ABSTRACT

Following international and national quality assessment requirements in the field of human monoclonal immunoglobulinopathies, this article reports practical procedures to be carried out by clinical laboratory professionals. In particular, it develops reagent quality control (checking of antiserum specificity) and should concern biologists and reagent manufacturers. The authors describe simple and clinical biology laboratory fitted tests for qualitative techniques (immunoelectrophoresis and immunofixation) and for quantitative techniques (liquid phase immunoprecipitation assays).

INTRODUCTION

In order to meet accreditation requirements, clinical biology professionals have to write down the quality rules they are used to apply in their work schemes. A number of general recommendations and guidelines have been published recently to help biologists and manufacturers in this way (1-9). In protein clinical biochemistry, and more precisely about monoclonal Ig, reagent evaluation requires a particular care. The wide variety of assay systems and antisera which are available for the biologist nowadays implies the use of procedures for selecting adequate reagents and choice criteria taking both quality and cost efficiency requirements into account. This work describes appropriate procedures that we have developed in our laboratory since the 1980s for the study of monoclonal Ig (10-11), at the request of manufacturers. This paper stresses on reagent quality control to be applied in immunoelectrophoresis (IEP), immunofixation (IF) and quantitative immunochemical techniques. As there is no reference techniques in this field, our first quality criterion is to use complementary techniques: the right test at the right time, according to each patient case, to stress on efficacy of tests (1). Among the numerous aspects to be controlled according to the present knowledge in this field, we would like to stress on the antisera specificity, because it determines the specificity of the results. As a matter of fact, in clinical biology practice, so many unexpected substances can disturb our analysis that we absolutely need to master each step of our procedures. Many efforts and improvements have been brought in this aim by manufacturers during the past 30 years, and the experience so acquired ought to be involved in the search of constant and total quality by all clinical laboratory professionals.

MATERIALS AND METHODS

As necessary for all immunoprecipitation techniques, exclusively polyclonal antisera have been used for this work. They were for in vitro analysis. They were provided by: ATAB, distributed by Orimbio, 75014 Paris, France; Behring, distributed by Dade-Behring S.A., Immeuble Le Berkeley, 19-29, rue du Capitaine Guynemer, 92081 Paris La Defense, France; Helena, France, 6 rue Charles-Cros, ZAE, 95320 Saint Leu la For�t Cedex, France; Sanofi Diagnostics Pasteur, 3 bd Raymond-Poincare, BP 3, 92430 Marnes la Coquette, France; and Dako S.A., rue des Charmes, BP 149, 78196 Trappes Cedex. According to national or international regulations (1, 9), we have checked that minimum registration information was mentioned on the inserted sheet. When we could not obtain the required information from the manufacturer, we have performed the described tests at the bench, and renewed them for each batch. All testing operations (complete analytical processes and results - according to the written working procedure -) have been recorded in a dedicated lab book. Start and end of use dates of all batches have been mentioned so as to ensure quality traceability for all analysis performed (1). Specifically chosen sera or control specimens have been used. They are spared in an appropriate and very carefully managed bank of sera.

As qualitative techniques, we have used IEP and IF. For IEP, common selection criteria for all antisera have been: they all appeared homogeneous and clear. We have frozen the content of any opened vial at -20 �C in aliquots, without any further unfreezing or re-freezing. We have discarded all antisera producing a persistent precipitate around the deposit gutter after washing.

For evaluation of an anti-total protein antiserum, we have preferred equine producers (horse, donkey, mule�). We have currently used a convenient mixture of different anti-total protein antisera from various brands, provided that they did not contain free antigens. The resulting mixture has been carefully controlled against a human normal serum pool (HNS). We have systematically avoided mixing antisera from different animal species.

We have explored the required specificities for an anti-human serum total protein antiserum according to previously published works (10). These main controls are: a test with two urine specimens containing Ig free light chains, one of the kappa type, the other of the lambda type, or with sera containing free light chains, see table I; a test with a monoclonal IgD containing serum; a test with a normal plasma and a test with an hemolysed serum. We have carried out a rough antibody titer control with a half diluted normal serum.

For pentaspecific (anti-IgG+IgA+IgM) or trispecific antisera (anti-gamma+alpha+mu chains), we have performed controls with a HNS, as in table I, as well as with two sera (or urine specimens) containing kappa or lambda free light chains. We have controlled anti-IgG (gamma+kappa+lambda) antisera with two monoclonal IgG containing sera, one of the kappa type, the other of the lambda type, and with two urine specimens containing kappa or lambda free light chains.

