandrosterone sulphate. There were no differences in zinc, copper, cortisol, hemoglobin, transferrin and ferritin concentrations, or in transferrin genetic subtypes. Author Keywords: Chronic fatigue; Cognition; Aluminum; Iron; DHEAS; Cholesterol Introduction Cognitive deficits are a well-established feature of chronic fatigue syndrome (CFS) [3, 11, 22 and 34]. Metals in the brain have been implicated in cognitive function and dysfunction: aluminum (Al) is a recognised neurotoxin, and has been found to affect cognitive function in humans [1 and 20]. Zinc (Zn) is co-released with neurotransmitters by neurons involved with memory formation [46] and Zn supplements have been demonstrated to delay the progression of Alzheimer's disease [28]. In addition, low iron (Fe) concentrations [5 and 27] have been found to affect cognition. Furthermore, a mutation of transferrin (Tf), the protein which carries both Fe and Al into the brain, has a higher allele frequency in diseases such as rheumatoid arthritis [6] and Alzheimer's disease [42]. The mutation, TfC2, is associated with diseases thought to be caused by free radical damage [6 and 43], presumably because this substitution in the Tf molecule renders the Fe atoms more reactive towards oxygen species. Richards et al. [31] found evidence of inappropriate increases in free radical generation in patients with rheumatoid arthritis as well as CFS. Hence the question arose whether there would be a difference in allele frequency of TfC2 in patients with CFS. Dehydroepiandrosterone (DHEA) and its sulphate (DHEAS) have also been found to have effects on cognition, metabolism, and the immune system [29 and 30]. DHEA has been shown to alleviate amnesia and enhance long-term memory retention in mice [32]. In humans, oral administration of DHEA alleviated fatigue [30] and increased REM sleep while enhancing electroencephalogram activity in the sigma frequency range, suggesting a membranous effect of DHEA at the gamma-aminobutyric acid (GABA)_A/benzodiazepine receptor [12]. The aim of the present study was to investigate whether concentration differences would be found of some metals and steroids that have been associated with cognitive deficits, in patients with CFS. Materials and methods Ethical approval for the study was obtained from the Ethics Committee of the University of Stellenbosch Medical School, which included approval of an information and consent form which had to be signed by all participants. PATIENTS Diagnosis The patients were diagnosed by a team consisting of a physician (PvZ) and a psychiatrist (FH), according to the Atlanta Centers for Disease Control (Atlanta, GA, USA) criteria for research purposes [22 and 34]. A complete physical examination, urine analysis, and the recommended blood tests [34] were done to exclude other diseases which could cause fatigue. Fifteen patients, 10 women and 5 men, met the criteria and were further questioned by a psychiatrist (FCVP) about their physical and psychological symptoms, including fatigue and lethargy, headache and lightheadedness, cognition, day/night sweating, fever and chills, restricted breathing, paraesthesia/acrothesia (pins and needles or numb fingers and hands), and sleep disturbances. Controls were age- and gender-matched healthy volunteers. Symptoms All the CFS patients had psychiatric and somatic symptoms. As for cognitive function: impaired memory, thought processes, and concentration; prone to disorientation/detachment. As for mood disturbances: irritability, anxiety, depression; prone to suicidal ideation. Premorbidly: A-type personality, inclined to do excessive amounts of work. Morbidly: patients were `difficult', avoidant, labile, had poor frustration tolerance and bouts of aggression, or were apathetic, dependent, socially inappropriate/disinhibited, and prone to road rage. A general feature was unusual paraesthesias: `burns' or a `numb spot' (especially on the thigh or foot). SERUM AL Blood was collected with syringes and transferred to Al-free plastic tubes, taking care to exclude contamination with external Al. Serum