WO2010123448A1

WO2010123448A1 - Method for reducing adsorption losses on metal oxide surfaces

Info

Publication number: WO2010123448A1
Application number: PCT/SE2010/050434
Authority: WO
Inventors: Mats INGANÄS; Johan ENGSTRÖM; Therese Eriksson; Tina Ekenkrantz
Original assignee: Gyros Patent Ab
Priority date: 2009-04-21
Filing date: 2010-04-21
Publication date: 2010-10-28

Abstract

The invention is related to a method for reducing adsorption losses of analytes during sample handling where the sample comes into contact with a metal oxide surface. The invention is further related to a device for sample handling comprising a channel, wherein the channel has a metal oxide surface which comes in contact with samples handled with the device. The invention is further related to a kit comprising a device for sample handling comprising a channel, wherein the channel has a metal oxide surface which comes in contact with the samples handled with the device, and a reagent.

Description

Method for coating a metal oxide surface

Field of the Invention

The present invention is related to a method for reducing adsorption losses of analytes during sample handling. In another aspect the present invention is related to a device for sample handling. In a further aspect the present invention is related to a kit comprising a device for sample handling comprising a channel, wherein the channel has a metal oxide surface which comes in contact with the samples handled with the device, and a reagent.

Background to the Invention

Within the field of analytical chemistry there is an increasing demand for lower detection limits. There are in principle a number of different ways to improve detection limits. One possibility is to increase measurement sensitivity, for example by increasing detector sensitivity. A further possibility is to reduce the measurement noise. Another possibility is to concentrate the analytes in the sample by the use of more efficient sample preparation procedures, for example including the use of chromatography or extraction techniques.

Automated systems for analysing samples most often include sample handling devices, such as autosamplers or autoinjectors. The use of automated systems usually improves the precision obtained in the analytical results, and this is a feature that traditionally has been used as an argument for using automated sample handling systems. However, it is also known that automated sample handling systems can be prone to carry-over effects, and in such cases the repeatability of the results as well as the detection limits decline. Carry-over is a secondary effect of undesired sample/analyte interactions with device components employed during sample handling. The present invention seeks to minimize these problems.

Summary of the invention

The present invention is related to a method for reducing adsorption losses of analytes during sample handling where the sample comes into contact with a metal oxide surface, characterized by adsorbing a first reagent, wherein the first reagent is a component of a first solution, to the metal oxide surface before, or simultaneously as, the sample is contacted with the surface, wherein the first reagent has a pi value higher than 8.

The present invention is further related to a method, wherein the first reagent is polylysine.

The present invention is further related to a method, wherein the polylysine has a molecular weight in the range from 1 kDa to 30 IcDa.

The present invention is further related to a method, wherein the polylysine has a concentration in the range from 0.1 μM to 0.1 niM.

The present invention is further related to a method, wherein the first reagent is a polymer comprising at least one amino group.

The present invention is further related to a method, wherein the first reagent is selected from a group containing: polyarginin, polyhistidine, polyetyleneimine (PEI), polyallylamine (PAA), and diethylaminoethyl dextran (DEAE Dextran).

The present invention is further related to a method further comprising the step of adsorbing a second reagent, wherein the second reagent is a part of a second solution, to the metal oxide surface before the sample is contacted with the surface, wherein the second reagent comprises a negatively charged group.

The present invention is further related to a method further comprising the step of adsorbing a second reagent, wherein the second reagent is a part of a second solution, to the metal oxide surface before the sample is contacted with the surface, wherein the second reagent is selected from a group containing: BSA, and casein.

The present invention is further related to a method, wherein the first solution is contacted with the metal oxide surface before the second solution is contacted with the metal oxide surface.

The present invention is further related to a method, wherein the first solution further comprises BSA and/or casein. The present invention is further related to a method, wherein the metal oxide surface is part of a channel of a sample handling device.

The present invention is further related to a method, wherein the sample handling device comprises a means (for example a needle or a capillary) for sampling and a means (for example a hydraulic liquid handler) for handling a liquid which is used to rinse the means for sampling before and/or after sampling, wherein the said liquid comprises the first reagent.