For Ig class monospecific antisera, we have chosen a rabbit origin. We have tested their monospecificity with an HNS. For anti-IgD antisera, we have checked their reactivity with an IgD myeloma serum, see table I. Then, we controlled them with a high concentration Ig containing serum (a polyclonal hypergammaglobulinemia serum such as that from a cirrhosis or septicaemia patient).

Evaluation of anti-Ig light chain antisera: two kinds of specific antisera are commercially available; when precessions about their specificity is lacking, they should be considered as anti-kappa or anti-lambda antisera detecting both complete Ig bound and free light chains. Their kappa or lambda bound light chain specificity has been controlled with known monoclonal Ig, see table I and figure 1a, and their free light chain specificity with free light chain containing urin, see table I. The second kind of anti-Ig light chain antisera is the kind that reveals unbound kappa or lambda light chains only. Their specificity has been checked with a HNS, as in table I and figure 1b.

For use in IF, the supplied antisera should fulfil the requirements needed for IEP antisera. Moreover, higher qualities are needed for these antisera to be fitted to IF. We used Ig class, type or fragment monospecific antisera only, as they are the only liable ones to identify monoclonal Ig. We have always preferred goat antisera.

We have evaluated their specificity in IF out by testing cross-reactivity between Ig and their various fractions as in IEP (10), see table I. As well as in IEP, for anti-free light chain antisera, we checked the absence of cross-reactivity either with light chains bound to Ig heavy chains, or with the other type, kappa or lambda as in figure 2. We have systematically included normal plasma, as in figure 3a, and a strongly hemolysed serum in our controls.

Evaluation of affinity and avidity: we have searched for anti-bound lambda light chain antisera lacking avidity for some monoclonal IgA, as in figure 4. We have performed variable antigen concentration/fixed time tests by applying various dilutions of the antigen for which they were labeled monospecific, at fixed time. So doing, we got their reactivity zone, expressed as antigen concentration, starting at the detection limit and ending at the antigen excess zone. We have also tried fixed antigen concentration/variable time tests . We have evaluated specificity, affinity and avidity at every batch change, and then regularly as a control of correct batch shelf life.

In quantitative techniques (BNA and Turbitimer, Behring), we have chosen rabbit or goat antisera. We have started qualitative evaluations of antisera by checking optical quality. We have preferred liquid form to lyophilised form, and controlled clarity by a simple visual investigation. When this step had been validated, we have confirmed this quality in the instrumental system used, by performing an antiserum blank (10).

For quantitative evaluations, we determined the limits of the quantitative reaction (detection limit, antigen excess) according to the involved assay type, achieving a complete precipitation curve from the detection limit up to the point where a signal decrease was observed (beginning of the antigen excess zone).

For internal quality control in qualitative techniques, we followed the current schemes: alternating the internal quality material (a HNS generally), specimens and antisera in IEP (10), introducing the analysed specimen itself as a single control material per film. In quantitative techniques, we performed the internal quality control scheme with instrument adapted specimens, provided by the manufacturer of the analytical system. These specimens must have been of human origin. For each of the Ig, acceptable values had been determined by scientific societies documents (14).

RESULTS

In all cases, we observed variations from one batch to another, even from a same manufacturer. The main adequate results we have obtained in selecting our IEP antisera are mentioned in the right column of table I.

For appropriate anti-human serum total protein antisera, the test with a HNS showed 15 to 20 arcs of clinically useful proteins, including IgG, IgA and IgM. For the test with Ig free light chains urin or serum specimens, as well as with a monoclonal IgD containing serum, we obtained one single additional arc, compared to the set of arcs yielded by the HNS. On the contrary, IEP of a normal plasma or of a hemolysed serum have not shown any additional arc.

We have declared such antisera as valid when a half diluted normal serum has exhibited the same, though thinner, precipitation arcs as an undiluted HNS.

For correct trispecific or pentaspecific antisera, the HNS test has given 3 arcs only. With free light chain containing sera, one more arc has been visible, linked up to others.

We have evidenced monospecificity of anti-gamma, alpha, mu, delta chain antisera by a single arc with a HNS, or with a polyclonal hypergammaglobulinemia serum, or with a matching monoclonal Ig containing serum (including IgD).

Figure 1a shows the results obtained while testing the bound light chain specificity of anti-kappa and anti-lambda antisera. A HNS gives one arc for each, and the kappa arc is about twice as thick as the lambda arc. But this last criterion is not an absolute one. Monoclonal IgG are correctly typed, and urin free light chains are detected. Figure 1b shows the results obtained while testing the free light chain specificity of anti-kappa and anti-lambda free light chain antisera. A HNS never gives any arc. Monoclonal IgG do not react with this sort of antisera, and urin free light chains are detected.

