ANNALS O F CLINICAL AND LABORATORY SCIEN CE, Vol. 26, No. 3Copyright © 1996, Institu te for Clinical Science, Inc.

Involvement of Analytical Chemistry in Chemical Spéciation of Metals in Clinical Samples*RITA CORNELIS, D.Sc.

Laboratory for Analytical Chemistry, Institute for Nuclear Sciences,

University o f Gent, Gent, Belgium

ABSTRACTThe different chemical species of the trace elements in a living system

are determinants for their physiological behaviour. Their study is neces­sary to improve the understanding of trace element kinetics and metabo­lism. In a complex matrix, such as biological fluids and tissues, some trace elements will occur as free or mononuclear ions; others as low molecular weight complexes, as reversible or irreversible macromolecular complexes. Spéciation investigations entail the separation of the compounds, followed by the measurement of the trace element in the different fractions. Frame- work-procedures are outlined and attention is drawn on the many difficul­ties that can be encountered. These include the complexity of the matrix, insufficient specificity of the separation of biocompounds, fortuitous con­taminations with trace elements, and cutting the original metal-protein binding. State of the art description is given for the spéciation studies of Al, As, Cd, Cr, Co, Cu, Hg, Ni, Pb, Pt, Se, Sn, and Zn.

IntroductionThe approach to “spéciation,” outlined

in this paper, comes from the field of ana­lytical chemistry. The objective is to measure trace element species in relation to physiological and metabolic processes associated with health and disease. In fact, much of the research done in other disciplines belongs to the domain of spé­ciation, although it will never be classi­

* Send reprint requests to: Rita Cornelis, D.Sc., Laboratory for Analytical Chemistry, Institute for Nuclear Sciences, University of Gent, Proefuin- straat 86, B-9000 Gent, Belgium.

fied under such heading. It seems that “spéciation” is a word introduced by the analytical chemists at home in total ele­ment determinations. As a new line of research was started, there had to be a way to mark it off from past activities. First thing to do is to give it a new name. So at some time, someone, or some group, “borrowed” the name “spécia­tion” from the biologists, or simply derived it from “species,” a term already currently used by nuclear and configura­tion chemists. At any rate, in the 1980s research on spéciation became fashion­able, and in the 1990s it is all over the literature.

2520091-7370/96/0300-0252 $01.80 © Institute for Clinical Science, Inc.

ANALYTICAL CHEMISTRY IN CHEMICAL SPECIATION OF METALS 2 5 3

Need for Speciation WorkKnowledge about the biochem ical

behaviour of the trace and, especially, the ultra-trace elements is blossoming. This is spurred not only by a scientific appetite for basic information, but also by pressing occupational, environmental, and economic needs to come up with explanations about mobility, storage, retention, and toxicity. It is a relatively new scientific field which is more at home in toxicology as it was there that the relationship between the toxicological effects and the chemical forms of the trace metals first was recognized. Many metals can form small organometallic compounds. For some of these, the toxic effects exceed those of the inorganic forms of the elements or of the com­pounds formed with large molecules. Methylmercury, triethyltin, and tetraeth­yllead are examples of this.

As m etal sp ec ies becom e better known, total element determinations of trace elements appear less relevant. In the past, this was the case for iron and iodine. Iron is perhaps the most widely studied element with the iron-binding proteins now being the prime clinically relevant parameters. The same is true for iodine, where the measurement has been replaced by that of the concentrations of the thyroid hormones T3 and T4 and other related hormonal compounds. Both iron and iodine provide good examples of elements whose role and pathways are now so well understood that the century- long interest in their total concentration has been completely superseded by that of components which need the elements for their biosynthesis and physiologi­cal functions.

A still different development is that of the biomarkers of exposure. These are measurements that reflect an event in a biological system, such as a human body. This may be exemplified with the meth­ylated As-species for measuring the expo­

sure to inorganic As. Another example is cadmium. In urine, Cd mainly reflects the amount stored in the body. Because most of the Cd in urine is probably bound to metallothionein, the changes in uri­nary metallothionein concentration par­allel those of Cd.D e f i n i t i o n o f Sp e c i a t i o n a n d D e l i m i t a t i o n o f t h e I n v o l v e m e n t o f T r a c e E l e m e n t A n a l y t ic a l C h e m i s t r y

As the usage of the terms “species” and “speciation” become more widespread, a general definition is needed. At the sec­ond speciation meeting in Loen in 1994, the following definition was embraced1: “Speciation is the occurrence of an ele­ment in separate identifiable forms, ie, chemical, physical or morphologi­cal state.”

Only the particular approach of chemi­cal speciation will be considered. It cov­ers the qualitative and quantitative mea­surements in body fluids and tissues of (1) organometallic compounds and (2) biologically active compounds to which the trace element is bound, including the quantification of the element in relation to those particular molecules. These investigations require first the separation of the biomolecules, followed by the determination of the trace element in the different fractions. It is assumed that the biom olecule and the trace elem ent, which are detected in the same fraction, are associated with one another. This is an indirect way to measure species, dif­ferent from the direct measurement of a specific property such as Co in vitamin B12. Its 3-dimensional structure was explained by X-ray diffraction crystallog­raphy by Dorothy Hodgkin in 1956 (Nobel Prize in 1964).