The present invention is further related to a method, wherein the analytes are handled using a solvent or a mixture of solvents having a first pH value, wherein the analytes have a pi value higher than the first pH value.

The present invention is further related to a method, wherein the metal oxide surface comprises negatively charged reactive sites when said surfaces is contacted with a solution having pH =^■ 7.

The present invention is further related to a method, wherein the first solution is mixed with or added to the sample before the sample is contacted with the metal oxide surface.

In a further aspect the present invention is related to a device for sample handling comprising a channel, wherein the channel has a metal oxide surface which comes in contact with samples handled with the device, wherein the metal oxide surface has a first reagent with a pi value higher than 8 adsorbed to the metal oxide surface.

The present invention is further related to a device further comprising a second reagent comprising a negatively charged group, wherein the second reagent is adsorbed to a layer of the first reagent.

In a further aspect the present invention is related to a kit comprising a device for sample handling comprising a channel, wherein the channel has a metal oxide surface which comes in contact with the samples handled with the device, and a first reagent with a pi value higher than 8. The present invention is further related to a kit further comprising a second reagent comprising a negatively charged group.

Brief description of the drawings

Fig. 1 illustrates a standard curve for VEGF suffering from significant carry over and imprecision in the lower part of the curve.

Fig. 2 illustrates a typical ^"carry over^" experiment performed, letting the capillaries aspirate blanks of different diluents and dispense in triplicates into the CD, followed by a standard wash of capillaries.

Fig. 3 illustrates the effect of poly-L- lysine in diluent when distributing low-to-high and high-to-low concentrations of VEGF interrupted by ordinary capillary wash (Fig. 3 a)). The same effect is seen when VEGF is diluted in Rexxip A and the needles are coated with poly- L-lysin (Fig. 3b).

Fig, 4 shows a comparison of standard for MIP- lβ analysed on two different instruments in Rexxip A and 0.01 mM poly-L-lysine diluent, respectively. Instrument 16 has never handled poly-L-Lysin.

Fig. 5 shows a comparison of two standard curves for a neutral analyte, hTNFa (pi = 7, mw = 17.5 kDa). For one of the curves, the analyte (hTNFa) was diluted in a diluent comprising polylysine. For the other curve, the analyte (hTNFa) was diluted in Rexxip A. The precision between the replicates were similar for both curves.

Fig. 6 shows a comparison of two standard curves for an acidic analyte, C-peptide (pi = 3.45, mw = 3 kDa). For one of the curves, the analyte (C-peptide) was diluted in a diluent comprising polylysine. For the other curve, the analyte (C-peptide) was diluted in Rexxip A. No significant differences between the two curves were detected, and the precision between the replicates were similar for both curves.

Fig. 7 shows a comparison of two standard curves for a basic analyte, hIL-8 (pi = 9.02, mw = 8.4 IcDa). For one of the curves, the analyte (hIL-8) was diluted in a diluent comprising polylysine. For the other curve, the analyte (hIL-8) was diluted in Rexxip A. No significant differences between the two curves were detected, and the precision between the replicates were similar for both curves.

Fig. 8 shows the results from a carry-over experiment wherein a basic analyte, hIL-8 (pi = 9.02, mw = 8.4 IcDa), was diluted in a diluent comprising polylysine, or in Rexxip A. The signals obtained for each diluent type was measured before and after the application of a high amount of analyte (hIL-8) to the analysis system. The results show that there are significant carry-over effects when Rexxip A is used as a diluent. However, when the diluent comprises polylysine, there are no significant carry-over effects.

Detailed Description of the Invention

Different analytes to be quantified using miniturized immunoassay in a Gyrolab Bioaffy^R CD (Gyros AB, Uppsala, Sweden) of a Gyrolab^® Workstation (Gyros AB, Uppsala, Sweden) display different physico- chemical properties. These are primarily reflections of the amino acid sequence of the polypeptide(s), and the nature of these different amino acids. Amino acids carry different side chains (R-groups). Depending on the R-group properties, amino acids can be classified into five different groups; polar, non-polar, aliphatic, positively and negatively charged amino acids.

