Method for determining active concentrations and/or kinetic interaction constants in complex biological samples by means of surface plasmon resonance

Abstract

The invention relates to a method for determining active concentrations of an analyte and optionally kinetic constants for the interaction of the analyte with a ligand in complex biological samples by means of surface plasmon resonance comprising the use of an auto-blank.

Claims

1. A method for determining, by surface plasmon resonance in complex biological samples, active concentrations of an analyte and optionally kinetic constants of interaction of the analyte with a ligand, comprising providing a surface plasmon resonance chip on which a capture agent specific to a ligand of the analyte is immobilized; capture, by the capture agent, of a control ligand that does not bind the analyte to be tested; passage of the sample over the chip, injected at a determined flow rate for a determined duration and obtaining a sensorgram for the control ligand; regeneration of the surface to remove the captured control ligand from the capture agent; capture, by the capture agent, of a ligand that binds the analyte to be tested; passage of the sample over the chip, injected at the same determined flow rate for the same determined duration and obtaining a sensorgram for the analyte; subtraction of the sensorgram obtained with the control ligand from the sensorgram obtained with the ligand that binds the analyte to be tested; and calculation of the active concentration of the analyte and optionally kinetic constants of interaction of the analyte with the ligand, and characterized in that the control ligand and the ligand that binds the analyte to be tested are of a similar mass and the control ligand and the ligand that binds the analyte to be tested are captured in equivalent amounts by the capture agent.

2. The method as claimed in claim 1, characterized in that the complex biological sample is selected from serum, plasma, urine, lavage liquids, ascites, biopsy eluates and cell culture media.

3. The method as claimed in claim 2, characterized in that the complex biological sample is serum or plasma.

4. The method as claimed in claim 1, characterized in that the sample is injected at least two different flow rates and that the active concentration of the analyte in the complex biological sample is calculated.

5. The method as claimed in claim 1, characterized in that the sample is injected at different concentrations and that the kinetic constants of interaction of the analyte with the ligand in the complex biological sample are calculated.

6. The method as claimed in claim 1, characterized in that the analyte-ligand pair is chosen from an antibody-antigen pair, a ligand-receptor pair or a xenobiotic-molecular target pair.

7. The method as claimed in claim 1, characterized in that the analyte is an anti-HLA antibody, the control ligand is an HLA antigen not recognized by the antibody to be tested and the ligand that binds the analyte to be tested is a target HLA antigen of the antibody to be tested.

8. The method as claimed in claim 7, characterized in that the sample undergoes one or more prior treatments chosen from a heat treatment, a treatment with dithiothreitol (DTT), a step of purification of the IgGs on a protein G resin, a step of concentration of the sample, a step of dialysis, in particular with a cut-off threshold of 100 kDa, and a combination of several of these treatments.

9. The method as claimed in claim 8, characterized in that the sample previously undergoes a combination of a heat treatment and a step of purification of the IgGs on a protein G resin.

10. The method as claimed in claim 9, characterized in that the method comprises a prior step in which the anti-HLA antibodies in the sample are detected.

11. The method as claimed in claim 8, characterized in that the method comprises a prior step in which the anti-HLA antibodies in the sample are detected.

Description

DESCRIPTION OF THE FIGURES

(1) FIG. 1. Successive steps leading to the expected result from the patient's serum.

(2) FIG. 2. CFCA auto-blank from a sample originating from the serum of a patient having anti-DQ2 antibodies (auto-blank DQ5). Dilution of the sample to 1/200, measured concentration 1.4 nM, concentration in the sample 270 nM, concentration in the serum after correction as a function of the IgG assays 460 nM, QC ratio 0.496.

(3) FIG. 3. SCK auto-blank from a sample originating from the serum of a patient having anti-DQ2 antibodies (auto-blank DQ5). Concentration used, deduced from the CFCA auto-blank: 2.16 nM, 10.8 nM and 54 nM. k.sub.a=1.493×10.sup.6 M.sup.−1.Math.s.sup.−1, k.sub.d=2.482×10.sup.−4 s.sup.−1, K.sub.D=1.663×10.sup.−10 M.