The speciation of unstable metal com­plexes by a direct experimental tech­nique is a difficult problem as the experi­mental method must not disturb the

2 5 4 CORNELIS

equilibria established in solution.2 It is unfortunate that the few direct methods which may prove useful, such as nuclear magnetic resonance and Raman spectros­copy, lack the sensitivity necessary to probe samples containing only trace lev­els of metal complex species. A similar shortcoming limits the possibilities of all so-called indirect speciation studies. As a main objective consists of the measure­ment of the trace elements, many bioac­tive compounds will escape detection because the amount of trace element they contain lies far below the detection limit of current analytical methods. Take again the example of Co in vitamin B12. As it is still a challenge to measure accurately the concentration of Co in serum at about 0.2 |uug/l, it is not feasible to measure the share of Co attributable to vitamin B12 at0.03 jxg/1.

Another example is the large family of metalloproteases. Generally, they also escape detection by the kind of proce­dures outlined in this paper because the trace element concentration, for instance that of Zn, lies below the detection limit of conventional methods.

In all those cases, measuring directly the property of a particular compound in the sam ple is much easier, eg, the enzym e activities of the sp eci­fic metalloenzymes.

Considering the limitations of such a working strategy, it looks as if only the link between the trace element and its main “bio-partners” can be defined. Nev­ertheless, it is an interesting field of research which collects useful informa­tion about some dominant species of an element in the body, and it surely is a valuable tool for kinetic and meta­bolic studies.

Classification of Trace Element Species in Clinical Samples

Elements documented in relation to their speciation in biological fluids and

tissues are: aluminum, arsenic, cadmium, chromium, copper, cobalt, mercury, n ick e l, lead , p latinum , se len iu m , and zinc.

The species can be categorized into five groups:1. Small organom etallic m olecu les,

either contaminants of food, water or air; overall they remain unchanged in the body. Examples: organotin orga- nolead compounds, methylmercury,

2. Biomarkers of exposure: detoxifying mechanisms change the species of the trace element from the inorganic to small organometallic compounds: eg, arsenite and arsenate are metabolized to monomethylarsonic acid (MMA) and dimethylarsinic acid (DMA).

3. Elements with different valence states that affect their biological action; eg, Cr(III)/(VI), Fe2+/Fe3 + .

4. The trace elem ent is an essential building block of the biomolecule: eg, Cu in ceruloplasmin, Co in cobala- min, or vitamin B12, metalloenzymes.

5. The trace elem ent forms a metal- ligand complex with a low molecular weight compound such as Al with cit­rate, or with a macromolecular bio­molecule; eg, Al and Cr to transferrin and albumin; arsenate to transferrin and hemoglobin; Zn to albumin and a2-macroglobulin. Their mechanism is governed by thermodynamic stability constants and by rates of formation and dissociation.

Group 5 also includes the trace ele­ments bound to unidentified proteins (at best some knowledge is available about the MM of the protein).

Within this scheme, three different lines of analytical procedures can be dis­cerned according to whether the specia­tion concerns:A. Small organometallic compounds, as

is the case for categories 1 & 2;

ANALYTICAL CHEMISTRY IN CHEMICAL SPECIATION OF METALS 2 5 5

B. Valence state studies, with element specific procedures (category 3); and

C. Complexes of the trace element with small and large biomolecules (cate­gory 5).

In the case of category 4, the com­pound is directly measured with specific biochemical techniques. Measurement of the trace element become redundant.

Separations of type A schematically consist of the separation of the species, usually with high performance liquid chromatography (HPLC), the retention time is characteristic for each species; as a rule the species are separated with very high resolution and elute as single com­pounds; the eluent is analyzed by its trace element detection system and so allows quantification of the species through comparison with calibrants. The spéciation of As-species in urine (figure 1) illustrates this. It shows chromato­graphic separation of arsenic metabolites

As(III), dimethylarsinic acid (DMA), monomethylarsonic acid (MMA) and As(V) achieved by Hakala and Puy.3

In the case of type B, separation of dif­ferent valence states, the procedures will be fashioned according to the element and the matrix.