Gyrolab Bioaffy^® CD can be used to perform multiplexed parallel analysis of a multitude of samples (for example 112). A Gyrolab Bioaffy^® CD is a disc having a compact disc format, wherein the disc comprises one or more microchannel structures suitable for transport and mixing of fluids. The CD can rotate so that fluids are propagated through the microchannel structures due to the centripetal force. Gyrolab Bioaffy^® CD's utilize only minute amounts of reagents allowing significantly reduced consumption of reagents compared to alternative procedures.

Several other specifics may also contribute to the overall properties of the molecule. These include for instance posttranslational modifications due to glycosylation and phosphorylation that may affect molecular properties, as well as changes of conformation of proteins due to specific interactions with biological counterparts, selective cleavage of polypeptide etc.

The net effect of these variables can be used to classify proteins as acidic, basic or hydrophobic. The net effect can easily be calculated from the specific amino acid sequence using established algorithms. One such parameter is the calculated isoelectric point (pi) which represents the pH at which the net charge is zero for the molecule. This process does not base calculations on the tertiary structure of the protein only incorporating amino acids that are exposed on the surface of the protein but rather the contributions from amino acids in a linearized polypeptide. Thus, the effect of factors such as different charge densities, concentration of aliphatic amino acids to a certain region of the three-dimensional protein is not well predicted in these calculations. Still it has proven valuable to investigate the average molecular properties of a protein to get an opinion of its expected behaviour in the context of quantitative analysis.

Gyrolab^® Workstation has been designed, as a part of the automated analytical process, to transfer reagents and samples from a microtiter plate placed in one compartment of the instrument into the respective inlet holes of the CD by the use of a robot equipped with metal capillaries into which liquids are aspirated, followed by dispensing the liquid into the CD inlet. The time a given aliquot resides in the capillary may differ due to the logistics in the process. Furthermore, depending on the number of replicates to be analysed in each sample, each singlet portion of the sample may be kept in the capillary for different length of time and the liquid plug is aspirated into the capillary to different heights (lengths). Thus, the first aliquot being aspirated is sucked longer into the capillary compared to the last aliquot. The last aspirated aliquot is also the first aliquot to be dispensed into the CD inlet.

Typically the sample volumes that are used when performing analysis on Gyrolab^R Workstation has been significantly reduced (20-1000 nl) compared to conventional immunoassays (20-100 μl). In order to maintain the accuracy of liquid handling in Gyrolab" Workstation, the inner diameter of the capillaries is kept small, typically 0.3 mm. This means that the sample plug that may come into physical contact with the metal surface is significant. In Table 1 sample plug lengths and the corresponding inner surface area of the capillary that is exposed to the sample under normal processing of CDs are illustrated. Table 1.

Sample plug length in relation to replicate volume in a 0.3 mm capillary and corresponding calculated surface area of inner capillary exposed to sample when processed accordingly.

Nominal analysis Replicate

CD type volume (nl) volume (nl) Sample plug length (mm)

Single- plicate Duplicate Triplicate

Bioaffy^®

20 HC 20 420 5.94 11.9 17.8

Bioaffy^®

200 200 420 5.94 1 1.9 17.8

Bioaffy^®

1000 1000 2000 28.31 56.6 84.9

Area exposed in capillary (mm²)

Bioaffy^®

20 HC 20 420 5.6 1 1.2 16.6

Bioaffy^®

200 200 420 5.6 11.2 16.6

Bioaffy^®

1000 1000 2000 26.6 53.3 80.0

In order to prevent corrosion of the surface of the capillary and to get a stable, reproducible and functional behaviour of the capillary, it is ^"passivated^" by exposure to nitric acid. Given the particular quality of steel in the capillary, this means that metallic chrome is oxidized to chromic oxide representing a durable and homogenous metal surface suitable for sample transferring purposes. During this process, the metal surface is oxidized and becomes permanently negatively charged under the typical conditions (physiologic conditions) Gyrolab^® Workstation is operated in. Thus, positively charged protein molecules may have the possibility to interact with the negatively charged metal surface of the capillary.