(4) FIG. 4. Comparison of the blank (light sensorgrams) and of the reaction (dark sensorgrams) of an SCK auto-blank from a sample originating from the serum of a patient having anti-DQ2 antibodies (auto-blank DQ5).

(5) FIG. 5. System making it possible to measure the concentration of B2M by CFCA SPR in complex media.

(6) FIG. 6. CFCA sensorgrams for the B2M assay under different conditions: conventional blank (PBS-T) and auto-blank in a serum treated with DTT and dialyzed, diluted to 1/1000, 1/100 and 1/10. The sensorgrams are in gray, the adjustment made by the analysis software is in black.

EXAMPLES

(7) The subject of the invention is the culmination of two years of experiments intended to make it possible to analyze complex media such as samples from patients and thus to eliminate the NSB that they systematically exhibit.

(8) The inventors first tested sera from patients diluted to 1/10 over the surface on which B2M-01 (mouse anti-B2M IgG2a) was immobilized (approximately 16 000 RU). All the sera showed strong interaction with the surface (table 1), whether or not the capture antibody was immobilized.

(9) TABLE-US-00001 TABLE 1 Observation of NSB with sera from patients (1/10 dilution in PBS-T) on the reference lane (empty lane 3) and the lane with 16 000 RU of immobilized B2M-01. Values in RU. Flow rate 25 μl/min, injection duration = 1 min. Injection Lane 3 (empty) Lane 4 (B2M-01) 4-3 Serum 001 195 1871 1650 Serum 003 59 2000 1900 Serum 004 44 1544 1513 Serum 005 61 1260 1195 Serum 007 208 1988 1779 Serum 010 87 1412 1300

(10) Various treatments were applied to the samples in order to attempt to reduce NSB (elimination of small proteins by dialysis or complement inactivation). These treatments only proved to be partially effective since residual non-specific binding remained, to varying degrees. The residual NSB might not have posed a problem if it had been identical for all the patients. In this case, a serum or a mixture of sera devoid of anti-HLA could have been used as blank. However, it was observed that residual NSB was highly variable among the samples.

(11) The inventors then undertook to purify the serum IgGs on protein G resins. The first results using purified IgGs diluted in PBS-T were not compelling. They even observed more NSB with the compositions purified on protein G, and that NSB was also variable on the reference lane as a function of the sample analyzed.

(12) A certain number of other tests were carried out: denaturing the serum with heat, exhausting NSB by passing the sample over lanes upstream of the test lane with immobilized anti-B2M, pre-saturation of the lane with a pure serum, addition of BSA to the running buffer, deactivation of the surface, immobilization of the anchor on the reference lane. All also proved to be insufficiently effective.

(13) At this stage of the investigations, the inventors accepted that it was impossible to reduce to zero the NSB obtained when samples originating from patients were used. Nonetheless, a crucial item of information obtained was that capturing HLA reduced NSB, most likely because the mass of proteins present on the chip was increased in comparison to the absence of capture.

(14) Thus, the inventors had the idea of using each sample as its own blank on a lane on which an HLA molecule, not recognized by the antibodies present in the sample, is captured in equivalent amounts to the target HLA molecule of these antibodies.

(15) The inventors first validated this method using monoclonal antibodies, used in derivatives of serum from non-anti-HLA-immunized patients, for which they determined the active concentration by CFCA. This was carried out for a mouse monoclonal anti-HLA class I antibody (anti-HLA-A2) (capture of the HLA molecules by an anti-B2M antibody, clone B2M-01) and a mouse monoclonal anti-HLA class II antibody (anti-DQ2) (capture of the HLA molecules by a commercial anti-DQ antibody, clone Tu169). The part of the process taking place on the SPR apparatus was as follows: 1—Immobilization of the capture antibody by chemical coupling on the SPR chip; 2—Carrying out the CFCA method: For the blank, firstly, an HLA antigen not recognized by the antibody to be tested is captured in a large amount on the surface, then the solution containing the antibody to be tested is injected at a determined flow rate for a determined duration. Finally, after a short period of dissociation, the surface is regenerated, which makes it possible to pass to a subsequent cycle. For the test itself, firstly, the HLA antigen recognized by the antibody to be tested is captured in a large amount on the surface, then the solution containing the antibody to be tested is injected at the same determined flow rate for the same determined duration. Finally, after a short period of dissociation, the surface is regenerated, which makes it possible to pass to the following cycle. The analysis software of the apparatus performs the correction by subtracting the sensorgram of the blank from the tests one; this is referred to as double-referencing.

