Methods for detecting and/or measuring anti-drug antibodies, in particular treatment-emergent anti-drug antibodies

Abstract

Described herein are methods, assays and techniques for detecting and/or measuring anti-drug antibodies that bind to a protein, polypeptide or other compound or molecule that comprises at least one immunoglobulin variable domain with an exposed C-terminal region.

Claims

1. Method for detecting and/or measuring in a sample anti-drug antibodies that bind to a polypeptide that comprises at least one variable domain of the heavy chain of a heavy-chain antibody (VHH domain), humanized VHH domain, or camelized heavy chain variable domain of a conventional antibody (VH domain) with an exposed C-terminal region with the sequence VTVSS (SEQ ID NO: 1), said method comprising at least the steps of: a) contacting said sample with a capturing agent that is immobilized on a support, wherein said capturing agent is or essentially consists of said polypeptide, under conditions such that anti-drug antibodies against said polypeptide can bind to said capturing agent to form a complex of the capturing agent and any captured anti-drug antibodies; b) optionally removing any components or constituents present in said sample that do not bind to the capturing agent; c) detecting or measuring any anti-drug antibodies that have bound to the capturing agent, by contacting the complex of the capturing agent and the captured anti-drug antibodies with a detection agent, under conditions such that said detection agent can bind to the complex of the capturing agent and the captured anti-drug antibodies, wherein said detection agent is or essentially consists of: (i) said polypeptide; (ii) a detectable tag or label bound to said polypeptide either directly or via a suitable linker and (iii) 1-5 amino acid residues that are linked to the exposed C-terminal end with sequence VTVSS (SEQ ID NO: 1) of the VHH domain, humanized VHH domain, or camelized VH domain, in which said sample is a sample of whole blood, serum, plasma, ocular fluid, bronchoalveolar fluid/bronchoalveolar lavage fluid (BALF), cerebrospinal fluid or another biological fluid.

2. Method according to claim 1, in which the detection agent has the amino acid sequence VTVSS(X).sub.n(SEQ ID NO:2) at its C-terminal end, in which n is 1 to 5; and in which each X is a naturally-occurring amino acid residue that is independently chosen.

3. Method according to claim 1, in which the sample has been obtained from a subject to which said polypeptide has been administered and wherein the polypeptide has been administered to a subject according to a regimen that is such that there is a risk or possibility that anti-drug antibodies against the polypeptide have been raised in the subject to which said polypeptide has been administered.

4. Method according to claim 1, in which the polypeptide has a half-life of at least 1 day in a human subject.

5. Method according to claim 1, in which the VHH domain, humanized VHH domain, or camelized VH domain with the exposed C-terminal region with the sequence VTVSS (SEQ ID NO: 1) is present at the C-terminal end of said polypeptide.

6. Method according to claim 1, in which the detectable tag or label comprised within the detection agent is a tag or label that can be detected using electrochemiluminescence techniques.

7. Method according to claim 2, wherein each X is independently chosen from the group consisting of alanine (A), glycine (G), valine (V), leucine (L) or isoleucine (I).

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) The invention will now be illustrated by means of the following non-limiting examples and Figures, in which:

(2) FIG. 1 schematically shows the principles of prior art bridging assay formats (such as ELISA-bridging format or ECL bridging format) for detecting ADA's against a conventional antibody.

(3) FIGS. 2A to 2D schematically show the use of an assay of the invention for detecting ADA's against a bivalent ISV construct, using an alanine-extended detection agent.

(4) FIG. 3 schematically show the use of an assay of the invention for detecting ADA's against a monovalent ISV.

(5) FIGS. 4A to 4C schematically show the use of an assay of the invention for detecting ADA's against a bivalent ISV construct, using an alanine-extended capturing agent.

(6) FIG. 5 is a table summarizing the different assay formats that were compared in Example 1.

(7) FIG. 6 shows the screening and confirmatory results of 171 test samples obtained in Example 2 for a conventional homogeneous ECL based bridging format (prior art) and the modified bridging format of the invention using an alanine-extended detection agent.

