JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 28 1 Progress With the Speciation of Aluminium and Silicon in Serum of Chronic Renal Patients Using Atomic Spectroscopic Techniques* K. Wrobel,t E. Blanco Gonzalez and A. Sanz-MedelS Department of Physical and Analytical chemistry University of Owiedo clJulian Cla weria 8 33006 Oviedo Spain In order to provide further evidence on the possible correlation between aluminium and silicon levels in the serum of renal failure patients and the possibility of the reduction of aluminium bioavailability by the presence of silicon in biological fluids the effects of different factors including storage conditions administration of desferrioxamine (DFO) and kidney transplantation on total aluminium and silicon contents and on their distribution in the same serum samples were examined and compared.Ultramicrofiltration was used for the separation of low molecular weight (LMW) and high molecular weight (HMW) serum fractions and electrother- mal atomic absorption spectrometry (ETAAS) for the determination of both elements. Consistent results were obtained showing that the distribution of aluminium between LMW and HMW serum fractions is a constant factor in the absence of DFO. It was observed that 11 52O/0 of the total aluminium in serum is ultrafiltrable and this value does not seem to be influenced by the total serum elemental concentration storage conditions particular renal pathology of the patients or kidney transplantation. However kidney transplantation induces a ‘clear-up’ of serum aluminium and silicon which is easily observable after a few months.Administration of DFO alters the speciation of aluminium by increasing its relative content in the LMW fraction up to 75&6% of the total element concentration in serum. Conversely distribution of silicon in serum proved to be affected only by the storage conditions. If the sample is stored properly (the pH maintained below 7.8) the ultrafiltrable silicon content results were consistent and reproducible (43ZfI3O/0 of total serum silicon in the LMW fraction was found to be ultrafiltrable). In any case silicon binding to serum proteins must be different to that observed for aluminium (which is mostly bound to transferrin). Moreover the observed distribution of aluminium between LMW and HMW serum fractions was neither related to silicon total concentration nor to the distribution of silicon in serum.Keywords Aluminium; silicon; speciation; ultramicrofiltration; electrothermal atomic absorption spectrometry Strong evidence exists for the toxicity of aluminium to renal failure patients.’ The aluminium body burden is mainly associ- ated with osteodystrophy microcytic anaemia and encepha- l~pathy.’.~ However the mechanism(s) of this toxicity remains ~nknown.~ It is clear that aluminium bioavailability and hence toxicity depends on the physicochemical form of the element as does its transportation within the body fluids and to the tissues where the toxic effects can be ob~erved.~ It has recently been suggested that silicon can reduce the bioavail- ability of aluminium by forming hydroxyaluminosilicates under physiological condition^.^,' Speciation studies of aluminium and silicon together in body fluids are therefore of fundamental importance to provide useful information on the possible interactions between both elements in a physiological environ- ment.The distribution of aluminium between low molecular weight (LMW) and high molecular weight (HMW) serum fractions has been studied in this laboratory using conventional ultrafiltration’ and ultramicrofiltration The dis- tribution of silicon was investigated by ultramicrofiltration and gel filtration.’ Reported results’ indicate that the presence of the main serum proteins (transferrin and albumin) affects the speciation of silicon in model solutions and also in uremic serum samples even if the character of the protein-silicon interaction seems to be rather different (less specific) than that observed for aluminium’ (which is chemically bound to transferrin4).The aim of the present work was to provide further exper- imental evidence of the possible relationship between the distribution of silicon and of aluminium in serum. Therefore the partition of both elements between LMW and HMW ~~~~ ~ ~ * Presented at the XXVIII Colloquium Spectroscopicum t On leave from Institute of Chemistry University of Warsaw $ To whom correspondence should be addressed. Internationale (CSI) York UK June 29-July 4 1993. Bialystok Branch Poland. serum fractions in the same samples was studied using ultramic- rofiltration as a separation technique [chosen for its simplicity small sample size requirements ( 1 ml) and easy control of ~ontamination’~] and electrothermal atomic absorption spectrometry (ETAAS) was used as a specific detector for aluminium and silicon in determinations in serum and in the ultrafiltrable fraction^.'