Type C consists of separation of the biomolecules for trace element determi­nation purposes. Generally spoken the resolution of separation is very poor. Illustrated in figure 2 are the speciation of Cr-species in plasma proteins. Each fraction may still contain tens of different molecules of which only a few do carry the trace element. In the case of chro­matographic separations, the ultraviolet (UV) detection of the biomolecule in the eluent is not specific. Therefore, each fraction must be split into a larger portion for trace element measurement and a small portion for the specific measure­ment of the biomolecule that presumably carries the trace element.4,5

9)UC(0

X3w.oV)-Q<

F i g u r e 1. Chromato­g raph ic se p a ra tio n of arsenic m etabolites (50 jjLg/1 of As). Injection vol­ume 50 |xl. HPLC on 2 Chrom -Sep colum ns in series packed with 5 (Jim C h r o m S p h e r C 18 reversed phase, followed by hydride genera tion atomic absorption spec­trometry. (Hakala E, Pyy L. J A n a l A to m Spec 1992;7:191-6. Permission to reprint figure granted by JM Gordon, editor of J A na l A to m Spec , T he Royal Society of Chemis­try, Cambridge, UK.)

Time/min

2 5 6 CORNELIS

Elution volume (ml)F ig u r e 2. Protein and 51Cr profiles of plasma of CAPD patient after instillation of olCr spiked perito­

neal dialysate (anion exchange chromatography on a MonoQ HR 16/10, eluent 0.05 Mol/1 NaCl in 0.025 Mol/1 Tris-HCl (pH 8) of the pooled fractions of unbound proteins from the cation exchange MonoS HR 16/10 in 0.025 Mol/1 Tris-HCl, pH 8). (Borguet F , Cornelis R, Delanghe J, Lambert MC, Lameire N. Clin Chim Acta 1995;2328:71-84. Permission to reprint figure granted by DW Moss, editor of Clin Chim Acta, London, UK.)

Frameworks to Identify and Measure Trace Element Species

The general scheme to do this kind of research consists of the two main steps: (1) isolation and identification of the spe­cies, and (2) quantification of the species.

A large array of techniques is at hand for the separation of species: size exclusion chromatography (SEC), ion exchange chromatography and affinity chromatog­raphy applied in a system of fast protein liquid chromatography (FPLC) or HPLC, electrophoresis, capillary electropho­

ANALYTICAL CHEMISTRY IN CHEMICAL SPECIATION OF METALS 2 5 7resis, dialysis, ultrafiltration, ultracen­trifugation, liquid-liquid extraction, electrod ep osition on H g-electrode, supercritical fluid extraction, etc.

In the development stage of doing separations of biomolecules it is wise to have them based on two different prop­erties: eg, molecular mass using solvent exclusion chromatography (SEC) and electric charge using ion exchange chro­matography. UV on-line detection of the proteins is not very specific and for qual­itative surveys this method has to be complemented with eg, electrophoretic measurements. For precise quantitative measurements of proteins, nephelome- try is one reliable possibility. Other methods are radial immunodiffusion, kinetic immunoturbidity.

Detection of the trace element is feasi­ble by either of the following techniques: flame atomic absorption spectropho­tometry (FAAS), hydride generation atomic absorption spectrophotometry (HGAAS), cold vapor atomic absorption spectrophotometry (CVAAS), electrother­mal atomic absorption spectrophotome­try (ETAAS), inductively coupled plasma optical emission spectroscopy (ICPOES), inductively coupled plasma mass spec­troscopy (ICPMS), magnetically induc­tive plasma optical emission spectros­copy (MIPOES), and direct current plasma optical emission spectroscopy (DCPOES). Ideally, the separation and detection techniques should be hyphen­ated. Sometimes radiotracer experiments are feasible, follow ed by gamma or beta measurements.

A Survey of Existing Knowledge about Trace Element Species

An overview of information on trace element species is included, but it is not an elaborate review of the existing litera­ture. Many valuable contributions in this field may not have been cited in this paper. Details about procedures used to

obtain “speciation-knowledge” are not specified. The reason is that they are variations on a similar theme: a particular type of colum n, a different buffer, another ingenious coupling system, a better sensitivity for detection of the trace element, a particular procedure of electrophoretic separations, etc. These details, although most relevant, would be outside the scope o f the p res­ent overview.A l u m i n u m

Speciation is basic in the understand­ing of its role in the etiology of a variety of neurological and skeletal disorders in man. The behavior of Al3+ species in bio­log ica l flu id s and c e lls has b een described by four different ligand-states or forms6: as free or mononuclear ions, low molecular weight complexes, revers­ible macromolecular complexes, and irre­versible macromolecular complexes.

A major group is the binding of Al3+ to low m olecular w eigh t com pounds, mainly citrate and polyphosphate spe­cies.6 The citrate complex provides the means by which Al3+ may pass through membranes; the polyphosphates may bind Al3+ in tissues. Reversible macro­molecular binding occurs when Al binds with transferrin. Based on stability con­stants, transferrin may be the ultimate carrier of Al3+ in plasma. As irreversible macromolecular complexes, the poly­phosphates in chromatid DNA have been postulated.

Most interesting studies are published on citrate and transferrin as the main com plexing agents for Al in serum. Ohman and Martin7 have recently sug­gested (based on stability constants) that of the Al3+ in human serum, —89 percent (±5 percent) binds to transferrin and ~11 percent (±5 percent) to citrate. In their theory, they also leave room for small amounts of other ligands.