Al was determined at the National Centre for Occupational Health in Johannesburg by means of graphite furnace atomic absorption spectrophotometry. FE STUDIES Concentrations of serum Fe and Tf iron binding capacity (TIBC; which is equivalent to Tf concentration) were done with a Beckman CX7 Analyser. Tf saturation was calculated according to standard equations. Hemoglobin (Hb) was measured with a Coulter Counter and ferritin was determined with a radioimmuno assay (RIA) method. Blood was taken at 0830 h from patients and controls in order to account for circadian rhythmicity. TF GENETIC SUBTYPES Tf genetic subtypes were determined by isoelectric focusing polyacrylamide gel electrophoresis according to a method described previously [42]. Composition of the gel: 2.91 g acrylamide, 0.09 g N, N'-methylene-bisacrylamide, 8.01 g sucrose; 2.7 ml Pharmalyte carrier ampholytes pH 4.5-5.4 (Amersham Pharmacia, Biotech UK Ltd, Buckinghamshire, UK) 0.3 ml Ampholine carrier ampholytes pH 5-7 (LKB) and 30 mul TEMED (N, N,N',N'-tetramethyl-ethylenediamine; Merck, Darmstadt, Germany). Anode and cathode solutions: 0.5 M H3 PO4 and 0.5 M NaOH, respectively. After prefocusing the gel for 30 min at 300 V, sample papers were applied 0.5 cm from the cathode. Power settings were 1000 V, 18 mA, and 8 W, for 4 h. The gel was stained for 10 min in 1 g Coomassie Brilliant Blue R250 (Merck) dissolved in 1:5:5 acetic acid:methanol:water (destaining solution) at 37 C, and destained in destaining solution for 1 h at 37 C. PLASMA ZN AND CU To determine plasma concentrations of Zn and Cu, blood was collected in plastic tubes with lithium heparin as anticoagulent. Zn and Cu were measured using a Varian Techtron Model 1200 atomic absorption spectrophotometer. SERUM DHEAS, CORTISOL, AND CHOLESTEROL An RIA method (Coat-a-Count; Diagnostic Products Corporation, Los Angeles, CA, USA) was used to determine serum levels of DHEAS in patients and controls. Serum cortisol was determined by a Technicon Immuno I and serum total cholesterol was determined by a Technicon DAX. STATISTICAL EVALUATION The results were analyzed using non-parametric statistics, i.e., the Mann-Whitney U-test and Spearman rank correlations. Results are given as the mean p/m the standard deviation. Results Serum Al concentrations in CSF patients were significantly higher (p<0.03; Mann-Whitney U-test) than in age- and gender-matched controls (10 women and 5 men in each group). Al concentration in patients was 0.314 p/m 0.193 mumol/l, and in controls 0.163 p/m 0.105 mumol/l (Fig. 1). Conversely, Fe concentrations in patients were significantly lower (p<0.02; Mann-Whitney U-test) than in controls: mean Fe concentration in patients was13.2 p/m 5.3 mumol/l, and in controls 19.1 p/m 6.4 mumol/l. In the female patients, Fe concentrations were highly significantly lower than in controls (p<0.0003; Mann-Whitney U-test; Fig 2). Mean Fe concentration in patients was 11.4 p/m 2.0 mumol/l and in controls 20.7 p/m 5.9 mumol/l. Tf saturation was also significantly lower (p<0.005; Mann-Whitney U-test; Table 1). However, interestingly, no differences were found in the concentrations of ferritin, Hb, or Tf concentrations (expressed as TIBC; Table 1). TABLE 1. Iron Status in Female Chronic Fatigue Syndrome Patients and Controlslegend -------------------------------------------------------------------- Patients Controls Difference -------------------------------------------------------------------- Iron (mumol/L) 12.9 (63.0) 21.1 (65.9) p<0.0003* Hb (g/dL) 14.1 (61.6) 13.7 (60.7) None Ferritin (ng/ml) 66.4 (632.4) 50.6 (648.9) None TIBC (mumol/L) 64.2 (610.3) 61.7 (611.3) None Tf saturation (%) 20.8 (67.2) 34.5 (69.4) p<0.005* -------------------------------------------------------------------- Hb, hemoglobin; TIBC, transferrin iron binding capacity; Tf, transferrin. * Significantly different, Mann-Whitney U-test. Although there seemed to be an inverse correlation between Al and Fe serum concentrations, this was not significant (Spearman rank correlation: r=-0.233, p=0.21). In a study by Huang et al. on haemodialysis