(8) FIG. 7 shows a comparison of the conventional homogeneous ECL-based bridging format with the modified homogeneous ECL-based bridging format in the absence of pre-existing antibodies.

(9) FIG. 8 shows the result of a comparative evaluation of the impact of interference on detection of ADA in a conventional bridging ADA assay versus the modified bridging ADA assay. The graphs show the detection of spiked positive control in interference containing samples, for both the conventional assay (Panel A) and the modified assay (Panel B).

(10) Experimental Part

EXAMPLE 1

Comparison of Different Assay Formats

(11) 5 different ADA assay formats for detecting ADA's (schematically shown in FIG. 5) were evaluated for the impact of protein interference on the respective assay performance (when used for measuring ADA's against a Nanobody), using a set of 171 test samples comprised of plasma or serum samples from 121 healthy volunteers and 50 rheumatoid arthritis patients (none of whom had previously been exposed to Nanobodies). The Nanobody-based protein used was SEQ ID No. 417 from WO06/122786.

(12) Each of the assay formats was evaluated on the prevalence and magnitude of interference in the respective assays as well as the impact of any interference on assay performance, including screening and confirmatory cut-point setting, assay sensitivity and the extent to which it was possible to reliably detect treatment emergent ADA in presence of interference.

(13) It was found that the modified ECL-based bridging format provided/allowed for: (i) assay sensitivity of less than 500 ng/ml, as recommended by the applicable guidelines for ADA assays: (ii) suitable cut-point setting (both screening and confirmatory) using appropriate statistics, again as recommended by standard guidelines for ADA assays; and (iii) even relatively low amounts of “true” ADA's could be detected in samples tested that (also) contained interfering factors (it is also expected that detection of treatment-emergent ADA's will still be possible at relatively high levels of interfering factors, although at very high levels sensitivity may be (somewhat) reduced

(14) All other four assay formats were impacted to a much larger degree by interference, such that detection of ADA was severely impeded in the majority of interference containing samples. Only in samples presenting low/very low levels of interference, could ADA's be reliably detected.

(15) The amount of ADA that could be detected is further dependent on the sensitivity of the assay: the sequential ECL based bridging format demonstrated the lowest sensitivity (estimated to be more than 5 μg/ml), whereas the sequential ELISA-based bridging format and the direct ELISA format were found to have a sensitivity estimated to be 600 ng/ml, which is higher than the sensitivity recommended by applicable guidelines. Also, for all assay formats tested, it was found that there is a certain degree of correlation between the level of interference (or baseline ECL signal) and the amount of ADA that can be detected. This can make it possible to define threshold values (based on ECL read outs) for maximum interference levels at which low levels of ADA's can still be reliably detected.

(16) The results are also summarized in Table 1.

(17) TABLE-US-00001 TABLE 1 overview of different assay format evaluated for impact of interference on assay performance. Detection of est. Sensitivity Cut-point 620 ng/ml pAb ADA assay format (+IF) setting (+IF) positive control Homogenous ECL-based OK Not OK Not OK bridging assay Sequential ECL-based Not OK OK Not OK bridging assay Sequential ELISA-based Not OK Not OK Not OK bridging assay Direct ELISA Not OK Not OK Technically not feasible Homogenous ECL-based OK OK OK modified bridging assay

EXAMPLE 2

Comparison of Homogeneous ECL-Based Bridging Format Using a C-Terminally Modified Detection Agent With a Homogeneous ECL-Based Bridging Format Using a Non-Modified Detection Agent

(18) Two homogeneous ECL-based bridging formats were compared. In one format, the capturing agent (biotinylated for binding to the support) and the detection agent (sulfo-tagged for ECL detection) had essentially the same sequence (SEQ ID NO. 417 from WO 06/122786). In the other format, the same capturing agent and detection agent were used as were used for the first format, but the detection agent had been modified by adding one additional alanine residue to the C-terminus of the Nanobody-based protein sequence.