(16) The inventors were able to test the auto-blank under four different conditions for each class of HLA in comparison with the running buffer (RB; this is PBS-T). For this purpose, they used complex media derived from sera from patients with the following NSB characteristics (dilution ½ in PBS-T with 10% of “NSB reducer”, a solution of dextran provided by GE Healthcare which is not however indispensable) (table 2):

(17) TABLE-US-00002 TABLE 2 NSB of the samples used to validate the auto-blank method for CFCA. Flow rate 25 μl/min, injection duration = 1 min Anchor B2M-01 Tu169 (anti-DQ) Treatment of Heat + Heat + DTT + Heat + Heat + DTT + the serum protein G dialysis protein G dialysis BOY 20 341 24 29 MOU 28 235 49 75 PIC 23/17 151/120 24 32 TAS 30/25 423/349 22 45

(18) This table highlights the heterogeneity of the NSB as a function of the samples and the pre-treatments but also of the anchor used. The following tests are applied to using different NSB samples, to which the method carried out in PBS-T was compared (table 3).

(19) TABLE-US-00003 TABLE 3 CFCA of two monoclonal antibodies in complex media using the auto-blank method, in comparison with the method in simple medium “RB” which defines the concentration target. The auto-blank HLA molecule was HLA-A11 for anti-HLA-A2, and HLA-DQ7 for anti-HLA-DQ2. The QC ratio is an indicator of quality, which demonstrates the presence of mass transport. The results are 100% reliable if the QC ratio is greater than 0.2. Class I (A2OL-A2/A11) DQ (DQ2OL-DQ2/DQ7) RB RB Sample PIC TAS (target) BOY MOU (target) Heat + Concentration 1.5 1.7 1.4 0.69 0.65 0.73 protein G measured (nM) QC ratio 0.549 0.509 0.558 0.446 0.464 0.399 Heat + Concentration 0.96 1.8 1.4 0.96 0.91 1.3 DTT + measured (nM) dialysis QC ratio 0.952 1 0.558 0.391 0.405 0.296

(20) This table demonstrates the reliability of the results obtained by virtue of the auto-blank, the difference to the target being of the order of a tenth of a nM.

(21) The inventors then tried this approach on 5 patients having anti-HLA DQ2 and DQ7 antibodies, the sera of which were treated with heat, DTT, and then dialyzed. The results are presented in table 4:

(22) TABLE-US-00004 TABLE 4 CFCA of anti-HLA DQ antibodies from samples originating from sera from patients. These sera were identified as containing anti-HLA DQ2 or DQ7 antibodies by the Luminex ® Single Antigen (SAFB) technique, the fluorescence value (MFI) for which is specified in the table. The treatment applied to the sera was a heat treatment (56° C. 30 min), DTT (5 mM 37° C. 30 min) then dialysis against PBS-T (100 kDa membrane, 1 hour, 6 hours then overnight). The total IgG assay carried out before and after treatment of the serum makes it possible to determine the initial concentration of the IgG anti-HLAs present in the serum. For passage on the SPR apparatus, the samples were used diluted to ½ in PBS-T with 10% of “NSB reducer”. For the anti-DQ2s (DQB1*02:01/DQA1*05:01), the auto-blank was carried out on captured HLA-DQ7, and vice versa for the anti-DQ7s. Corrected Concentration concentration Initial IgG Final IgG Target SAFB MFI (nM) (nM) (g/l) (g/l) BEL ½ DQ2 11237 1.6 2.18 9.65 7.07 BRI ½ DQ2 20507 7.6 12.71 11.00 6.58 LAP ½ DQ2 15000 3.3 4.90 9.02 6.08 PLA ½ DQ7 20627 3.4 5.46 8.33 5.19 THI ½ DQ7 21464 2.1 2.83 11.60 8.61

(23) It is interesting to note that the inventors did not observe a perfect relationship between the active concentration of the anti-HLA-DQs obtained by SPR and the MFI by SAFB. This suggests that the data obtained by SPR could provide different information. This could be crucial for patients.