patients, where a significant inverse correlation was found between serum Al and Fe, as well as between Al concentration and ferritin levels, it was hypothesised that Fe deficiency may be related to Al accumulation in dialysis patients [21]. In order to clarify the possibility of a competitive binding of Al and Fe to Tf, calculations were carried out to model serum conditions. Taking into account the reported Fe-Tf binding constants (log K1=21.4 and log K2=20.4) [16 and 40] and those for Al-Tf (log K1=13.5 and log K2=12.5) [17 and 18], Tf seems to be able to bind Fe ~8 orders of magnitude stronger than Al. The comparable mumol/l level of both Fe and Al in serum indicates that no real competition can be expected between the two metal ions for the Tf binding sites. Fe is bound very strongly to Tf and hence even the Fe bound to the low molecular mass fraction of serum (the most efficient potential binder is citrate) is rather low. At the same time a considerable part of Al (20-25%) is bound to the low molecular mass binders, mostly to citrate and phosphate. Accordingly, as the serum Tf concentration does not show significant alteration between patients and controls (Table 1), the significantly lower Fe saturation level of Tf would simply mean that ~30% more Tf binding sites are available for Al in patients, but not at all to displace bound Fe from the Fe-Tf complex. The TfC2 allele frequency in the CFS patients was 0.160, which was not different from control values of 0.136. There were no differences in plasma Zn and Cu in the CFS patients and controls. Zinc: patients 17.1 p/m 2.3 mumol/l; controls 17.7 p/m 2.6 mumol/l. Copper: patients 22.8 p/m 4.4 mumol/l; controls 24.6 p/m 7.8 mumol/l. The female CFS patients had significantly lower mean serum levels of DHEAS (p < 0.005; Mann-Whitney U-test) and higher cholesterol levels (p<0.001; Mann-Whitney U-test) than controls (FIG. 3 and FIG. 4). Mean DHEAS level in patients was 2.37 p/m 1.95 mumol/l and in controls 4.50 p/m 1.72 mumol/l. Mean cholesterol level in patients was 6.87 p/m 1.04 mmol/l and in controls 5.15 p/m 0.69 mmol/l. There was a significant inverse correlation (Spearman rank correlation: r= -0.481, p=0.03) between the DHEAS and cholesterol concentrations (Fig. 5), which could be linked to the inhibition of cholesterol synthesis by DHEA (see Discussion). Morning cortisol levels were not different from controls: patients 398 p/m 345; controls 361 p/m 169 nmol/l. Normal levels for women in our laboratory are: DHEAS 2.17-15.2 mumol/l; cortisol (a.m.) 171.1-800.4 nmol/l and cholesterol 3.80-5.70 mmol/l, respectively. The cholesterol results were similar to those found by Bates et al.[4]. Lower DHEA and DHEAS and unchanged cortisol levels in CFS patients were also found by Scott et al. [37] and DHEAS deficiency in Japanese patients by Kuratsune et al. [23]. In the present study, there was also a significant positive correlation (Spearman rank correlation: r=0.574, p=0.008) between DHEAS and Fe concentrations in the female patients (Fig. 6) which may possibly be explained through a cytokine mechanism (see Discussion) Discussion In the present study, significantly higher Al concentrations and lower Fe concentrations were found in patients with CFS than in controls. In the literature, high concentrations of Al and low Fe have been associated with cognitive deficits. Howard [20] found that children with marginally raised serum Al values were more likely to suffer from learning problems or hyperactivity, and adults from memory disturbances than people with low serum Al concentrations. Human studies on Fe deficiency anaemia show a negative effect on mental performance and psychomotor function [27]. Beard [5] speculates that the biological basis for the effects of Fe deficiency on cognitive dysfunction may be expressed through the dopamine pathway, because Fe deficiency caused significant decreases in brain dopamine D2 receptor densities in rats [5]. Such a dysfunction of neurones would lead to deficits in attention, arousal, and affect [5]. Huang et al. [21] found that body Fe