Any interaction between analyte molecules in the sample and the capillary which is used for liquid transfer may distort sample integrity and contribute to either loss of analyte due to adsorption to the capillary or redistribution of analyte molecules between the different replicates of the sample that is being analysed. Loss of analyte due to adsorption of analyte to the capillary wall will lead to inaccurate quantification of analyte and possible also increase the risk of facing ^"carry over^" of analyte between samples, particularly if rinsing of needles is insufficient between processing of different samples. Redistribution of analyte molecules caused by weak interactions between analyte molecules and the capillary surface may cause increased imprecision between replicate determinations and negatively affect the analytical performance of the system.

There arc several options to prevent negative analytical consequences of analyte to surface interactions. One is to reduce the analyte to surface interactions by presenting the sample in a suitable diluent that is designed to minimize non-wanted interactions between sample and capillary. Another is to introduce sample plugs in the aspiration process that are only used to preserve integrity of those particular sample aliquots that are intended to use for analysis of replicate samples. One obvious consequence is that more sample volume has to be used to achieve this goal.

Based on the hypothesis that the metal surface is negatively charged experiments have been performed to eliminate the negative consequences of analyte interaction with the metal surface.

In order to neutralize potential effects of negative charges on the metal surface of the capillary on sample integrity, correspondingly positively charged polymers such natural proteins, synthetic polyamincs, other types of polymers such as polylysin, polyarginin, polyhistidine, polyctyleneimine (PEl), polyallylamine (PAA), diethylaminoethyl dextran (DEAE Dextran) could be used. Potentially these positively charged polymers have to be used together with BSA and/or caseins or polyanionic polymers in the diluents. There are several types of usage of such compounds; they can be incorporated into the hydraulic liquid that is used in the pumping system, they can be used for coating capillaries at regular intervals and potentially they can be incorporated into diluents that are used for dilution of samples and/or preparation of samples prior to analysis.

In Fig. 1 an illustrative example of carry over and analytical imprecision when a standard curve of Vascular Endothelial Growth Factor (VEGF, catalogue number 293-VE, R&D

Systems, UK) is analyzed without poly-L-lysin present on the needles: A carry-over effect can be seen. Here VEGF was serially diluted in steps of 2 in Rexxip A (Gyros AB, Uppsala, Sweden) to form calibration solutions used to obtain 15 standard points and one blank in a range from 1000-0.06 ng/ml. The calibration solutions were transferred to the CD using 8 capillaries in 2 sample transfer rounds beginning with the highest concentration. Biotinylated polyclonal anti-VEGF (R&D Systems, UK) from goat was used as a capturing reagent at a concentration of 0.1 nig/ml (the capturing reagent was diluted in 15 niM phosphate buffer, 150 niM NaCl, 0.02% NaN₃, 0.01 % Tween 20). Alexa-labelled polyclonal anti-VEGF (R&D Systems, UK) from goat was used as a detection reagent at a concentration of 12.5 nM (the detection reagent was diluted in Rexxip F (Gyros, Uppsala, Sweden)).

Fig. 1 illustrates a standard curve for VEGF suffering from significant carry over and imprecision in the lower part of the curve. In the first round of sample transfer, needles 1-8 transfer samples with high concentrations of VEGF. The second round of sample transfer, needle 1 -8 is used again for samples with low concentration of VEGF. For each concentration triplicate samples arc analysed. In round two the triplicates show significant imprecision due to carry over. Looking closer at these replicates the first replicate have lower signal than the last dispensed replicate.