(24) It should be noted that there are large “buffer leaps” when using the serum treated with heat+DTT+dialysis, and that the time required to obtain a sample that is ready to be used on the Biacore platform is 24 hours (after identifying the antibodies by SAFB which lasts half a day). More recently, the inventors opted for the second pre-treatment mentioned, that is to say heat treatment of 1 ml of serum (56° C. for 30 min) then purification on protein G resin (1 hour), and finally concentration of the purified composition to return to an equivalent volume (1 to 1.2 ml) in PBS-T. Assaying the total IgGs before and after purification makes it possible to return to the initial concentration of the anti-HLA antibodies.

(25) It is important to note that these two treatments make it possible to do away with anti-HLAs of IgM isotype that may be present in the serum of patients at the same time as the IgGs. These IgMs were identified as potentially interfering molecules in SAFB, but could also have interfered with the analysis of the anti-HLAs of IgG isotype by SPR.

(26) FIG. 1 details the successive steps leading to the expected result from the patient's serum. The workflow presented in FIG. 1 shows that the time for carrying out the examination is entirely compatible with a clinical application, since the characterization of anti-HLA antibodies is not an urgent examination.

(27) In the end, it is therefore entirely possible to obtain a result in two working days. The example of a patient having anti-DQ2 antibodies, studied by this process, is presented below. FIGS. 2 and 3 present the results from CFCA then SCK auto-blank. The fit of the sensorgrams to the kinetic interaction model is highly satisfactory, as shown by the superposition of the experimental curves and theoretical curves.

(28) FIG. 4 demonstrates the level of NSB during sample injection with increasing concentration, and the importance of using the auto-blank which makes it possible to correct this NSB. Indeed, strong binding to the lane clearly appears in the absence of the target HLA molecule.

(29) The method is entirely reproducible (tables 5 and 6).

(30) TABLE-US-00005 TABLE 5 CFCA auto-blank from a sample originating from the serum of a patient having anti-DQ2 antibodies (auto- blank DQ5). Dilution of the sample to 1/200 or 1/100 from frozen aliquots tested on different days. Concentration Concentration Dilution measured (nM) sample (nM) QC ratio Aliquot 1 200 1.4 270 0.496 Aliquot 2 200 1.5 290 0.526 Aliquot 3 100 3.2 320 0.439

(31) TABLE-US-00006 TABLE 6 SCK auto-blank from a sample originating from the serum of a patient having anti-DQ2 antibodies (auto-blank DQ5). Tests carried out from frozen aliquots on different days. k.sub.a (1/Ms) k.sub.d (1/s) K.sub.D (M) Aliquot 1 1.49 × 10.sup.6 2.48 × 10.sup.−4 1.66 × 10.sup.−10 Aliquot 2 1.36 × 10.sup.6 2.37 × 10.sup.−4 1.74 × 10.sup.−10 Aliquot 3 1.24 × 10.sup.6 2.38 × 10.sup.−4 1.90 × 10.sup.−10

(32) The inventors repeated the use of this approach on 10 additional patients having anti-HLA DQ2, DQ4, DQ6, DQ7, DQ9 antibodies and having received different types of transplant (lungs, heart, kidney, liver). The results are presented in table 7.