Serum of dialysis patients has been investigated with SEC and ultrafiltration

2 5 8 CORNELIS

by several research centers. One of the first studies on Al speciation in serum with SEC was done in the group of J. Savory at the beginning of the 1980s.8 Recently, very interesting analytical work has been published by D ’Haese and coworkers9 on speciation of Al in serum, using HPLC-ETAAS. A major merit is their achieving contamination free analytical procedures and quantita­tive recovery of the compounds at clini­cally relevant concentrations. This is a very challenging task, considering that the total concentration of serum alumi­num in healthy individuals is about 1.3 to1.6 |xg/l.10 Values are much higher in dialysis patients, up to 30 jxg/1 sometimes 100 ug/1.10

A r s e n i c

For arsenic, most of the studies deal with arsenic speciation in urine to evalu­ate occupational and environmental exposure.3,11,12,13,14 The species are arse- nite (AsO3 - ), arsenate (AsO|~), MMA monomethylarsonic acid (MMA), dimeth- ylarsinic acid (DMA), arsenobetain (AsB), and arsenocholine (AsC). They are presently measured in urine. Whereas AsB and AsC are considered nontoxic species absorbed from the diet, mainly from seafood, the presence of arsenite, arsenate, MMA, and DMA are evidence for exposure to toxic arsenite and arsen­ate from food, water or air. Both MMA and DMA are the relevant biomarkers of exposure. Measurement of total As in irrelevant. The latter value amounts to about 10 (Jig/1.10

Speciation studies in serum and packed cells are more difficult because of very low concentrations in healthy, unex­posed individuals. Values of total arsenic in healthy individuals vary between 1 and 5 jxg/1 serum. Similar concentrations are found in packed cells.10

A low cost, on-line procedure for the measurement of the toxic inorganic,

MMA, DMA species and the non toxic species, AsB, AsC, in serum of uraemic patients was recently published by Zhang et al.15

Besides the presence of the As-com- pounds mentioned, in vivo radio tracer studies of rabbits with 74As-labelled arsenate revealed the binding of As with transferrin. In their packed cells arsenate is bound to haemoglobin.5,16

C a d m iu m

The element cadmium has been found to form ligands with proteins of many dif­ferent molecular masses. The group of the metallothionein is without any doubt the best documented .17,18 This topic appears mainly focused on biomarkers of effect of cadmium, especially in relation to its nephrotoxicity, and does not include measurements of the Cd concen­trations in the different markers.19

There are, however, some exceptions. Carson describes the binding of Cd to human a 2-m acroglobu lin ,20 and Michalke et al21 report on the speciation of Cd in human breast milk where it was measured in the fraction corresponding to metallothionein.

c h r o m iu m

The valence state of Cr in the body is Cr(III). There is so much reducing capacity in saliva and in the stomach that any normal intake of Cr(VI) from food and drinks will be readily reduced to Cr(III).22,23 When Cr(VI) is inhaled, it is transported into erythrocytes as chromate and easily crosses the cell membranes. In the cell, it is reduced to Cr(III), and it is this reduction (or perhaps intermediate valence states) that causes damage to deoxyribonucleic acid (DNA).

In plasma or serum, Cr is present at about 0.1 to 0.2 (xg/1,10 and is bound to

ANALYTICAL CHEMISTRY IN CHEMICAL SPECIATION OF METALS 2 5 9large biomolecules, mainly to transferrin and to albumin. In figure 2 is shown a chromatogram of plasma proteins with the concomitant Cr-elution profile. In the case of serum from dialysis patients, a shift was observed from albumin to pre­albumin. An attempt to identify this com­pound revealed the binding of Cr to a yet unidentified molecule with a MM of about 5 000 D .24

In urine Cr is present as a Cr(III) com­pound, probably bound to small ligands.*C o b a l t

Besides the fact that cobalt is present in the body partly as cobalamines, no fur­ther know ledge about other species seems available for the time being.C o p p e r

The best documented compound of copper is ceruloplasmin, an a2-glycopro- tein of MM 132 000 D, the main Cu-con- stituent in serum.25,26 Wirth and Lind- ler27 found an additional Cu transport protein in plasma, named transcuprein (MM 270 000 D). There also occurs some binding to albumin.28 More difficult to fit into the scope of trace element measure­ment procedures are the many Cu-en- zymes (cytochrome c oxidase, lysyl oxi­dase, Cu-Zn dismutase). The copper concentration in serum or plasma of healthy individuals ranges from 0.8 to 1.4 mg/1.10 In urine, the mean concentration ranges from 15 to 36 fig/day.10 Copper is present as chloride or chlorocomplexes.29 In tissue, copper binding to liver metal- lothionein is reported in the literature.30

M e r c u r y

The main item about mercury is to dis­tinguish the inorganic from the short-

* Cornelis R. Unpublished results.

chain alkyl (mainly methylmercury) spe­cies. The method based on the so-called Magos principle is probably the best- known and most accessible.31 Samples are treated with L-cysteine and trichloro­acetic acid. The protein-free supernatant is treated with SnCl2 and CdCl2 for the determination of total Hg or with SnCl2 for selective reduction of inorganic vapor. The Hg° is next m easured w ith co ld vapor a to m ic a b so r p ­tion spectrometry.