parameters were highly significantly inversely correlated with serum Al in haemodialysis patients. Because Fe and Al are both carried by Tf, the question could be posed whether an increase in Al was a compensatory effect of Fe deficiency; however in the present study the inverse correlation between Al and Fe was not significant. The high difference in the Fe or Al binding capability of Tf, to the advantage of the former, makes competitive binding of Al ions negligible. Accordingly, iron deficiency may be related to Al accumulation only in a way that the unsaturated binding sites of Tf can bind more Al, and thus more Al can be transported into the brain. However, high Al load would not result directly in iron deficiency or would not lower Tf saturation because of Fe displacement by Al, which does not occur. On the other hand an increase in Tf saturation would hinder Al binding and transport to the brain. Al is a neurotoxin that enhances lipid peroxidation (membrane damage) caused by Fe [14 and 43]. Neurological abnormalities suggesting lesions in white matter of subcortical brain areas have been detected by magnetic resonance imaging in patients with CFS [24]. Central nervous system involvement including difficulty with concentration, attention, and memory may thus be related to neuronal membrane damage caused by free radicals. Even though Fe deficiency has been linked to peripheral symptoms such as paraesthesia [33] and visual problems [25] that improved markedly upon Fe treatment [45], we still feel that the problem in CFS is central rather than peripheral. Beutler et al. [8] found that iron supplementation was of benefit to chronically fatigued women who had Fe depletion of the bone marrow but not to women with adequate bone marrow Fe. It may thus be possible that individuals have an "optimum" Fe level. The reason for the low Fe values observed, especially in the women, was not obvious because the patients had adequate nutrition. In addition, the difference could not be ascribed to loss of Fe through menstruation, because menstrual histories were unremarkable in both groups of women, while of the 10 women with CFS, 5 had had a hysterectomy and none in the control group. Although serum Fe concentrations and Tf saturation were significantly lower than in controls, there were no differences in associated serum protein concentrations. These results were surprising because Fe deficiency anaemia is normally associated with increased Tf and decreased ferritin concentrations, while anaemia of chronic disease is associated with decreased Tf concentrations. This would seem to indicate that although serum Fe is low, this fact is not communicated to the brain in order to increase Fe uptake, suggesting inadequate synthesis of Tf in response to a decrease in Fe concentration. Because there was no difference in Tf genetic subtypes between patients and controls, the defect seems to lie in the regulation of Tf synthesis rather than in the structure of the Tf molecule or its Fe binding capacity. In addition, there may be a continual loss of Fe that cannot fully be replenished by normal intake. In trained athletes, sweating during training can result in low Fe levels [9]; hence the excessive sweating experienced by CFS patients may have the same result. In the present study, significantly lower DHEAS levels were found in CFS patients. Impaired production of steroids has previously been demonstrated in CFS [10]. Decreased DHEA(S) has also been implicated in Alzheimer's disease [41], in severe illness, systemic lupus erythematosus, and anorexia nervosa [19]. Plasma DHEAS levels correlate with many parameters of physical and mental well being, for example, in HIV-positive patients, DHEAS levels decrease as the disease progresses towards AIDS, the decrease occurring as symptoms (including dementia [36]) develop [19]. Immunological abnormalities in CFS [13, 15 and 39] may also be associated with decreased levels of DHEAS [29]. In addition, DHEA and DHEAS are excellent free radical scavengers [44]. The significant inverse correlation between DHEAS and