^^'^ The influence of some important factors including sample storage previous administration of desferrioxamine (DFO) to patients and kidney transplantation on silicon and aluminium distribution (speciation) in serum was investigated.The results obtained from all these experi- ments are given and discussed mainly from the point of view of likely (or otherwise) interactions between silicon and alu- minium that could reduce the bioavailability of aluminium (i.e. its LMW fraction).Experimental Apparatus The ETAAS determination of aluminium and silicon concen- trations in the serum samples and their ultrafiltrable fractions was carried out using a Model 3030 Perkin-Elmer atomic absorption spectrometer with a Model HGA-500 graphite furnace equipped with an autosampler Model AS-40 and a PR-100 printer. Ultramicrofiltration experiments were carried out using the Amicon micropartition system (MPS-l) which has been described elsewhere,’ fitted with Amicon YMT membranes (nominal cut-off 30 000 Da). All the instrumentation was housed in a clean room equipped with a filtered laminar air supply to avoid sample contami- nation from aluminium and silicates in dust. Reagents All chemicals used in this work were of analytical-reagent grade. Stock solutions containing 1.000 g 1-’ of silicon and282 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL.9 aluminium were provided by Merck. Intermediate standard solutions of 100mgl-l of the element were prepared by diluting 10ml of the appropriate stock solutions with dilute (1 + 20) nitric acid to 100 ml. Working standard solutions were prepared daily by appropriate dilution of the intermediate solution with Milli-Q water (Millipore). Clinical Samples A serum pool from persons with normal renal function and serum samples from seven patients with end-stage renal failure on regular haemodialysis therapy were provided by the Hospital Central de Asturias in Oviedo Spain. In this latter case samples were always collected before the corresponding dialysis session.During this session DFO (mg kg-' of body mass) was administered and then for each patient 48 h later (before the next dialysis session) the blood was sampled again. Serum samples from 12 renal failure patients initially col- lected before kidney transplantation and then collected again 1 month and 6 months after the surgical operation were provided by the Hospital Universitario 'Marques de Valdecilla' de Santander in Santander Spain. All patients had their kidney function normalized after the first month from the transplantation (urinary creatinine level < 2 g per day). The normal serum pool used as a reference was spiked with aluminium and silicon aqueous standards up to final total elemental contents of 40 pg 1-l and 0.40 mg l-l respectively. For the experiments on the influence of storage conditions and pH of the sample on the partitioning of silicon and aluminium in serum the pH was adjusted with 1 mol I-' sodium hydroxide solution (Merck).Procedures Risks of sample contamination with external aluminium and silicon during sampling transport storage fractionation and determination steps were minimized following the detailed procedures described in previous ~ o r k . ~ ~ ~ * ~ ~ ~ ~ Samples were collected in polystyrene tubes that had been treated with 10% v/v nitric acid for 24 h before use. All components of the MPS-1 ultramicrofiltration system were also soaked for 24 h in 10% v/v nitric acid and rinsed with Milli-Q water before use. The YMT membranes were washed by ultrafiltering twice with 1 ml of 0.1 mol 1-I sodium hydroxide solution and then with ultrapure water until the washings were 'free' from aluminium and silicon (as established by ETAAS).The operating conditions for ETAAS determination of alu- minium and silicon in serum and ultrafiltrates were the same as described previ~usly.'~.~~ Serum samples were diluted with Milli-Q water 1 + 1 and 1 + 3 for total aluminium and silicon determinations respectively. No pre-treatment was necessary for ultrafiltrable serum fractions and in all cases calibration was accomplished by the corresponding standard aqueous solutions of aluminium and silicon. For the separation of LMW and HMW serum fractions 1 ml aliquots of the serum samples were introduced into the ultramicrofiltration cell and then centrifuged (1800g) for 15 min at room temperature.Results and Discussion The procedures described above were applied to an investi- gation of the total aluminium and silicon serum contents in different situations and their distribution (speciation) between the LMW and HMW serum