Inhibition of interleukin-5 induced false positive anti-drug antibody responses against mepolizumab through the use of a competitive blocking antibody.

Mepolizumab, a humanized IgG1 monoclonal antibody that blocks native homodimeric interleukin-5 (IL-5) from binding to the IL-5 receptor, has recently been approved for treatment of severe eosinophilic asthma. Our initial immunogenicity assay method for phase I and II studies utilized a bridging electrochemiluminescence format with biotin and ruthenium-labelled mepolizumab linked by anti-drug antibodies (ADA). We discovered that IL-5 significantly increased in dosed subjects from a phase II study and that the increased IL-5 was in the form of a drug-bound complex. We demonstrated that the elevated drug-bound IL-5 produced false-positive response in the in vitro ADA assay, in which drug-bound IL-5 dissociated and then bridged mepolizumab conjugates to yield positive signal. To eliminate the IL-5 interference, we compared two strategies: a solid-phase immunodepletion of IL-5 and an in-solution IL-5 immunocompetition. We identified the best competitive antibody for each purpose. We found both methods demonstrated similar effectiveness in reducing the false positive signal in IL-5 spiked samples; however, the in-solution immunocompetition for IL-5 had fewer false positives in study samples. Additionally, the in-solution immunocompetition method was experimentally simpler to execute. We modified the ADA assay by adding a pre-treatment step with a mepolizumab competitive anti- IL-5 antibody. Using this new method, we retested clinical samples from two phase II studies (MEA112997 and MEA114092). The confirmed ADA positive incidence was reduced from 29% and 61% to 1% and 8% with the modified in-solution immune inhibition method. Target interference is a fairly common problem facing immunogenicity testing, and target-induced false positive cannot be distinguished from true ADA response by the commonly used drug competitive confirmation assay. The approach and method used here for resolving target interference in ADA detection will be useful for differentiating between a true ADA response and target induced false positive as well as similar challenges in other programs.

Development of an ultra-sensitive single molecule counting assay for the detection of interleukin-13 as a marker for asthmatic severity.

OBJECTIVE: Interleukin-13 (IL-13) has been difficult to quantify within human serum due to low abundance. Available assays have not been sensitive enough to detect IL-13 at the femtogram level. Thus, there are inconsistencies within the published literature as to IL-13 concentrations in normal or disease populations. To better understand IL-13 serum concentrations, a highly sensitive immunoassay was developed and used to determine concentrations from asthmatics with varying clinical severities. METHODS: A single molecule counting (SMC) method was used to analyze serum samples from a total of 145 individuals (60 severe asthma, 60 moderate asthma, 60 mild asthma and 23 healthy donors). RESULTS: IL-13 concentrations correlated with severity of asthma, with overlapping ranges. Mean IL-13 levels were highest in severe asthma. Mean IL-13 levels in moderate asthma population were second highest followed by mild asthma with the lowest IL-13 concentration. IL-13 concentrations in healthy donors were similar to the mild asthmatic population. The average concentrations of IL-13 in severe, moderate, mild and healthy donors were 1.286pg/mL, 0.672pg/mL, 0.508pg/mL and 0.155pg/mL respectively. CONCLUSION: Severe asthma patients have elevated levels of IL-13.

False-positive immunogenicity responses are caused by CD20+ B cell membrane fragments in an anti-ofatumumab antibody bridging assay.

An electrochemiluminescent (ECL) bridging assay to detect anti-ofatumumab antibodies (ADA) in human serum samples was developed and validated. Using this assay format, clinical samples were first screened to identify potential ADA positive samples, which were then further tested by adding excess drug, confirming the positive signals as drug specific. However, when the method was implemented into clinical studies for ADA testing, a high positive rate was observed in the pre-dose samples collected from patients with chronic lymphocytic leukemia (CLL). Since the positive signals were not associated with ofatumumab (Ofa) treatment, and diminished after treatment, it was suspected that matrix interference might be responsible, resulting in false-positive responses. We performed a series of experimental investigations to identify, characterize, minimize or eliminate the possible false-positive responses. One possible source was identified to be CD20 (the target of Ofa) present on cell membrane fragments (CMFs). The false-positive responses caused by CD20(+) CMFs could be reduced by solid-phase immunodepletion, ultracentrifugation, or inhibited by adding another anti-CD20 antibody (rituximab). As a consequence, the ADA method was modified to minimize the matrix interference caused by CD20(+) CMFs and, then, validated for sample testing.

Development of an enzymatic assay for the detection of neutralizing antibodies against therapeutic angiotensin-converting enzyme 2 (ACE2).