Figure 3a shows the results yielded by a non-specific anti-bound and free lambda light chain antiserum: this antiserum cross-reacts with fibrinogen, giving a false positive result with a plasma. Figure 2 shows a case of non specificity for an anti-free lambda light chain antiserum. The test has been performed with a patient serum containing both a complete monoclonal IgGlambda and free lambda light chains, evidenced by figure 3a. A reference test in figure 3b shows the proper result to be displayed for the same patient with a truly free lambda light chain antiserum, without cross-reaction with the IgG bound lambda light chains. On figure 3c, we can see results yielded by a supposedly anti-free lambda light chain antiserum, which actually reveals both bound and free lambda light chains.

Figure 4 gives an example of avidity requirements needed in IF: although the anti-bound and free lambda light chain antiserum had shown adequate properties in IEP, it lacks avidity for certain monoclonal IgA. Various dilutions of this patient serum at fixed time, and various contact time at constant IgA concentration have been tried, without any success. We have evaluated antisera as good ones when the detection limit so obtained allows to detect down to 0,5 g/l monoclonal Ig concentrations.

In quantitative techniques, antisera have been considered as valid according to the antiserum blank test when its absorbency value has been significantly lower than that of the lowest calibration point.

We have evaluated results of internal quality control of qualitative techniques following our own usual requirements. These validation criteria notably refer to the aspects of the precipitation arcs (length, thickness, outline sharpness, intensity, relative position to the gutter or deposit well, surrounding disturbing precipitation's�), to the homogeneous transparency of the reaction plate background (a proof of the proper washing processes and of the stain good conservation state) and to the staining intensity.

DISCUSSION

We easily explain batch to batch variations in the results of these tests by the fact that these reagents are not chemically defined substances. This is the reason why selecting procedures must be undergone upon each batch changing. We have frozen occasionally used antisera, in order to prevent repeated polluting risks, and never repeated unfreezing/re-freezing to avoid a significant loss of activity.

While choosing an anti-total protein antiserum for IEP, we have preferred equine producers, because of their precipitation curve shape (10). As a matter of fact, the equivalence zone being broader, it is easier to find it out, which is a real advantage as one tries to get optimal arcs for many proteins. Moreover, the non-reactivity zone observed for low concentrations of antigen raises the detection limit of the reaction, which is not a disadvantage here, considering the screening use and the clinical relevance of this test. We usually mix different anti-total protein antisera from various brands, provided that they do not contain free antigens which could eventually react with antibodies contained in one of the reactants. We perform this process to enrich an antiserum batch with antibodies which are present at a too low titter, so as to homogenize the various reaction levels. However, we never mix antisera from different animal species, to avoid interspecies reactions.

Although it is impossible to determine the titter of the various antibodies in a polyspecific antiserum, we perform the described rough control. This allows us to manage a wide enough reactivity zone for the study of sera likely to show important quantitative variations of their constituents.

We have tried trispecific or pentaspecific antisera in experimental procedures, but not validated their use because of a too low information/cost efficiency ratio. The three precipitation arcs overlapping, it is not always possible to conclude about the heavy chain responsible for the observed abnormality in case of monoclonal Ig. Besides that, anti-total protein antiserum is much more informative, and identifying the monoclonal Ig implies the use of strictly class and type monospecific antisera (6).

In contrast to what happens for anti-total protein antisera, a single equivalence zone is used at a time for monospecific antisera. Thus, for class monospecific antisera, we have chosen rabbit antisera because the narrowness of their equivalence zone allows a better evaluation of the protein concentration variations so studied. Moreover, they provide a better and useful detection limit to the reactions.

The tests we have designed for isotype monospecific antisera intend to prove the absence of reactivity against undesirable proteins, and mainly other Ig classes or types. As usual for these immunochemical techniques, we need to be careful about interpretation of results: a truncated precipitation arc leads to suppose an antigen excess phenomenon. Thus, we have tried revealing this arc or arc fragment, by changing the antigen/antibody ratio, with dilutions of either specimen or antiserum. Considering practicability and reliability of the analytical process, we find it easier to choose antisera which provide good results without any dilution. However, when necessary, we adjust antigen/antibody ratios by varying dilutions of antigens (specimens) only, preferably to antisera. Nevertheless, considering the very wide variety of monoclonal Ig (for a single class), some antisera might be too weak whatever the working conditions (particularly the antigen/antibody ratio). Then we prefer avoiding the use of such antisera: as a matter of fact, these antisera might have been produced from a pool of monoclonal Ig, the sampling of which could always be inadequate. We think that our analysis should give more reliable results if we use monospecific antisera that have been produced from a pool of purified polyclonal Ig.