fractions in the same samples under various conditions. The observed influence of serum storage conditions previous administration of DFO to patients and kidney transplantation is presented in the following sections. Influence of Serum Storage Conditions on Ultrafiltrable Serum Aluminium and Silicon Preliminary experimentsg on the distribution of aluminium and silicon were carried out using aged uremic serum samples stored for a long time (more than 3 months) in a refrigerator. It was found that results for the distribution of silicon were rather scattered 12-50% of the total concentration of silicon was present in the ultrafiltrable fraction depending on the sample analysed. Conversely the ultrafiltrable fraction of serum aluminium in the same aged samples ranged from 9 to 13% (mean 11 +2%).The pH was measured in all analysed serum samples and it varied between 7.5 and 8.6. It was observed that the lowest values of silicon in the ultrafiltrable fraction of those serum samples always occurred for pH values above 8. In order to investigate this point further a fresh normal serum pool (adequately spiked to obtain final total concentrations of 40 pg 1-l of aluminium and 0.40 mg I-' of silicon) was prepared and divided into five aliquots which were stored under con- trolled but differing conditions as indicated in Table 1.It was verified that additions of spikes (1-8 pl) did not change the pH of the serum pool (20ml). The distribution of aluminium and silicon in those sera aliquots were then studied following procedures given previously. The results have been summarized in Table 1 which shows that virtually the same relative distri- bution of aluminium was observed for any of the sample storage conditions tested (11+2% of the total serum alu- minium is present in the ultrafiltrate). The distribution of silicon in the fresh and in adequately frozen serum aliquots was virtually the same [ultrafiltrable silicon around 43% of the total concentration of silicon in the sample (Table l)]. However if samples are not frozen the ultrafiltrable fraction of serum silicon decreased to 40% in samples stored at 4°C for 24 h and the level dropped to around 17% for samples stored at room temperature for 24 h (Table 1). It was observed that unfrozen samples at room temperature had elevated pH values (pH>8 Table 1) probably owing to some loss of the C02 content.Finally when the pH of fresh serum samples was artificially increased with dilute sodium hydroxide solution to pH 8.6 a low value of silicon content in the LMW fraction of serum (ultrafiltrable fraction 18_+2%) was also found while the distribution of aluminium remained virtually unchanged as is shown in Table 1. Influence of Administration of DFO on Total and Ultrafiltrable Serum Aluminium and Silicon Levels Desferrioxamine (DFO) is a drug presently used to reduce the total body burden of aluminium in renal failure patients,l4 for aluminium de-toxification. The administration of this drug helps to clear up other trace elements from the body of such patients as has been shown recently for c0ba1t.l~ Therefore the effect of administration of DFO on the total elemental concentrations of silicon and aluminium in serum and on the partioning of these elements between LMW and HMW serum Table1 Influence of sample storage conditions on sample pH and on ultrafiltrable aluminium and silicon in normal spiked serum pool ~ ~~~~ Amount ultrafiltrable (%) Storage condition n pH Aluminium Silicon Stored (- 20°C) Fresh sample 5 7.4-7.6 11.6 1.9 46.5 f 1.8 up to 3 months 5 7.5-7.8 11.9 f 2.3 43.3 1.3 Stored (4°C) 24 h 5 7.6-7.8 11.2_+ 1.4 41.1 k0.3 Stored (25°C) 24 h 5 8.0-8.6 12.6 & 2.3 16.5 1.5 Fresh sample + NaOH* 5 8.6 12.2& 1.8 18.0+ 1.5 MeankSD 11.9L0.5 33.1A14.6 *pH was artificially adjusted to pH = 8.6.JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY MARCH 1994 VOL.9 283 fractions was investigated. To do so ultramicrofiltration experi- ments on seven different uremic serum samples from patients who were undergoing de-toxification therapy with DFO (mg kg-' of body mass) was undertaken. The results observed for all these experiments are summarized in Fig. 1 (a) and (b). As shown in Fig. l(a) 48 h after DFO infusion both total and ultrafiltrable aluminium levels in serum increased abruptly from a mean of 56 k 30 to a mean of 170 + 65 pg 1-l and from 7.3 k 3.1 to 107 & 50 yg l-' respectively.As expected,'v'' the relative amount of aluminium in the ultrafiltrate increased more than the total aluminium indicating that the LMW content of aluminium represents up to 74.9&6.3% of the total serum aluminium concentration in DFO treated sera.','' In the case of silicon the total element serum concentration in the patients was found to be 0.95 