Fe found in the present study could possibly be linked to the aberrations in cytokine production in CFS [13]. DHEA(S) has an immunoregulating function, its primary target being stimulation of the TH-1 subclass of the CD4+ T cell population, leading to increased lymphokine interleukin-2, while inhibiting interleukin-6 production [29]. Higher interleukin-6 levels have been found in patients with CFS [13] and in anaemia of chronic disease of unknown origin [26]. A prolonged activation of the immune system may also cause continual sequestration of Fe, rendering it unavailable for other body functions [7 and 38], while simultaneously increasing oxidative damage [2]. The high cholesterol values found in patients with CFS may be due to a mechanism of negative feedback control in the synthesis of DHEA from cholesterol: DHEA inhibits isoprenylation by depletion of endogenous mevalonate, a precursor of cholesterol synthesis [35]. Thus, decreased concentrations of DHEA may be accompanied by increased cholesterol concentrations. CONCLUDING REMARKS CFS manifests with both cognitive and somatic symptoms. Metals and adrenal steroids have been shown to play a role both in brain function and dysfunction, as well as the modulation of cytokines. Hence the results obtained in the current study reinforce the idea that the widespread symptoms of CFS would be better explained by a model indicating that the primary dysfunction is central rather than peripheral in origin. The pivotal part of this dysfunction may be the low serum Fe, reflecting exhaustive utilization (externally by sweating and internally by chronic inflammation), but not adequately replenished from food sources or body stores for some unknown reason. Normally a shortage of iron would result in an increased production of Tf to meet the body's demands, but in CFS the message for increased Fe uptake is thwarted. Acknowledgements We gratefully acknowledge the financial support given by the Provincial Administration: Western Cape and the Medical Research Council of South Africa. We also thank Dr Ben van der Walt for help with the manuscript. Figure captions Fig. 1. Serum aluminum (Al) concentrations in chronic fatigue syndrome (CFS) patients and controls (10 women and 5 men). The patients had a significantly increased mean serum Al concentration (p<0.03; Mann-Whitney U-test). Fig. 2. Serum iron (Fe) concentrations in female chronic fatigue syndrome (CFS) patients and controls (10 in each group). The mean Fe concentration in the patients was highly significantly lower (p<0.0003; Mann-Whitney U-test) than in the controls. Fig. 3. Serum dehydroepiandrosterone sulphate (DHEAS) concentrations in female chronic fatigue syndrome (CFS) patients and controls (10 in each group). The patients had a significantly lower mean DHEAS concentration than controls (p<0.005; Mann-Whitney U-test). Fig. 4. Serum cholesterol concentrations in female chronic fatigue syndrome (CFS) patients and controls (10 in each group). The patients had significantly higher cholesterol concentrations than controls (p<0.001; Mann-Whitney U-test). Fig. 5. Significant negative correlation between dehydroepiandrosterone sulphate (DHEAS) and total cholesterol in 10 female chronic fatigue syndrome (CFS) patients and 10 age- and gender-matched controls (Spearman rank correlation: r=-0.481, p=0.03). Fig. 6. Significant positive correlation between dehydroepiandrosterone sulphate (DHEAS) and iron concentrations in 10 female chronic fatigue syndrome (CFS) patients and 10 age- and gender-matched controls (Spearman rank correlation: r=0.574, p=0.008). References 1. A.C. Alfrey, Aluminum intoxication. N. Engl. J. Med. 310 (1984), pp. 1113-1115. 2. B.M. Babior, Oxidants from phagocytes: Agents of defense and destruction. Blood 64 (1984), pp. 959-966. 3. E.M. Barker, B.M. de Bruijn, E.J. Immelman, D.G.C. Presbury, A.D.P. van den Berg, D.A. Green and V.J. Pinkney-Atkinson, Clinical guidelines: Chronic fatigue syndrome. Committee for Science and Education, Medical Association of South Africa. South Afr. Med. J. 85 (1995), pp. 780-782. 