(19) The same 171 test samples used in Example 1 were again tested in a screening and confirmatory set up.

(20) When the homogeneous modified bridging format according to the invention was used, it was found that the number of test samples presenting high ECL signals was greatly reduced (FIG. 6). Furthermore, variation in ECL-signals is less; varying from 180 to <2,500. A commonly used statistical method (box plot analysis for outlier identification) could be used to successfully define cut-points on the screening and confirmatory data set.

(21) Panel A in FIG. 6 shows the screening results of the 171 test samples in both assay formats. Screening cut-points are indicated by dotted lines. Representative samples having high interference levels in the homogeneous ECL-based bridging assay (which samples were also used in experiments referred to in Example 1) are indicated by stars or squares respectively, with all other samples being indicated by dots.

(22) Panel B in FIG. 6 shows results of the confirmatory set up compared to screening values of the homogeneous ECL-based bridging assay: screening results (ECL signal) were plotted versus the % reduction obtained in the confirmatory assay. A positive correlation is found between the screening ECL signal and % reduction obtained in the confirmatory assay.

(23) The obtained confirmatory results show a continuous distribution (i.e., % reduction varies from <20% to >90% reduction), indicating that the interference is present in varying amounts in many test samples. This makes it difficult to define a distinct interference-negative population which is a prerequisite for commonly used statistical methods defining screening and confirmatory cut-points. As screening and confirmatory cut-point setting is based on the 95th percentile/confidence interval and 99th percentile/confidence interval, respectively, of the response of the blank, non-treated population, the calculated cut-point will be highly dependent on the distribution of the ECL signals within the validation sample set and therefore cut-point setting becomes arbitrary.

(24) Panel C in FIG. 6 shows results from screening and confirmatory results in the modified bridging format (i.e. using an alanine-extended detection agent): screening results (ECL signal) were plotted versus the % reduction obtained in the confirmatory assay. The number of test samples showing high screening ECL values is lower, allowing cut-point setting via commonly used statistical methods.

(25) The data from FIG. 6 shows that whereas a conventional ECL bridging format is significantly impacted by interference, the modified bridging assay using an alanine-extended capturing agent is impacted to a far lesser extent. Also, whereas for a conventional bridging ADA assay format the setting of the cut point would to a large extent be more or less arbitrary, relevant cut points could be set for the modified bridging ADA format using commonly used statistical methods.

(26) Also, the sensitivity of both assay formats in the absence of interference was compared, using data obtained for a positive control serum (rabbit serum raised against SEQ ID NO. 417 from WO06/122786) as a reference. In the absence of interference, the sensitivity of both assays was comparable as is shown in FIG. 7. This confirms that the differences between the two assay formats in the presence of interference are indeed due to the presence of the interference.

(27) The results above were further confirmed by a separate experiment in which different test samples with varying interference levels were spiked with the rabbit positive control serum used to obtain the data shown in FIG. 7 and were evaluated in both assay formats (results shown in FIG. 8). Again, it was found that for the conventional bridging assay format, it was not possible to (reliably) detect ADA's in samples with high interference level; whereas for the modified bridging format involving the use of an alanine-extended capturing agent it was found that this assay could successfully detect low amounts of ADA in the presence of interference (see the data shown in FIG. 8). The bars indicate the ECL signal of different matrix samples spiked with positive control polyclonal serum spiked at concentrations of 0 and 620 and 5580 ng/ml. The black bars represent results obtained with interference-negative matrix. The gray bars numbered 1-6 depict ECL signals of the positive control pAb spiked at 3 different concentrations into example samples containing varying amounts of interference. Panel A shows the results obtained in the homogeneous bridging assay and panel B the results obtained in the homogeneous modified bridging assay. The black horizontal line shows an example of suitable threshold ECL values for each of these assays, i.e. the ECL value of the test sample that still allows detection of an acceptable amount of ADA in the presence of interference).