k0.65 mg 1-l before administration of DFO and 0.97f0.60 mg 1-l 48 h after administration [Fig. l(b)]. Thus it appears that the total silicon content in serum was not affected by the DFO therapy. Moreover in Fig. l(b) it is shown that the LMW or ultrafiltr- able fractions contained 43.3 f 1.3% of the total serum silicon before and 44.2+4.7% after administration of DFO. Thus the conclusion to be drawn is that distribution of silicon in serum is also unaffected by DFO therapy.It should be mentioned that the results presented in Fig. 1 were obtained for samples stored at -20°C at pH<7.8. In parallel with the above experiments some unfrozen samples (pH > 8) were also tested and lower silicon contents were found in the ultrafiltrable serum fraction (about 20% of total silicon serum concen- tration) supporting previous findings on the effect of age and pH of sera samples (storage conditions). Influence of Kidney Transplantation on Total and Ultrafiltrable Serum Aluminium and Silicon Levels Serum samples were collected from 12 uremic patients before and after kidney transplantation. These samples were analysed for total and ultrafiltrable aluminium and silicon following the previously described procedures in order to evaluate the effect of kidney transplantation on the total content and distribution of silicon and aluminium in serum.The results obtained have been plotted in Fig. 2. As shown by Fig. 2(a) the total serum aluminium content decreased from a mean of 55.4f4.2 before the transplantation to 35.8+2.6pgl-' one month later and to 19.9k9.5 pg1-' six months after the transplantation. Similar decreases were observed for the aluminium content in the ultrafiltrable sera fraction of the patients undergoing operation (mean values 8.1k4.5 4.8k2.6 and 3.2f0.3 pgl-' respectively) [Fig. 2(b)]. Therefore the percentage of ultrafiltrable or LMW aluminium remained virtually unchanged during the whole period of time after transplantation that was investigated (before surgical operation 11.2+ 1.3 12.3 +2.7% after one month from trans- plantation and 12.3 f 1.8% after six months from the transplan- tation) [Fig.2(a)]. Considering the silicon contents it can be seen from Fig. 2(b) that the total concentration in the same serum samples decreased from 1.30k0.60 mg 1-' before the operation to 1.10+0.50 mg 1-1 one month and to 0.73f0.21 mgl-I six months after transplantation. As regards the experiments on silicon distribution unfortunately the serum samples were transported unfrozen from Santander to Oviedo (about a 12 h journey). Therefore the distribution of silicon was probably altered with respect to that actually existing in the fresh sera of the patients. Fairly elevated pH values (pH > 8) were found in all of these samples.This would explain why the relatively ultrafiltrable silicon content was 18.1 4.9 before operation 19.0 f 5.6 one month afterwards and 20.8 & 2.8% six months after operation that is the type of values observed are consist- ent for 'aged' serum samples (see in Table 1 the samples stored for 24 h at 25 "C) Conclusions It has been shown that the total aluminium and silicon contents in serum and the distribution of these elements between LMW and HMW serum fractions are affected in different ways by 200 c I - o 150 I s 0 E 100 00 3 50 n C A Pre-DFO Post-DFO 80 Pre-DFO Post-DFO Fig.1 Influence of administration of DFO on A total and B ultrafiltrable levels of (a) aluminium and (b) silicon in serum; C shows the ultrafiltrable percent of the element.Uremic serum samples 7.4 < pH < 7.6 - 01 a 3 40 a 0 4- 3 20 0 Before After After transplantation 1 month 6 months 50 (bl Before After After transplantation 1 month 6 months Fig.2 Influence of kidney transplantation on A total and B ultrafiltrable levels of (a) aluminium and (b) silicon; C shows the ultrafiltrable percent of the element. Serum pH> 8284 JOURNAL OF ANALYTlCAL ATOMIC SPECTROMETRY MARCH 1994 VOL. 9 the factors studied here indicating different speciation mechan- isms for both elements. While as expected,8.'0g16 significant increases in the total aluminium serum content were observed after administration of DFO (from 56&30 to 170+65 pg I-' which is in agreement with earlier reports on mobilization of aluminium from body deposits owing to the formation of a stable Al-DFO com- plexs*10*'6) the observed total silicon serum level was not affected by treatment with DFO.After kidney transplantation with normalization of kidney functions the physiological excretion begins and indeed total aluminium and silicon levels in serum samples have been demonstrated to decrease (from 55.4 + 4.2 before the operation to 19.9k9.5 pg1-I six months after the operation for alu- minium and from 1.30&0.60 to 0.73 k0.21 mg 1-' for silicon in