We have been using poly-L-Lysine, where the R-group has a pKa of 10.53 to prevent VEGF, a positively charged (at pFI •= 7.2) dimeric molecule with a calculated pi of 9.16, to interact with the metal surface of the capillary Poly-L-Lysine is commercially available at reasonable cost and in product formats that differ in molecular size. So far poly-L-Lysine 15-30 IcD and 1-5 IcD has been tested at concentrations of 0.01 to 0.1 mM. The results obtained so far will be reviewed below.

Fig. 2 illustrates a typical ^"carry over^" experiment performed, letting the capillaries aspirate blanks of different diluents and dispense in triplicates into the CD, followed by a standard wash of capillaries. In the next step VEGF at lμg/ml was aspirated into the capillaries, and dispensed in triplicates in the CD, and followed by a standard wash of capillaries. Thirdly, blank diluents were once more aspirated and dispensed in triplicates into the CD. The analytical process of quantifying VEGF was completed and the response values for the final set of blanks were plotted. Four different diluents were used; Rexxip A, and specially prepared diluents where one of the components in Rexxip A, casein, was exchanged for poly- L-lysm at 0.01 mM or 0.1 mM concentrations or for poly-L-arginine at a concentration of 0.1 nM. For each diluent two needles are used dispensing triplicates. For example for the Rexxip A diluent, needle 1 and 2 dispenses triplicate of samples 1-3 and 3-6 (x-axis numbers). The VEGF response seen in the final set of blanks for the Rexxip A buffer was reduced by a more than 2 orders of magnitude using poly-L-lysin and > 1 order of magnitude using poly-L- arginine.

Tn order to verify the positive effects, a complete standard curve of VEGF covering concentrations from 100 to less 0.02 ng/ml of VEGF was tested using 0.01 niM poly-L-lysine in the diluent aspirating samples from low to high concentrations in the first round, and after ordinary washing, reusing the same capillaries aspirating samples from low to high VEGF concentrations. The curves follow each other well, even at low concentrations suggesting essentially no carry over of analyte. This is seen even if poly-L-Lysin only is present on the needles from previous runs and not in the diluent, see Fig. 3 b). These two curves should be compared to the curve in Fig. 1 which indicates carry-over effects.

Fig. 3 illustrates the effect of poly-L- lysine in diluent when distributing low-to-high and high-to-low concentrations of VEGF interrupted by ordinary capillary wash (Fig. 3 a)). The same effect is seen when VEGF is diluted in Rexxip A and the needles are coated with poly- L-lysin (Fig. 3 b)).

Control experiments indicated strongly that once capillaries were exposed for poly-L-lysine, the effect remained for a long period of time. Thus experiments performed in the same run or in the same instrument using the same capillaries subsequent to first exposure to poly-L- lysine are likely to be affected by previous experiments and hence the results should be interpreted with caution.

This observation opens up a discussion on the mechanisms involved in reducing interactions between analyte and capillary. It is likely though, that a possible coating effect of poly-L- lysine on capillaries remains after a significant time as well as exposure to liquids selected to dcsorb bound poly-L-lysine.

One obvious problem that has to be ruled out is that any new component in diluents must be inert with regards to the immunochemical principles during assaying. Furthermore, it should also be compatible with other analytes, also hydrophobic analytes, analytes having high pi values, and hydrophobic analytes having high pi values . For that reason a standard curve of recombinant human MIP- lβ cultured in E. CoIi (catalogue number 271 -BME, R&D Systems, UK) was prepared in poly-L-lysine diluent (Fig. 4). No such inferences could be demonstrated. Fig. 4 shows a comparison of standard for MIP-I β analysed on two different instruments in Rexxip A and 0.01 niM poly-L-lysine diluent, respectively. Instrument 16 has never handled poly-L-Lysin. Biotinylated monoclonal mouse anti-MIP-lβ subclass IgG₂B (catalogue number MAB-271, R&D Systems, UK was used as a capturing reagent at a concentration of 0.1 mg/ml (the capturing reagent was diluted in 15 niM phosphate buffer, 150 niM NaCl, 0.02% NaN₃, 0.01% Twcen 20). Alexa-labelled polyclonal goat anti-MIP-lβ (R&D Systems, UK) was used as a detection reagent at a concentration of 50 nM (the detection reagent was diluted in Rexxip F (Gyros, Uppsala, Sweden)). The curves has the same shape indicating no signi ficant influence of polylysin in the assay. The difference between the curves can be explained by the use of different Gyrolab Workstations.