Distinguishing between bound and free light chains and free light chains is of the utmost importance, because free light chains must be positively identified, and not merely supposed by the absence of abnormalities with all class specific antisera and the presence of an abnormality with anti-bound and free light chain antisera (10). Moreover, they allow the assessment of free light chain presence even when they are associated to complete monoclonal Ig.

Specificity is more difficult to define in IF than in IEP. Actually, whereas the presence of a single arc in IEP is a good monospecificity criterion for an antigen or an antibody, the situation is different in IF, since a single band can correspond to an exact overlapping of two (or more) different antigens which are not separated by electrophoresis. Similarly, the aspect of IF bands is far less varied than the characteristics of IEP arcs. Indeed, the latter can be described by their length, thickness, outlines, or by the distances separating them from the antiserum gutter and from the deposit well. The IF bands provide only two description elements: width and intensity. We give an example of this in figure 3, showing how a double error (use of a plasma + use of a non-specific antiserum) could lead to an erroneous result in IF, while IEP allows a clue through the aspect of the arcs, see figure 3b. A similar cross-reactivity between fibrinogen and IgM had also been described previously (13). For this reason, besides tests of cross-reactivity between Ig and their various fractions (10), we have included non-reactivity tests against undesirable proteins, which are already causes of errors in electrophoresis (11) for the specificity evaluation of antisera dedicated to monoclonal Ig identification. These accidental proteins, simulating a monoclonal Ig, are mainly fibrinogen and haemoglobin. Thus it is highly suggested to check the antisera with a normal plasma and a strongly hemolysed serum. As well as in IEP, the use of anti-free light chain antisera in IF should be strictly submitted to their monospecificity: no cross-reactivity either with light chains bound to Ig heavy chains, or with the other type, kappa or lambda, see figure 2.

Owing to the very short time of reaction used in IF, compared to that necessary for IEP, all IEP fitted antisera are not necessarily appropriate for IF, as in figure 4. As a matter of fact, their reactivity (or affinity) should be evidenced in a short time. This property is usually called avidity. As they have a better avidity than rabbit antisera, goat antisera are the appropriate antisera to be used in this technique. They also have a broader equivalence zone. The fact that some anti-bound and free lambda light chain antisera do not react with some particular lambda type monoclonal IgA had been pointed out a long time ago (12). We confirmed it regularly in our daily practice. This is why we included the above described test while selecting our antisera.

The dilution process described also makes it possible to determine optimal reactivity conditions for each antiserum (still as antigen concentration), in order to calculate the dilutions to be used for each studied biological fluid. The optimal antigen concentration so obtained, corresponding to the equivalence zone, should be as low as possible to get obvious bands with very low concentration monoclonal Ig. Usually, these conditions yield a very sharp precipitation curve, but this is not a disadvantage, since antigen excess situations are actually not very dangerous in this technique.

We have used quantitative techniques to quantify polyclonal Ig as markers of remaining circulating immunity, and not to quantify monoclonal Ig (11). We have preferred rabbit or goat antisera because of the very low detection limit they provide to the reaction. The nephelometry grade quality of these reagents should be mentioned on the package, so the manufacturer should comply to it, and clearing processes must not to be necessary at first opening of flask for properly stored reagents. We have not developed procedures to control the composition of antisera, since we had shown (10) that (whole serum or purified antibodies) has no influence on the properties needed for clinical use. As the manufacturer is supposed to control the monospecificity of these antisera, it is not of common use to perform it at the bench. It is evidenced by a single arc with a HNS and with a polyclonal hypergammaglobulinemia serum in IEP. Its characteristics should be identical to those usually observed for the protein matched to the announced specificity.

To comply to internal quality control requirements (1, 9) for qualitative techniques, biologists have to check each step of the process (electrophoresis, antigen-antibody reaction, staining, reporting procedures), as well as each of the work unit (film, slide�). For IEP, the usual schemes (10) allow this constant checking. For IF, where current practice is to introduce a single (but necessary) control material per film, to check the electrophoresis process, this material being the analysed specimen itself, the above requirements are not fulfilled (8). To reach an internal quality control level similar to that performed in IEP, we should design a scheme (for each film) allowing a reaction between a control material (a HNS for instance) and each of the involved antisera, during the same analytical process as the assayed specimens. This rigorous quality control scheme of an IF technique cannot be currently carried out in clinical biology and this is one of the points where this technique is significantly less reliable than IEP. This adds one more reason why it is not possible to encourage an exclusive use of IF. It is better to consider this technique as a complementary technique of IEP. The results to be reached in these tests have been set according to our own traditional standards, which is a common way of doing for many a laboratory like ours. We think that to fully meet the present quality needs, our criteria of choice should be stated in the laboratory working procedures (1), along with practical examples of targets. For this technique, it should be particularly stressed on that written procedures should specify decisions to be taken in case of abnormalities, and that a very careful follow-up should be ensured.