Up to date research uses chromato­graphic separations, after extraction of the species, and coupling to AAS or AES systems. The species of interest are Hg2*, (CH3)Hg+, and (C2H5)Hg+. The matrixes analyzed are blood, urine, and hair.32,33 The total Hg concentration in serum of healthy individuals amounts to0.5 fjLg/1; in packed cells, the level is about 5 |xg/kg. In urine of non-exposed individuals, the Hg level lies at the 1 to 10 (xg/1.10

N ic k e l

In 1988, Nomoto and Sunderman34 defined some Ni-components in human serum. They found an association with a2-macroglobulin, albumin and ultra fil­terable constituents, such as amino acids, binding Ni. Similarly, N ielsen et al showed that nickel-binding occurs with albumin, a2-macroglobulin and other proteins such as prealbumin, c^-anti- trypsin, o^-lipoprotein.35 Reference val­ues for Ni in serum are <0.3 jxg/1.10

Very interesting findings on nickel- binding proteins are reported using a protein blotting method with radioactive 63Ni tracer.36,37

L e a d

The analyzed lead species are mainly organic alkyllead, contaminants from anthropogenic origin: R4Pb, R3PbX, R2PbX2; R = CH3, C2H5; X = anion. These

2 6 0 CORNELIS

compounds have not yet been widely investigated in clinical samples. This may be due not only to the lim ited amount of material typically available for this type of study, but also to the analyti­cal difficulties related to this type of spe- ciation work.38 The concentrations in clinical samples are also very low. For instance, in urine of workers who are occupationally exposed, trimethyllead and triethyllead lie below the detection limits of 0.43 and 0.45 jxg/1, respectively.P l a t in u m

The in terest paid to platinum is focused on the derivatives of cisplatin, the cancer treatment drug.39,40 Experi­ments with rats (in vivo) and mice (in vitro) revealed the binding of Pt anti­tumour compounds to plasma proteins with high and low molecular weight. Over time and for certain compounds, transfer occurs to the low MM proteins.40

Se l e n i u m

Selenium may be the hottest topic in speciation world at present. Among the many research teams working on this topic are the groups of Whanger and co-workers at the Oregon State Univer­sity ,41 Bratter and co-workers,42 and Behne and co-workers, both at the Hahn- Meitner Institut Berlin,43 and Akesson at the University of Lund, Sweden.44 The list is not complete by far. The previously mentioned authors describe how Se in human serum is associated with glutathi­one peroxidase, selenoprotein P, and concanavalin .44,45,46,47 The serum or plasma-Se concentrations in adults vary between 0.04 and 0.16 mg/1 depending on the Se intake from food and bever­ages.10 Total Se is not indicative for the Se status of an individual. (If so, residents of New Zealand who have 48 |J.g/L Se in serum would all suffer from Se-defi- ciency, as they are at about half of the

value in Western Europe.) Defining the clinically essential Se-species would, therefore, be very interesting. It is likely that m easurem ent of selen ium may become superfluous, and that selenopro­tein P, Se in glutathione peroxidase, or another compound will become the func­tional indicators for the Se-status in man.

Interesting results were obtained by Ducros et al48 on sera of patients who received stable isotope 74Se-selenite sup­plementation. Besides glutathione per­oxidase, two Se-peaks were noted after SEC in unequal proportions over a period of 24 h. No labelling of red blood cells was found.

In the erythrocytes, Se is mainly asso­ciated with hemoglobin and for a small amount with glutationperoxidase; Se is present in packed cells at the same con­centration as in serum.

The species distinguished in urine are selenomethionine, trimethylselenonium ion (TMSe + ) [detoxification metabolite in urine], selen ite S eO f- , selenate S e O f- , se len o a m in o a cid s, se len o choline, proteins with different MM.49,50 The Se concentration in urine is at about 100 |a,g/l, but it is very dependent on the Se intake.10

As for tissue, very interesting studies have been published by Behne et al43 on the distribution and characteristics of new-mammalian Se-containing proteins. The results were obtained with electro­phoretic separations and autoradiography on experimental animal proteins.