4. D.W. Bates, D. Buchwald, J. Lee, P. Kith, T. Doolittle, C. Rutherford, W.H. Churchill, P.H. Schur, M. Wener, D. Wybenga, J. Winkelman and A.L. Komaroff, Clinical laboratory test findings in patients with chronic fatigue syndrome. Arch. Intern. Med. 155 (1995), pp. 97-103. 5. J. Beard, One person's view of iron deficiency, development, and cognitive function. Am. J. Clin. Nutr. 62 (1995), pp. 709-710. 6. L. Beckman and G. Beckman, Transferrin C2 as an enhancer of cyto- and genotoxic damage. Prog. Clin. Biol. Res. 209B (1986), pp. 221-224. 7. Beisel, W. R.; Blackburn, G. L.; Feigin, R. D.; Keusch, G. T.; Long, C. L.; Nichols, B. L. Proceedings of a workshop: Impact of infection on nutritional status of the host. Am. J. Clin. Nutr. 30:1203-1371, 1439-1566; 1977. 8. E. Beutler, S.E. Larsh and C.W. Gurney, Iron therapy in chronically fatigued, nonanemic women: A double-blind study. Ann. Intern. Med. 52 (1960), pp. 378-394. 9. J.C. Chatard, I. Mujika, C. Guy and J.R. Lacour, Anaemia and iron deficiency in athletes. Practical recommendations for treatment. Sports Med. 27 (1999), pp. 229-240. 10. M.A. Demitrack, J.K. Dale, S.E. Straus, L. Laue, S.J. Listwak, M.J.P. Kruesi, G.P. Chrousos and P.W. Gold, Evidence for impaired activation of the hypothalamic-pituitary-adrenal axis in patients with chronic fatigue syndrome. J. Clin. Endocrinol. Metab. 73 (1991), pp. 1224-1234. 11. F. Friedberg, L. Dechene, M.J. McKenzie, II and R. Fontanetta, Symptom patterns in long-duration chronic fatigue syndrome. J. Psychosom. Res. 48 (2000), pp. 59-68. 12. E. Friess, L. Trachsel, J. Guldner, T. Schier, A. Steiger and F. Holsboer, DHEA administration increases rapid eye movement sleep and EEG power in the sigma frequency range. Am. J. Physiol. 268 (Endocrinol. Metab. 31) (1995), pp. E107-E113. 13. S. Gupta, S. Aggarwal, D. See and A. Starr, Cytokine production by adherent and non-adherent mononuclear cells in chronic fatigue syndrome. J. Psychiatr. Res. 31 (1997), pp. 149-156. 14. J.M.C. Gutteridge, G.J. Quinlan, I. Clark and B. Halliwell, Aluminium salts accelerate peroxidation of membrane lipids stimulated by iron salts. Biochem. Biophys. Acta 835 (1985), pp. 441-447. 15. M.S. Harbuz and S.L. Lightman, Stress and the hypothalamo-pituitary-adrenal axis: Acute, chronic and immunological activation. J. Endocrinol. 134 (1992), pp. 327-339. 16. W.R. Harris, Binding and transport of aluminium by serum proteins. Coord. Chem. Rev. 149 (1996), pp. 347-365. 17. W.R. Harris and J. Sheldon, Equilibrium constants for the binding of aluminium to human serum transferrin. Inorg. Chem. 29 (1990), pp. 119-124. 18. W.R. Harris, G. Berthon, J.P. Day, C. Exley, T.P. Flaten, W.F. Forber, T. Kiss, C. Orvig and P.F. Zatta, Speciation of aluminium in biological systems. J. Toxicol. Environ. Health 48 (1996), pp. 543-568. 19. P.J. Hornsby, Biosynthesis of DHEAS by the human adrenal cortex and its age-related decline. Ann. N.Y. Acad. Sci. 774 (1995), pp. 29-46. 20. J.M.H. Howard, Clinical import of small increases in serum aluminum. Clin. Chem. 30 (1984), pp. 1722-1723. 21. J.Y. Huang, C.-C. Huang, P.S. Lim, M.-S. Wu and M.-L. Leu, Effect of body iron stores on serum aluminium level in hemodialysis patients. Nephron 61 (1992), pp. 158-162. 22. H.I. Kaplan and B.J. Sadok. Synopsis of psychiatry (8th ed.),, Lippincott Williams and Wilkens, Philadelphia (1998). 23. H. Kuratsune, K. Yamaguti, M. Sawada, S. Kodate, T. Machii, Y. Kanakura and T. Kitani, Dehydroepiandrosterone sulfate deficiency in chronic fatigue syndrome. Int. J. Mol. Med. 1 (1998), pp. 143-146. 24. G. Lange, J. DeLuca, J.A. Maldjian, H. Lee, L.A. Tiersky and B.H. Natelson, Brain MRI abnormalities exist in a subset of patients with chronic fatigue syndrome. J. Neurol. Sci. 171 (1999), pp. 3-7. 25. Y. Matsuoka, S. Hayasaka and K. Yamada, Incomplete occlusion of central retinal artery in a girl with iron deficiency anemia. Ophthalmologica 210 (1996), pp. 358-360. 26. M. Neidhart, P. Bruhlmann, S. Gay and B.A. Michel, Activation of CD4+ and CD8+ T-lymphocytes in bone marrow associated with reduced erythropoiesis in patients with chronic inflammation and anaemia. Schweiz. Med. Wochenschr. 