CONCLUSIONS

Quality Assessment in the field of monoclonal Ig is widely dependant on reagent performances in the necessary antigen-antibody reactions. Antisera selecting and quality checking are essential processes to be carried out before use of these techniques. As these antigen-antibody reactions follow the immunoprecipitation rules, the antibodies must be polyclonal (antisera). They should look homogeneous and clear. It is often useful to check that the species of the producer animal is appropriate to the performed technique. Specificity tests are particularly important to undergo, especially for anti-free light chains. The authors suggest that users check specificity for each new batch, if this cannot be proved by the manufacturer. Moreover, for quantitative techniques, it is advised to control that antisera have optical qualities consistent with use in nephelometry or turbidimetry. For these last techniques, it is very useful to know the detection limit and the antigen excess zone linked to the antiserum. As these tests are designed to ensure the analysis quality for a fairly long period of work, it remains necessary to introduce a control specimen in each working series (even if it includes only one specimen), in order to check each step of the analytical process.

ACKNOWLEDGEMENTS

The authors are grateful to all technicians of our laboratory for excellent technical assistance

REFERENCES

1. Jansen, R. T. P. et al. European Communities Confederation of Clinical Chemistry (EC4): Essential Criteria for Quality Systems in Medical Laboratories. (1997) Eur. J. Clin. Chem. Clin. Biochem. 35, 123-132. MEDLINE
2. Tomar, R. H. College of American Pathologists Conference XXXII on Guidelines for Laboratory Evaluation and Use of Antinuclear Antibodies and Laboratory Diagnosis and Monitoring of Monoclonal Gammopathies. Introduction to the Development of Guidelines. (1999) Arch. Pathol. Lab. Med. 123, 103.
3. Goeken, J. A. and Keren, D. F. Introduction to the report of the Consensus Conference on Monoclonal Gammopathies. (1999) Arch. Pathol. Lab. Med. 123, 104-105. MEDLINE
4. Keren, D. F., Alexanian, R., Goeken, J. A., Gorevic, P. D., Kyle, R. A. and Tomar, R. H. Guidelines for Clinical and Laboratory Evaluation of Patients With Monoclonal Gammopathies. (1999) Arch. Pathol. Lab. Med. 123, 106-107. MEDLINE
5. Alexanian, R., Weber, D. and Liu, F. Differential Diagnosis of Monoclonal Gammopathies. (1999) Arch. Pathol. Lab. Med. 123, 108-113. MEDLINE
6. Kyle, R. A. Sequence of Testing for Monoclonal Gammopathies. Serum and Urine Assays. (1999) Arch. Pathol. Lab. Med. 123, 114-118. MEDLINE
7. Kallemuchikkal, U. and Gorevic, P. D. Evaluation of Cryoglobulins. (1999) Arch. Pathol. Lab. Med. 123, 119-125. MEDLINE
8. Keren, D. F. Procedures for the Evaluation of Monoclonal Immunoglobulins. (1999) Arch. Pathol. Lab. Med. 123, 126-132. MEDLINE
9. Arrete du 26 novembre 1999 relatif la bonne execution des analyses de biologie medicale. J. O. Rep. Fr. du 11 decembre 1999, 18441-18452.
10. Pontet, F., Diemert, M. C., Pressac, M. and Bienvenu, J. Study of six antisera used for immunonephelometric assay of human IgG. (1992) Eur. J. Clin. Chem. Clin. Biochem. 30, 145-152. MEDLINE
11. Pontet, F., Doche, C. and Bienvenu, J. Projet de recommandations pour l'etude des immunoglobulinopathies monoclonales au laboratoire. (1997) Ann. Biol. Clin. 55, 486-490. MEDLINE
12. Solomon, A. Light Chains of Human Immunoglobulins. (1985) Methods in Enzymology 116, 101-121. MEDLINE
13. Register, L. J. and Keren, D. F. Hazard of Commercial Antiserum Cross-Reactivity in Monoclonal Gammopathy Evaluation. (1989) Clin. Chem. 35, 2016-2017. MEDLINE
14. Vassault, A., Grafmeyer, D., de Graeve, J., Cohen, R., Beaudonnet, A. and Bienvenu, J. Analyses de biologie medicale: specifications et normes d'acceptabilite � l'usage de la validation de techniques. (1999) Ann. Biol. Clin. 57, 685-695. MEDLINE