T in

The analyzed tin species are mainly organic Sn contaminants from anthropo­genic origin, such as trimethyl, tributyl- and triphenyltinchloride. Up to now, studies have been dealing with environ­mental problems.51 Considering that the total concentration of tin in the serum of unexposed individuals is < |xg/l,52 it

ANALYTICAL CHEMISTRY IN CHEMICAL SPECIATION OF METALS 2 6 1will take some time before these con­ta m in a n ts can b e m e a su r e d in human samples.Z in c

For zinc there is binding to albumin, a2-macro-globulin, dissolved species, and proteins with different MM.26,53,54 More difficult, if not impossible, to assess with trace element measurements are the ~300 enzymes that require Zn to modu­late their activities and the Zn-finger pro­teins and their involvement in genetic expression. The serum or plasma total Zn concentrations amount to 1 mg/1.10 Zinc in packed cells is one order of magnitude higher10 and in urine is present as chlo­rides or chlorocomplexes.29 The total concentration in urine may vary between 50 to 1,000 (xg/day, depending on the Zn intake.10

Experimental Problems Encountered in Trace Element Spéciation Work

The many factors that are liable to influence chemical spéciation studies of trace e lem ents in c lin ica l samples have been very thoroughly reviewed by G ardiner .55,56 A long the same line of thoughts, other researchers have described their own experimen­tal problems.4,5,57

Many separation techniques do not meet the requirements for trace element determinations, and they all pose huge hazards of contamination and loss of the trace element. Developments in biotech­nology are focusing on systems that pro­cess very small quantities of biocom­pounds, offering increasing resolution and specificity. As the proteins carry just one or a few atoms per molecule, the diluted fractions of the isolated proteins more often than not contain trace ele­ment quantities far below the detection limit of the analytical methods. There­fore, separation of proteins for spéciation

studies has to be done on a preparative scale with less resolution and question­able specificity. To improve the detec­tion sensitivity of the trace elements, 100 times enrichment procedures in the frac­tions have to be developed, or new, more sensitive detection should become avail­able. High resolution ICP-MS is the new­comer on the horizon.

Contamination hazards and insuffi­cient detection sensitivity for the ele­ment are the two major obstacles in spé­ciation work. They can be partially remedied by doing feasibility studies with in vivo and in vitro labelled radio­active compounds. Radiotracers virtually eliminate the immense hazard of distort­ing the results with exogenous amounts of the trace element. Because they are not radioactive, they escape detection. This eases to a great extent the determi­nation of very low concentrations in the various fractions. Another interesting line is the in vivo application of enriched stable isotopes, with mass spectrometry as the trace element detection method.48

In the case of supplementation studies, it is to be expected that the increases in tissue concentration observed at high trace element intakes w ill be mainly caused by non-specific incorporation of the element into many proteins. High accuracy is evidently requested for all quantitative measurements. This can only be assured by the availability of rep­resentative reference materials.58 As spé­ciation results are only just beginning, it will be a long time before universally accepted certified reference materials will become available.Conclusion

Analytical chemistry is most relevant in the development and application of trace element spéciation to toxicological, environmental, and occupational expo­sure issues. Trace element spéciation already provided useful insights in the toxicological, m etabolic and kinetic

2 6 2 CORNELIS

understanding of different trace element species. It can be postulated that gradu­ally, as the understanding about trace element species is growing, these will become the relevant parameters for clini­cal decisions.References

1. Thomassen Y, Nieboer E. Foreword. The sec­ond international symposium on speciation of elements in toxicology and in environmental and biological sciences. Analyst 1995;120:30N.

2. Fernando Q. Metal speciation in environmental and biological systems. Environm Hlth Per­spectives 1995;103:13-6.

3. Hakala E, Pyy L. Selective determination of toxicologically im portant arsenic species in urine by high-performance liquid chromatogra- phy-hydride generation atomic absorption spectrometry. J Anal Atom Spec 1992;7:191-6.

4. Cornelis R, Borguet F, De Kimpe J. Trace ele­ments in medicine. Speciation: the new fron­tier. Anal Chim Acta 1993;283:183-9.

5. Cornelis R, De Kimpe J. Elemental speciation in biological fluids. J Anal Atom Spectr 1994;9: 945-50.

6 . Macdonald TL, Martin RB. Aluminium in bio­logical systems. Trends Bioch Sci 1988;13: 15-9.

7. Ohman LO, Martin RB. Citrate as the main small molecule binding Al3+ in serum. Clin Chem 1994;40:598-601.

8 . King SW, Savory J, Wills MR. Aluminium dis­tribution in serum following hemodialysis. Ann Clin Lab Sc 1982;12:143-9.

9. Van Landeghem GF, D ’Haese PC, Lamberts LV, De Broe ME. Quantitative HPLC/ETAAS hybrid method with an on-line metal scavenger for studying the protein binding and speciation of aluminum and iron. Anal Chem 1994;66: 216-22.

10. Cornelis R, Heinzow B, Herber RFM, Molin Christensen J, Poulsen OM, Sabbioni E, Tem­pleton DM, Thomassen Y, Vahter M, Vester- berg O. Sample collection guidelines for trace elements in blood and urine. Pure Appl Chem 1995;67:1575-608.

11. Vahter M. What are the chemical forms of arse­nic in urine, and what can they tell us about exposure? Clin Chem 1994;40:679-80.