(33) TABLE-US-00007 TABLE 7 CFCA and SCK of anti-HLA DQ antibodies from samples originating from sera from patients. These sera were identified as containing anti-HLA DQ antibodies by the Luminex ® Single Antigen (SAFB) technique, the fluorescence value (MFI) for which is specified in the table (alpha chain associated with beta chain specified in brackets where relevant). Corrected Initial Final Target Blank SAFB Concentration concentration IgG IgG k.sub.a k.sub.d K.sub.D HLA-DQ HLA-DQ MFI (nM) (nM) (g/l) (g/l) (1/Ms) (1/s) (M) DQ2 (05) DQ5 15570 5.2 7.85 10.4 6.90 5.71 × 10.sup.6 5.59 × 10.sup.−4 9.79 × 10.sup.−11 DQ2 (05) DQ5 13802 15 24.15 11.3 7.02 2.43 × 10.sup.6 1.37 × 10.sup.−4 5.61 × 10.sup.−11 DQ7 DQ5 20442 65 100.6 6.47 4.18 1.19 × 10.sup.6 1.76 × 10.sup.−4 1.50 × 10.sup.−10 DQ7 DQ5 18172 14 22.3 13.2 8.3 2.13 × 10.sup.6 1.80 × 10.sup.−3 8.59 × 10.sup.−10 DQ7 DQ5 20763 11 17.8 11.7 7.23 5.38 × 10.sup.6 6.83 × 10.sup.−4 1.27 × 10.sup.−9  DQ2 (05) DQ5 23076 280 459.9 13.4 8.17 8.15 × 10.sup.6 2.35 × 10.sup.−4 2.88 × 10.sup.−10 DQ2 (02) DQ6 18807 9.7 15.8 12.1 7.41 3.15 × 10.sup.6  1.4 × 10.sup.−3 4.58 × 10.sup.−10 DQ4 DQ5 18401 27.0 40.0 7.49 5.06 1.72 × 10.sup.6 4.29 × 10.sup.−4 2.49 × 10.sup.−10 DQ6 DQ5 21668 8.6 39.8 10.5 2.27 1.97 × 10.sup.6 5.81 × 10.sup.−4 2.95 × 10.sup.−10 DQ9 DQ6 20000 14.0 21.3 4.29 2.82 1.52 × 10.sup.6 5.93 × 10.sup.−4 3.89 × 10.sup.−10

(34) The inventors also applied this approach to a system other than the anti-HLA antibodies, namely assaying beta-2 microglobulin (B2M) in complex media. The kinetic constants of interaction of B2M with its corresponding antibody were not measured since they do not have any clinical or biological significance.

(35) The part of the process taking place on the SPR apparatus was as follows (FIG. 5): 1—Immobilization of the capture antibody by chemical coupling on the SPR chip (mouse anti-IgG2a, clone R11-89, BD Biosciences); 2—Carrying out the CFCA method:

(36) For the blank, firstly an antibody not recognizing B2M (clone OKT3, anti-CD3 antibody, BD Biosciences) was captured in a large amount on the surface, then the solution to be tested containing B2M was injected at a determined flow rate for a determined duration. Finally, after a short period of dissociation, the surface was regenerated, which made it possible to pass to a subsequent cycle.

(37) For the test itself, firstly an antibody recognizing B2M (clone B2M-01) was captured in a large amount on the surface, then the solution containing B2M was injected at the same determined flow rate for the same determined duration. Finally, after a short period of dissociation, the surface was regenerated, which makes it possible to pass to the following cycle.

(38) The analysis software of the apparatus performed the correction by subtracting the sensorgram of the blank from that of the test; this is referred to as double-referencing.

(39) The concentration of a recombinant B2M solution (Thermo Fisher) diluted to 1/10 000 in running buffer (PBS-T) was firstly measured by capture CFCA, using running buffer as blank. The concentration measured was 4.3 nM.

(40) Capture CFCA was then carried out in auto-blank in a complex medium (serum from a healthy subject, heat-treated, DTT-treated, then dialysed against PBS-T over 100 kDa membrane) used at 3 different dilutions and thereby giving rise to different levels of NSB. The NSB values recorded on the OKT3 antibody were as follows (table 8).

(41) TABLE-US-00008 TABLE 8 Amount of NSB generated by the complex media used for validating the auto-blank in complex media for assaying B2M. NSB at flow rate of NSB at flow rate of Complex medium 5 μl/min (RU) 100 μl/min (RU) Serum 1/10 342 575 Serum 1/100 71 152 Serum 1/1000 14 26

(42) The concentrations of B2M measured in these samples were as follows (table 9).

(43) TABLE-US-00009 TABLE 9 Assaying B2M by auto-blank capture CFCA within different complex media. Concentration Medium measured (nM) PBS-T 4.3 Serum 1/1000 4.0 Serum 1/100 4.0 Serum 1/10 4.7

(44) The CFCA sensorgrams are presented in FIG. 6.