Fig. 5 shows a comparison of two standard curves for a neutral analyte, recombinant hTNFa (pi = 7, mw = 17.5 kDa, product number 210-TA, R&D Systems, UK). For one of the curves, the analyte (hTNFa) was diluted in a diluent comprising polylysine. For the other curve, the analyte (hTNFa) was diluted in Rexxip A. The precision between the replicates were similar for both curves. Biotinylated monoclonal mouse anti-hTNFa (product number AHC3419, Invitrogcn, Biosource, US) was used as a capturing reagent at a concentration of 0.1 mg/ml (the capturing reagent was diluted in 15 mM phosphate buffer, 150 mM NaCl, 0.02% NaN₃, 0.01 % Twecn 20). Alexa-labelled monoclonal mouse anti-hTNFa (product number 551220, BD Biosciences, CA, USA) was used as a detection reagent at a concentration of 25 nM (the detection reagent was diluted in Rexxip F (Gyros, Uppsala, Sweden)). The difference between the curves can be explained by the use of different Gyrolab Workstations.

Fig. 6 shows a comparison of two standard curves for an acidic analyte, C-peptide (pi = 3.45, mw = 3 kDa). For one of the curves, the analyte (C-peptide) was diluted in a diluent comprising polylysine. For the other curve, the analyte (C-peptide) was diluted in Rexxip A. No significant differences between the two curves were detected, and the precision between the replicates were similar for both curves. Biotinylated monoclonal mouse anti-C-peptide (code number OAO 13, DakoCytomation, Copenhagen, Denmark) was used as a capturing reagent at a concentration of 0.1 mg/ml (the capturing reagent was diluted in 15 mM phosphate buffer, 150 mM NaCl, 0.02% NaN₃, 0.01% Tween 20). Alexa-labelled monoclonal mouse anti-C-peptide (code number 09583, DakoCytomation, Copenhagen, Denmark) was used as a detection reagent at a concentration of 25 nM (the detection reagent was diluted in Rexxip F (Gyros, Uppsala, Sweden)).

Fig. 7 shows a comparison of two standard curves for a basic analytc, recombinant hIL-8 (pi = 9.02, mw ~^"~ 8.4 kDa, catalogue number 208-1L, R&D Systems, UK) expressed in E. Coll. For one of thc curves, the analyte (hϊL-8) was diluted in a diluent comprising polylysine. For the other curve, the analyte (hIL-8) was diluted in Rexxip A. No significant differences between the two curves were detected, and the precision between the replicates were similar for both curves. Biotinylated monoclonal mouse anti-hIL-8 (catalogue number MAB208, R&D Systems, UK) was used as a capturing reagent at a concentration of 0.1 mg/ml (the capturing reagent was diluted in 15 mM phosphate buffer, 150 mM NaCl, 0.02% NaN₃, 0 01% Twccn 20). Alexa-labellcd polyclonal goat anti-hIL-8 (catalogue number AF-208-NA, R&D Systems, UK) was used as a detection reagent at a concentration of 12.5 nM (the detection reagent was diluted in Rexxip F (Gyros, Uppsala, Sweden)).

Fig. 8 shows the results from a carry-over experiment wherein a basic analyte, hIL-8 (pi = 9.02, mw = 8.4 kDa), was diluted in a diluent comprising polylysine, or in Rexxip A. The signals obtained for each diluent type was measured before and after the application of a high amount of analyte (hIL-8) to the analysis system. For each diluent eight needles are used each dispensing two replicates of blank before high anlalyte transfer and three triplicates of blank after high analytc sample transfer. The replicate number on the x-axis represents the following needles (needle 1, replicate number 1 -3; needle 2 replicate number 4-6; neddle 3 replicate number 7-9 etc.) The results show that there are significant carry-over effects when Rexxip A is used as a diluent. The trend that the third replicate has the highest signal is significant using the Rexxip A buffer. This indicates that the carry-over is larger for blank replicated that has been aspirated a longer distance in the needle. However, when the diluent comprises polylysine, there are no significant carry-over effects.