sera). The relative distribution of aluminium between LMW and HMW serum fractions in the absence of DFO was found to be a constant factor:" approximately 11+2% of the total aluminium is ultrafiltrable. This partitioning does not seem to be influenced by the storage conditions of the sample the sample ageing the total serum aluminium concentration the silicon level the particular renal pathology of the patient or kidney transplantation.After administration of the DFO both the total concentration of aluminium and the relative content of the element in the ultrafiltrate increased [Fig. l(a)] mainly because of the mobilization of aluminium from body deposits4 and also because of the formation of an ultrafiltrable A1-DFO complex (which has a molecular weight of 587 Da and hence should be present in the ultrafiltrable fraction of the serum) which should be formed slowly from the Al-transferrin c o m p l e ~ . ~ ~ ~ ~ ~ ' ~ Conversely the factor that seems mainly to influence silicon partition between LMW and HMW serum fractions is sample storage conditions.When the serum samples are stored prop- erly the ultrafiltrable silicon content results are consistent and fairly reproducible (ultrafiltrable silicon 43 & 3% of total silicon serum content) in a normal spiked serum pool and also in serum from renal patients before and after the administration of DFO. However the relative silicon content in the ultrafil- trates observed decreases to approximately 20+ 5% of the total silicon concentration if samples were stored without freezing. These observations seem to support earlier sugges- t i o n ~ ~ that silicon-serum protein interactions are less specific than those for aluminium and that they are probably more physical in nature (e.g. owing to adsorption of silicic acid on proteins). Finally in the light of these results and previous experimental evidence the silicon content in serum does not seem to be related to the aluminium content both total aluminium and silicon and their corresponding distribution between the LMW and HMW serum fractions do not seem to be inter-related.As observed with various serum samples containing very different silicon levels,' the ultrafiltrable aluminium fraction remained virtually constant irrespective of the silicon concentration (in other words a change of 'ultrafiltrable aluminium' fraction ,with increasing silicon concentration by forming aluminosil- icates at the physiological pH of was not observed here). The authors thank Dr. J. B. Cannata from the Unidad de Investigacion (Hospital Central de Asturias in Oviedo) and Dr.M. D. Fernandez Gonzalez from the Hospital Universitario 'Marques de Valdecilla' Santander for helpful suggestions and for providing the clinical samples. Financial support for this research from the Comision Interministerial de Ciencia y Tecnologia (CICYT) is gratefully acknowledged. K.W. thanks the Spanish Ministry of Education and Science for the provision of a post-doctorate grant. References 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 Bertholf R. L. Wills M. R. and Savory J. in Handbook on Toxicity of Inorganic Compounds ed. Seiler H . G. Sigel H. and Sigel A. Marcel Dekker New York 1988 pp. 55-64. Bia M. J. Cooper K. Schnall S. Duffy S. Hendler E. Maluche H. and Solomon L. Kidney Int. 1989 36 852. Alfrey A. C. Legendre G.R. and Kaehny W. D. New Engl. J. Med. 1976 294 184. Sanz-Medel A. Fairman B. and Wrobel K. in Element Speciation in Bioinorganic Chemistry ed. Caroli S. Wiley New York in the press. Venturini M. and Berthon G. J. Inorg. Biochem. 1989 37 69. Birchall J . D. Chem. Br. 1990 26 141. Fahal I. H. Yaqoob M. McClelland P. Williams P. S. Roberts N. B. Birchall J. D. Ahmad R. and Bell G. M. poster presented at the 1993 Winter Conference on Plasma Spectrochemistry Granada Spain January 1993. Perez Parajon J. Blanco Gonzalez E. Cannata J. B. and Sanz- Medel A. Trends Elem. Med. 1989 6 41. Wrbbel K. Blanco Gonzalez E. and Sanz-Medel A. J . Anal. At. Spectrom. 1993 8 915. Wrobel K. Blanco Gonzalez E. and Sanz-Medel A. Trends Elem. Med. 1993 10 97. Sanz-Medel A. and Fairman B. Microchim. Acta 1992 109 7 . Sanz-Medel A. Rodriguez Roza R. Gonzalez Alonso R. Nova1 Vallina A. and Cannata J. B. J. Anal. At. Spectrom. 1987,2 177. Perez Parajon J. and Sanz-Medel A. J. Anal. At. Spectrom. 1994 9 111. Cannata J. B. Nefrologia 1992 XXII 295. Perez Parajon J. and Sanz-Medel A. Quim. Anal. 1993 12 30. Day J. P. in Aluminium and Other Trace Elements in Renal Disorders. ed. Taylor D. 1986 Bailliere Tindall London Garcia Alonso J. I. Lopez Garcia A. PCrez Parajbn J. Blanco Gonzalez E. Sanz-Medel A. and Cannata J. B. Clin. Chim. Acta 1990 189 69. pp. 184-192. Paper 31044 74B Received July 27 1993 Accepted October 21 1993