(45) Materials and Methods

(46) HLA Molecules and Antibodies

(47) The monoclonal anti-HLA class I antibodies of murine origin used were: pan-class I W6/32 and anti-beta2-microglobulin (clone B2M-01) antibodies (ThermoFisher Scientific, Rockford, Ill.) and an anti-HLA-A2 antibody (One Lambda, Inc, Canoga Park, Calif.). The monoclonal anti-HLA class II antibodies of murine origin used were: pan-DQ antibody (clone Tu169) (BD Biosciences, Le Pont de Claix, France) and an anti-HLA-DQ2 antibody (One Lambda, Inc, Canoga Park, Calif.). The purified HLA molecules HLA-A*02:01 (A2), A*11:01 (A11), DQB1*02:01/DQA1*05:01 [DQ2(05)], DQB1*02:01/DQA1*02:01 [DQ2(02)], DQB1*03:01/DQA1*05:05 (DQ7), DQB1*03:03/DQA1*02:01 (DQ9), DQB1*04:01/DQA1*03:03 (DQ4), DQB1*05:01/DQA1*01:01 (DQ5) and DQB1*06:03/DQA1*01:03 (DQ6) were produced by One Lambda. An anti-CD3 antibody (clone OKT3, BD Biosciences) and mouse anti-IgG2a antibody (clone R11-89, BD Biosciences) were used.

(48) Surface Plasmon Resonance (SPR) Experiments

(49) The SPR experiments were carried out at 25° C. on a Biacore™ T200 (GE Healthcare Life Sciences, Uppsala, Sweden) with CM5 biochips (Biacore™). The sensorgrams were analyzed with the Biacore T200 Evaluation Software. The capture antibodies were immobilized with chemical coupling by amines, using a mixture of N-hydroxysuccinimide and N-ethyl-N′-dimethylaminopropyl carbodiimide according to the supplier's recommendations (GE Healthcare), after having been diluted in a solution of sodium acetate (10 mM, pH 5), followed by deactivation of the surface by injection of an ethanolamine solution (1 M, pH 8.5, GE Healthcare). A lane left without immobilization was used to perform a double correction of the sensorgrams. The peaks still present after this step were not removed since they do not affect the results. Unless indicated otherwise, the samples were prepared in PBS-T 0.05% which constituted the running buffer. The surface was regenerated by injection of a glycine solution (10 mM, pH 2.1, GE Healthcare) for 1 minute at 25 μl/min. For the tests for evaluation of non-specific binding (NSB), the samples were injected over the surface for 1 min at 25 μl/min, unless indicated otherwise.

(50) Measurement of the Active Concentration

(51) The calibration-free concentration analysis (CFCA) experiments were carried out after preliminary capture of the HLA ligands (900 s at 2 μl/min for class I, 840 s at 2 μl/min for class II) which were used in running buffer at a dilution enabling a high level of capture, in an equivalent amount for the same class. The anti-HLA antibodies were injected for 50 s at 5 μl/min then at 100 μl/min. All the samples and blanks were injected in duplicate. The coefficient of diffusion of the anti-HLA antibodies was calculated with the following formula:

(52) $D=342.3\times 1/\left(\sqrt[3]{\mathrm{MW}}\times f\times \mathrm{hrel}\right)\times {10}^{-11}$
in which MW is the molecular weight (150 000 Da), f is the coefficient of friction (1.2 for globular proteins), and hrel is the relative viscosity (0.89 at 25° C.). The criteria for validating CFCA were those recommended by Biacore™, that is to say a sufficiently large initial association slope at a slow flow rate (greater than 0.3 RU/s at 5 μl/min) and sufficiently different sensorgrams between the two flow rates, as indicated by a “QC ratio” fit of greater than 0.2. The QC ratio is the Q quotient reflecting the degree of limitation of mass transport:

(53) $Q=\frac{\mathrm{initial}\mathrm{association}\mathrm{slope}\mathrm{at}\mathrm{the}\mathrm{fast}\mathrm{flow}\mathrm{rate}}{\mathrm{initial}\mathrm{association}\mathrm{slope}\mathrm{at}\mathrm{the}\mathrm{slow}\mathrm{flow}\mathrm{rate}}\times \sqrt[3]{\frac{\mathrm{speed}\mathrm{of}\mathrm{the}\mathrm{slow}\mathrm{flow}\mathrm{rate}}{\mathrm{speed}\mathrm{of}\mathrm{the}\mathrm{fast}\mathrm{flow}\mathrm{rate}}}$