12. Vahter M, Lind B. Concentrations of arsenic in urine of the general population in Sweden. Sci Total Environm 1986;54:1-12.

13. Miirer AJ, Abildtrup A, Poulsen OM, Chris­tensen JM. Effect of seafood consumption on the urinary level of total hydride-generating arsenic compounds. Instability of arsenobe- taine and arsenocholine. Analyst 1992; 117: 677-80.

14. Larsen H, Pritzl G, Hansen SH. Speciation of eight arsenic compounds in human urine by high-performance liquid chromatography with

inductively coupled plasma mass spectrometric detection using antimonate for internal chro­matographic standardization. J Anal Atom Spec 1993;8:557-63.

15. Zhang X, Cornelis R, De Kimpe J, Mees L. Arsenic speciation in serum of uraemic patients based on HPLC-HGAAS w ith on-line UV photo-oxidation digestion. Anal Chim Acta 1996;319:177-85.

16. De Kimpe J, Cornelis R, Mees L, Vanholder R. Binding, biotransformation and tissue distribu­tion of ip-injected carrier-free 74As-labelled arsenate in rabbits. Submitted to Fund Appl Toxicol, 1996.

17. Kági JHR, Kojima Y, eds. Metallothionein. 2nd In ternational M eeting on M etallothionein, University of Zurich, 1985.

18. Lauwerys RR, Bernard AM, Roels A, Buchet JP. Cadmium: exposure markers as predictors of nephrotoxic effects. C lin C hem 1994;40: 1391-4.

19. Michalke B. Speciation of cadmium and zinc in human breast milk samples. In: Brátter P, Ribas B, Schramel P, editors. Trace Element Analyti­cal Chemistry in Medicine and Biology, vol. 6 . Madrid: Consejo Superior de Investigaciones. 1994;397—409.

20. Carson SD. Cadm ium b ind ing to hum an a 2-macroglobulin. Biochim Biophys Acta 1984; 791:370-4.

21. Nordberg M, Trojanowska B, Nordberg GF. S tudies on m etal-b inding pro teins of low molecular weight from renal tissue of rabbits exposed to cadmium and mercury. Environ Physiol Biochem 1974;4:149-58.

22. De Flora S, Serra D, Camoirano A, Zanacchi P. Metabolic reduction of chromium, as related to its carcinogenic properties. Biol Trace Element Res 1989;21:179-87.

23. De Flora S, Badolati GS, Serra D, Picciotto A, Magnolia MR, Savarino V. Circardian reduction of chromium in the gastric environment. Muta­tion Res 1987;192:169-74.

24. Borguet F, Cornelis R, Delanghe J, Lambert MC, Lameire N. Study of the chromium bind­ing in plasma of patients on continuous ambu­latory peritoneal dialysis. Clin Chim Acta 1995; 2328:71-84.

25. Percival SS, Harris ED. Copper transport from ceruloplasmin: characterization of the cellular uptake mechanism. Am J Physiol 1990;258: C140-6.

26. Chilvers DC, Dawson JB, Bahreyni-Toosi MH, Hodgkinson A. Identification and determina­tion of copper- and zinc-protein complexes in blood plasma after chromatographic separation on DEAE-Sepharose CL-6B. Analyst 1984;109: 871-6.

27. Wirth PL, Linder MC. Distribution of copper among components of human serum. J Natl Cancer Inst 1985;75:277-84.

28. Masuoka J, Saltman P. Zinc(II) and copper(II) binding to serum albumin. J Biol Chem 1994; 269:2557-61.

29. Weber G. Investigation of metal-species (Ca,

ANALYTICAL CHEMISTRY IN CHEMICAL SPECIATION OF METALS 2 6 3Mg, Zn, Fe, Cu, Pb, Cd, Sn) in urine by HPLC- AAS. J Trace Elem Electrolytes Health Dis 1988;2:61-5.

30. Presta A, Green AR, Zelazowski A, Stillman MJ. Copper binding to rabbit liver metallothionein. Formation of a continuum of copper(I)-thiolate stoichiometric species. Eur J Biochem 1995; 227:226-40.

31. Magos L. Atomic absorption determination of total, inorganic and organic mercury in blood. J Assoc Off Anal Chem 1972;55:966-71.

32. Brunmark P, Skarping G, Schütz A. Determina­tion of methylmercury in human blood using capillary gas chromatography and selected ion- monitoring. J Chromatogr 1992;573:35-41.

33. Bulska E, Emteborg H, Baxter DC, Freeh W. Spéciation of mercury in human whole blood by capillary gas chromatography with a micro­wave-induced plasma emission detector system following complexometric extraction and buty­lation. Analyst 1992;117:657-63.

34. Nomoto S, Sunderm an FW Jr. Presence of nickel in alpha-2 macroglobulin isolated from human serum by high performance liquid chro­matography. Ann Clin Lab Sc 1988;18:78-84.