Summarizing, it seems that addition of poly-L-lysine to dilution buffers significantly reduces the carry over for VEGF and possibly improving the precision of assay for VEGF. Whether these effects can be extrapolated for other analytcs with different molecular properties remains to be further studied. The limited studies that have been performed so far for such analytes indicate no overt negative effects. The mechanism(s) by which the effect of poly-L-lysine is exerted remains to be elucidated. However, the experiments performed so far indicates that poly-L-lysine coats the capillary through charge interaction and that this coating is rather durable.

Claims

1 . Method for reducing adsorption losses of analytes during sample handling where the sample comes into contact with a metal oxide surface, characterized by adsorbing a first reagent, wherein the first reagent is a component of a first solution, to the metal oxide surface before, or simultaneously as, the sample is contacted with the surface, wherein the first reagent has a pi value higher than 8.

2. Method according to claim 1 , wherein the first reagent is polylysine.

3. Method according to claim 2, wherein the polylysine has a molecular weight in the range from 1 kDa to 30 kDa.

4. Method according to any one of claims 2-3, wherein the polylysine has a concentration in the range from 0.1 μM to 0.1 mM.

5. Method according to claim 1, wherein the first reagent is a polymer comprising at least one amino group.

6. Method according to claim 1, wherein the first reagent is selected from a group containing: polyarginin, polyhistidine, polyetyleneimine (PEI), polyallylamine (PAA), and diethylaminoethyl dextran (DEAE Dextran).

7. Method according to any one of claims 1 -6, further comprising the step of adsorbing a second reagent, wherein the second reagent is a part of a second solution, to the metal oxide surface before the sample is contacted with the surface, wherein the second reagent comprises a negatively charged group.

8. Method according to any one of claims 1-6, further comprising the step of adsorbing a second reagent, wherein the second reagent is a part of a second solution, to the metal oxide surface before the sample is contacted with the surface, wherein the second reagent is selected from a group containing: BSA, and casein.

9. Method according to any one of the claims 7-8, wherein the first solution is contacted with the metal oxide surface before the second solution is contacted with the metal oxide surface.

10. Method according to any one of claims 1-6, wherein the first solution further comprises BSA and/or casein.

1 1 . Method according to claim 1 , wherein the metal oxide surface is part of a channel of a sample handling device.

12. Method according to any one of the claims 1-11 , wherein the analytes are handled using a solvent or a mixture of solvents having a first pH value, wherein the analytes have a pi value higher than the first pH value.

13. Method according to any one of the claims 1-12, wherein the metal oxide surface comprises negatively charged reactive sites when said surfaces is contacted with a solution having pH = 7.

14. Method according to claim 1, wherein the first solution is mixed with or added to the sample before the sample is contacted with the metal oxide surface.

15. Method according to claim 11, wherein the sample handling device comprises a means (for example a needle or a capillary) for sampling and a means (for example a hydraulic liquid handler) for handling a liquid which is used to rinse the means for sampling before and/or after sampling, wherein the said liquid comprises the first reagent.

16. Device for sample handling comprising a channel, wherein the channel has a metal oxide surface which comes in contact with samples handled with the device, characterized in that the metal oxide surface has a first reagent with a pi value higher than 8 adsorbed to the metal oxide surface.

17. Device according to claim 16, further comprising a second reagent comprising a negatively charged group, wherein the second reagent is adsorbed to a layer of the first reagent

18. Kit comprising a device for sample handling comprising a channel, wherein the channel has a metal oxide surface which conies in contact with the samples handled with the device, and a first reagent with a pi value higher than 8.

19, Kit according to claim 18, further comprising a second reagent comprising a negatively charged group.