(54) Measurement of the Kinetic Parameters

(55) The kinetic parameters were determined on the same lanes as for the CFCAs, but by capturing a small amount of HLA (less than 100 RU) in order to avoid the kinetic artefacts sometimes observed with high ligand densities. The experiments were carried out at 25 μl/min using the single cycle kinetics (SCK) method: three increasing concentrations of the antibody were injected successively without regeneration between the injections. All the samples and blanks were injected in duplicate. The association and dissociation constants, k.sub.a and k.sub.d, respectively, were determined by direct adjustment of the sensorgrams according to a Langmuir 1:1 interaction model. The dissociation equilibrium constant, K.sub.D, was calculated to be equal to k.sub.d/k.sub.a.

(56) Human Sera

(57) The human sera used came from healthy subjects, from patients registered on the organ transplant wait list or patients who had received grafts. A milliliter thereof was used in different states, that is to say with or without pretreatment, with the possibility of having combined several pretreatments. The heat treatment consisted in incubation of the serum at 56° C. for 30 min. The treatment with dithiothreitol (DTT) consisted in the addition of DTT to the serum in order to reach a concentration of DTT of 5 mM then heating at 37° C. for 30 min. The serum IgGs were purified on Sepharose beads coupled to protein G (ThermoFisher Scientific), according to the supplier's recommendations. The purified compositions were concentrated using Amicon Ultra-4 10 kDa (ThermoFisher Scientific) centrifuges, twice over in order to return to a volume of 1150 μl in running buffer. The dialysis of the samples was carried out against the running buffer using Float-A-Lyzer G2 dialysis devices with a 100 kDa cut-off (Spectrum Laboratories, Rancho Dominguez, Calif.). All the samples were filtered over 0.45 μm filters before use. The total IgGs were assayed on sera treated and not treated by immunonephelometry on a BNII automated device (Siemens Healthcare Diagnostics, Marburg, Germany).

(58) SAFB Tests

(59) The sera were tested with the Luminex® Single Antigen kits after treatment in EDTA (10 mM final) according to the supplier's recommendations (One Lambda) and were analyzed on a Luminex 100® (Luminex, Austin, Tex.). The fluorescence intensities (MFI for mean fluorescence intensity) were standardized with the “baseline” formula (Fusion® software, One Lambda, Inc.).

(60) Materials and Methods for Table 2

(61) 15 000 RU of antibodies were immobilized for the B2M-01 lane (class I), and 18 000 RU for the Tu169 lane (DQ). The human sera used originated from non-anti-HLA-immunized patients registered on the transplant wait list. These sera had undergone a heat treatment then purification of the IgGs on protein G and concentration in order to obtain 1 ml of purified composition, or else heat treatment followed by treatment with DTT and dialysis against running buffer. The samples treated were diluted to ½ in running buffer with 10% final of NSB reducer (GE Healthcare).

(62) Materials and Methods for Table 3

(63) 15 000 RU of antibodies were immobilized for the B2M-01 lane (class I), and 18 000 RU for the Tu169 lane (DQ). The human sera used originated from non-anti-HLA-immunized patients registered on the transplant wait list. These sera had undergone a heat treatment then purification of the IgGs on protein G and concentration twice in order to obtain 1.15 ml of purified composition in running buffer, or else heat treatment followed by treatment with DTT and dialysis against running buffer. The A2OL (for anti-HLA-A2 One Lambda) and DQ2OL (for anti-DQ2 One Lambda) monoclonal antibodies were diluted to 1/2000 and 1/4000 in the treated sera diluted to ½ in running buffer with 10% final of background noise reducer, NSB reducer (GE Healthcare). For the A2OL, the blank was carried out by injecting this antibody over the surface on which a large amount of HLA-A11 was captured (1103+/−16 RU), and the test itself over a surface on which a large amount of HLA-A2 was captured (1014+/−18 RU). For the DQ2OL, the blank was carried out by injecting this antibody over the surface on which a large amount of HLA-DQ7 was captured (545+/−26 RU), and the test itself over a surface on which a large amount of HLA-DQ2 was captured (523+/−16 RU).