35. Nielsen JL, Poulsen OM, Abildtrup A. Studies of serum protein complexes with nickel using crossed immunoelectrophoresis. Electrophore­sis 1994;15:666-71.

36. Lin SM, Hopfer SM, Brennan SM, Sunderman FW Jr. Protein blotting method for detection of n ickel-binding proteins. Res Comm Chem Pathol Pharmacol 1989;65:275-88.

37. Haspel J, Sunderman FW Jr, Hofper SM, Hen- jum DC, Brandt-Rauf PW, Weinstein IB, Nish- imura S, Yamaizumi Z, Pincus MR. Res Comm Chem Pathol Pharmacol 1993;79:131^40.

38. Baxter DC, Freeh W. Spéciation of lead in envi­ronmental and biological samples. Pure Appl Chem 1995;67:615-48.

39. Zhao Z, Tepperman K, Dorsey JG, Elder RC. Determination of cisplatin and some possible m etabolites by ion-pairing chromatography with inductively coupled plasma mass spectro- metric detection. J Chrom Biomed Appl 1993; 615:83-9.

40. Perera A, Jackson H, Sharma HL, Charles A, McAuliffe CA, Fox BW. A comparative binding of platinum anti-tumour compounds to plasma proteins in the rat (in vivo) and mouse (in vitro), Chem Biol Interactions 1992;85:199- 213.

41. W hanger PD, Xia Y, Thomson CD. Protein technics for selenium spéciation in human body fluids. J Trace Elem Electrolytes Health Dis 1994;8:1-7.

42. Bràtter P, Gercken B, Tomiak A, Rôsick U. Combination of HPLC and ICP-AES for the spéciation of selenium and other trace ele­ments in body fluids. In: Bràtter P, Schramel P, editors. Trace elem ent analytical chemistry in m edicine and b io logy, vol 5. B erlin: de Gruyter, 1988:119-35.

43. Behne D, Weiss-Nowak C, Kalcklôsch M, West- phal C, Gessner H, Kyriakopoulos A. Studies

on the distribution and characteristics of new mammalian selenium-containing proteins. Ana­lyst 1995;120:823-5.

44. Huang W, Akesson B, Svensson BG, Schütz A, Burk RF, Skerfving S. Selenoprotein P and glu­tathione peroxidase (EC 1.11.1.9) in plasma as indices of selenium status in relation to the intake of fish. Brit J Nutr 1995;73:455-61.

45. Hasegawa T, Mihara M, Nakamuro K, Sayato Y. Distribution and chemical form of selenium in mice after administration of selenocystine. Biol Pharm Bull 1994;17:1215-9.

46. Daher R, Van Lente F. Concanavalin A-bound selenoprotein in human serum analyzed by graphite furnace atomic absorption spectrome­try. Clin Chem 1994;40:62-70.

47. Deagan JT. Determination of the distribution of selenium betw een glutathione peroxidase, selenoprotein P, and albumin in plasma. Anal Biochem 1993;208:176-81.

48. Ducros V, Richard MJ, Favier A. The distribu­tion of selenium in human plasma proteins for 24 hours after ingestion of 74Se (in sodium sele­nite form). J Inorg Biochem 1994;55:157-63.

49. Hasunuma R. Analysis of selenium metabolites in human urine using ion exchange chromatog­raphy. Bull Environ Contam Toxicol 1993;50: 19-23.

50. Hasunuma R. Selenium metabolite levels in human urine after dosing selenium in different chemical forms. Bull Environ Contam Toxicol 1993;51:756-63.

51. Kuballa J, Wilken RD, Jantzen E, Kwan KK, Chau YK. Spéciation and genotoxicity of butyl compounds. Analyst 1995;120:667-73.

52. Versieck J, Cornelis R. Trace elem ents in human plasma or serum. Boca Raton: CRC Press Inc, 1989.

53. Gardiner PE, Ottaway JM, Fell GS, Burns RR. The application of gel filtration and electrother­mal atomic absorption spectrometry to the spé­ciation of protein-binding zinc and copper in human blood serum. Anal Chim Acta 1981;124: 281-94.

54. Franke U. Speziation von Zink in Serum mit Hilfe chromatographisher Methoden. Metho- denen tw ick lung und k lin isch e S tud ie an Hàmodialysepatienten. Aachen: Verlag Shaker, 1994.

55. Gardiner PE. Considerations in the preparation of biological and environm ental reference materials for use in the study of the chemical spéciation of trace elements. Fres J Anal Chem 1993;345:287-90.

56. Gardiner PHE. Factors that influence the dis­tribution of trace element-containing chemical species in living systems before and after sam­ple collection. J Trace Elem Electroly tes Health Dis 1993;7:1-8.

57. Behne D. Spéciation of trace elements in bio­logical materials: trends and problems. Analyst 1992;197:555-7.

58. Quevauviller P. Certified reference materials for specific chemical forms of elements. Analyst 1995;120:597-602.