128 (1998), pp. 1618-1623. 27. E. Pollitt, Iron deficiency and cognitive function. Annu. Rev. Nutr. 13 (1993), pp. 521-537. 28. F.C.V. Potocnik, S.J. Van Rensburg, J.J.F. Taljaard and R.A. Emsley, Zinc and platelet membrane microviscosity in Alzheimer's disease. The in vivo effect of zinc on platelet membranes and cognition. South Afr. Med. J. 87 (1997), pp. 1116-1119. 29. W. Regelson, R. Loria and M. Kalimi, Dehydroepiandrosterone (DHEA) - The "Mother Steroid". I. Immunologic action. Ann. N.Y. Acad. Sci. 719 (1994), pp. 553-563. 30. W. Regelson and M. Kalimi, Dehydroepiandrosterone (DHEA) -The multifunctional steroid. II. Effects on the CNS, cell proliferation, metabolic and vascular, clinical and other effects. Mechanism of action?. Ann. N.Y. Acad. Sci. 719 (1994), pp. 564-575. 31. R.S. Richards, T.K. Roberts, R.H. Dunstan, N.R. McGregor and H.L. Butt, Free radicals in chronic fatigue syndrome: Cause or effect?. Redox Rep. 5 (2000), pp. 146-147. 32. E. Roberts, L. Bologna, J.F. Flood and G.E. Smith, Effects of dehydroepiandrosterone and its sulfate on brain tissue in culture and on memory in mice. Brain Res. 406 (1987), pp. 357-362. 33. M.K. Rusak and V.N. Livshits, Paresteziia slizistoi obolochki polosti rta, obuslovlennaia defitsitom zheleza. Stomatologiia (Mosk.) 57 (1978), pp. 30-33. 34. A. Schluederberg, S.E. Straus, P. Peterson, S. Blumenthal, A.L. Komaroff, S.B. Spring, A. Landay and D. Buchwald, Chronic fatigue syndrome research. Ann. Int. Med. 117 (1992), pp. 325-331. 35. S. Schulz and J.W. Nyce, Inhibition of protein isoprenylation and p21ras membrane association by dehydroepiandrosterone in human colonic adenocarcinoma cells in vitro. Cancer Res. 51 (1991), pp. 6563-6567. 36. R.B. Schwartz, A.L. Komaroff, B.M. Garada, M. Gleit, T.H. Doolittle, D.W. Bates, R.G. Vasile and B.L. Holman, SPECT imaging of the brain: Comparison of findings in patients with chronic fatigue syndrome, AIDS dementia complex, and major unipolar depression. Am. J. Roentgenol. 162 (1994), pp. 943-951. 37. L.V. Scott, F. Salahuddin, J. Cooney, F. Svec and T.G. Dinan, Differences in adrenal steroid profile in chronic fatigue syndrome, in depression and in health. J. Affect. Disord. 54 (1999), pp. 129-137. 38. A. Shenkin, Trace elements and inflammatory response: Implications for nutritional support. Nutrition 11 (1995), pp. 100-105. 39. C. Song and B.E. Leonard. Fundamentals of psychoneuroimmunology, John Wiley & Sons, Chichester (2000). 40. H. Sun, M.C. Cox, H. Li and P.J. Sadler, Rationalisation of metal binding to transferrin: Prediction of metal-protein stability constants. Structure and Bonding 88 (1997), pp. 71-102. 41. T. Sunderland, C.R. Merril, M.G. Harrington, B.A. Lawlor, S.E. Molchan, R. Martinez and D.L Murphy, Reduced plasma dehydroepiandrosterone concentrations in Alzheimer's disease [letter]. Lancet ii 8662 ()1989), p. 570. 42. S.J. Van Rensburg, M.E. Carstens, F.C.V. Potocnik, A.K. Aucamp and J.J.F. Taljaard, Increased frequency of the transferrin C2 subtype in Alzheimer's disease. Neuroreport 4 (1993), pp. 1269-1271. 43. S.J. Van Rensburg, W.M.U. Daniels, F.C.V. Potocnik, J.M. Van Zyl, J.J.F. Taljaard and R.A. Emsley, A new model for the pathophysiology of Alzheimer's disease: Aluminium toxicity is exacerbated by hydrogen peroxide and attenuated by an amyloid protein fragment and melatonin. South Afr. Med. J. 87 (1997), pp. 1111-1115. 44. S.J. Van Rensburg, W.M.U. Daniels, J.M. Van Zyl and J.J.F. Taljaard, A comparative study of the effects of cholesterol, beta-sitosterol, beta-sitosterol glycoside, dehydroepiandrosterone sulphate and melatonin on in vitro lipid peroxidation. Metab. Brain Dis. 15 (2000), pp. 257-265. 45. J.C. Walker, D. Selva, G. Pietris and J.L. Crompton, Optic disc swelling in Crohn's disease. Aust. N.Z. J. Ophthalmol. 26 (1998), pp. 329-332. 46. G.L. Westbrook and M.L. Mayer, Micromolar concentrations of Zn2+ antagonize NMDA and GABA responses of hippocampal neurons. Nature 328 (1987), pp. 640-644. -------- (c) 2001 Elsevier ScienceDirect