(64) Materials and Methods for Table 4

(65) Five sera from anti-HLA DQ2- or DQ7-immunized patients were heat-treated followed by a treatment with DTT and dialysis against running buffer. The IgGs were assayed in the serum before and after treatment. The Single Antigen test was carried out on serum treated with EDTA (SAFB MFI). The CFCA was carried out on these samples diluted to A in running buffer with 10% final of NSB reducer. For the patients with antibodies targeting HLA-DQ2, the blank was carried out by injecting the sample over the surface on which a large amount of HLA-DQ7 was captured, and the test itself over a surface on which a large amount of HLA-DQ2 was captured, and vice versa for patients with antibodies targeting HLA-DQ7. The levels of capture were 807+/−19 for HLA-DQ2 and 801+/−23 for HLA-DQ7, on a lane on which 20 000 RU of Tu169 were immobilized. The corrected concentration corresponds to the concentration initially present in the serum, obtained by deduction of the assays of total IgGs in the sera (initial IgG) and the treated samples (final IgG).

(66) Materials and Methods for FIG. 2

(67) A serum from an anti-HLA DQ2-immunized patient was heat-treated then underwent purification of the IgGs on protein G and concentration twice in order to obtain 1.15 ml of purified composition in running buffer. The CFCA was carried out on this sample diluted to 1/200 in running buffer with 10% final of NSB reducer. The blank was carried out by injecting the sample over the surface on which a large amount of HLA-DQ5 (818+/−2 RU) was captured. The test itself was carried out over a surface on which a large amount of HLA-DQ2 was captured (790+/−3 RU) by 15 000 RU of immobilized Tu169.

(68) Materials and Methods for FIGS. 3 and 4

(69) The SCK was carried out on the same sample as FIG. 2 at concentrations deduced from CFCA, injected in this order: 2.16 nM, 10.8 nM and 54 nM for 60 s at 25 μl/min, without regeneration between the injections. After the third injection, a 400 s period of dissociation was applied. The blank was carried out by injecting the sample over the surface on which a small amount of HLA-DQ5 (78+/−0 RU) was captured. The test itself was carried out over a surface on which a small amount of HLA-DQ2 was captured (82.5+/−0.7 RU) by 15 000 RU of immobilized Tu169.

(70) Materials and Methods for Table 7

(71) The treatment applied to the sera was exposure to heat then purification of the IgGs on protein G and concentration twice in order to obtain 1.15 ml of purified composition in running buffer. The CFCA was carried out on these samples diluted from Y to 1/200 in running buffer with 10% final of NSB reducer. The blank was carried out by injecting the sample over the surface on which a large amount of HLA-DQ not recognized by the patient's serum was captured by 10 300 RU of immobilized Tu169. The test itself was carried out over the same surface on which an amount of target HLA-DQ was captured in an equivalent amount to the antigen used for the blank. The SCKs were carried out on the same samples and analysis lanes at increasing concentrations deduced from the CFCA, for 60 s at 25 μl/min, without regeneration between the injections. After the third injection, a 400 s period of dissociation was applied. The blank was carried out by injecting the sample over the surface on which a small amount of HLA-DQ not recognized by the patient's serum was captured. The test itself was carried out over the same surface on which an amount of target HLA-DQ was captured in an equivalent amount to the antigen used for the blank.

(72) Materials and methods for FIG. 6

(73) A mouse clone anti-IgG2a antibody (R11-89, BD Biosciences) was immobilized at a level of 13 000 RU by chemical coupling on a CM5 chip. The concentration of a beta-2 microglobulin (B2M) solution diluted to 1/10 000 in running buffer (PBS-T) was firstly measured by capture CFCA, capturing 2400 RU of an antibody recognizing B2M (clone B2M-01) and using running buffer (PBS-T) as blank. The measurement of the B2M concentration was then repeated using the auto-blank method and by diluting it in 3 different complex media giving different levels of non-specific binding (“serum 1/1000”, “serum 1/100” and “serum 1/10”). The auto-blank condition consisted in the capture, at an equivalent level to the B2M-01 antibody, of an antibody of IgG2a isotype that does not bind B2M (clone OKT3, anti-CD3 antibody, BD Biosciences).