Therapeutic proteins have the potential to elicit immune responses in animals and humans (Mire-Sluis et al., 2004; Yu et al., 2006; Shankar et al., 2008). Contributors to the response could include product related factors such as chemical modifications, impurities that co-purify with product, contaminants, formulation, aggregates, and clinical factors such as dose concentration, dosing frequency, route of drug administration, rate of administration, patient underlying disease, concomitant medication, and genetic status among others (Patten and Schellekens, 2003). Further, an immune response triggered by a therapeutic enzyme may neutralize the endogenous counterpart resulting in a decrease or depletion of the therapeutic and endogenous enzymes imposing safety concerns for patients. Therefore, monitoring of anti-drug antibody (ADA) and neutralizing antibody (NAb) responses to both the recombinant therapeutic enzyme and endogenous enzyme is important during early development and subsequent clinical studies. Testing considerations for NAb detection against therapeutic enzymes have been published mostly for lysosomal storage diseases (Wang et al., 2008). NAb cross-reactivity to the endogenous counterpart has also been characterized (Sominanda et al., 2010). Here, we describe an enzymatic NAb assay which detects neutralizing antibodies to both recombinant and endogenous angiotensin-converting enzyme 2 (ACE2). NAb assay sensitivity was optimized by selecting the assay incubation time as 20 min with an enzyme concentration of 0.5 µg/mL. Four anti-ACE2 antibodies out of a commercial panel of 18 were found to have neutralizing capabilities based upon their ability to abrogate ACE2 enzymatic activity. We demonstrated assay specificity by small peptide inhibitors specific for ACE or ACE2. DX600, an ACE2 specific inhibitor did not cross-react with ACE. Conversely, captopril, an inhibitor of ACE did not inhibit ACE2. The assay specificity for ACE2 neutralizing antibodies was further demonstrated by the lack of reactivity of two species control antibodies and 14 anti-ACE2 antibodies. Moreover, we demonstrated assay specificity to human endogenous ACE2 from human epithelial cells. Three human cell lines (Calu-3, Caco-2, Huh-7) were evaluated for the cell surface expression of ACE2 by flow cytometry and Western blot. Subsequently, whole cell lysates, cell culture supernatant, and live cells were evaluated in the assay. Results demonstrated that Calu-3 had elevated levels of ACE2 compared to Caco-2 or Huh-7. Calu-3 also demonstrated elevated ACE2 enzymatic activity in all three sources and could be inhibited by the ACE2 specific inhibitor DX600 as well as the neutralizing antibodies for the recombinant ACE2. Thus, we describe here a method to detect NAb against a therapeutic enzyme and assess NAb cross-reactivity to the native endogenous enzyme. The approach of method development described here could be applied for the assessment of NAb responses to other enzymatic therapeutics.

Development and validation of a cell-based SEAP reporter assay for the detection of neutralizing antibodies against an anti-IL-13 therapeutic antibody.

A cell-based bioassay capable of detecting neutralizing antibodies (NAb) specific to a therapeutic anti-IL-13 monoclonal antibody was developed, validated and used to analyze normal human and asthma serum samples. At the time of this study, a neutralizing assay was unavailable for anti-IL-13 antibody therapeutics with sufficient rigor for validation. Thus, we describe here a method and considerations for validation. The assay used IL-13 responsive HEK293 cells transfected with a secreted embryonic alkaline phosphatase (SEAP) reporter gene. Cells were plated at 5.4×10(4) per assay well due to 90% confluence on the subsequent day. Optimal IL-13 and anti-IL-13 concentrations were determined to be 600 pg/mL and 900 ng/mL respectively. We demonstrated the assay's cut point, sensitivity, specificity/cross reactivity, selectivity/matrix interference, and precision. Also, we demonstrated how the drug inhibitory concentration (IC(50), IC(75), and IC(90)) can affect sensitivity and dynamic range/assay window. We characterized the differences in assay response between serum samples of normal population and asthma population. Asthma samples demonstrated an elevated OD ratio in average compared to normal samples. Thus, separate cut points were needed and calculated to be 1.78 and 2.43 for normal and asthma serum, respectively. The assay sensitivity was 670 ng/mL with the positive control (affinity purified rabbit anti-drug polyclonal antibodies). Potential false positives resulting from endogenous serum cytokines including IL-13, IL-4, and Interferon alpha (INF-α) were evaluated and the results indicated that the interfering concentrations for these cytokines are much higher than the respective physiological concentrations. Based on these data, the risk of false positive by endogenous cytokines was considered to be low. In addition, irrelevant anti-drug positive control antibodies were evaluated for assay specificity and did not demonstrate neutralizing capability. Further, no matrix interference in the intended patient population was found when using a final assay serum concentration of 16.7%. The validated assay had acceptable intra- and inter- assay precision in that all %CVs were ≤25%. Overall, this assay successfully proceeded through validation and was used to determine NAb responses within serum samples.

Alloxan is an inhibitor of O-GlcNAc-selective N-acetyl-beta-D-glucosaminidase.

We have previously shown that streptozotocin (STZ) inhibits O-GlcNAc-selective N-acetyl-beta-d-glucosaminidase (O-GlcNAcase), the enzyme that removes O-GlcNAc from proteins. In light of this observation, we explored the possibility that the diabetogenic toxin alloxan, an O-GlcNAc transferase (OGT) inhibitor, might also inhibit O-GlcNAcase. Alloxan inhibited islet O-GlcNAcase with a dose-response much like that of STZ. Similar to STZ, islet O-GlcNAcase was more susceptible to alloxan inhibition than was brain O-GlcNAcase. Alloxan directly inhibited recombinant O-GlcNAcase activity with a dose-response very similar to that of STZ. Subsequent LC/MS/MS analysis revealed that alloxan modified the tryptic digest pattern of the enzyme. One tryptic peptide LGCFEIAK(894-901) was modified by alloxan. Two other tryptic peptides, LDQVSQFGCR(158-167) and SFALLFDDIDHNMCAADK(168-185), both N-terminal active site peptides, were absent after alloxan treatment. Together, these data demonstrate that alloxan is an inhibitor of O-GlcNAc-selective N-acetyl-beta-d-glucosaminidase, with inhibition corresponding to an altered tryptic digest pattern of N-terminal active site peptides.

The diabetogenic antibiotic streptozotocin modifies the tryptic digest pattern for peptides of the enzyme O-GlcNAc-selective N-acetyl-beta-d-glucosaminidase that contain amino acid residues essential for enzymatic activity.

Streptozotocin (STZ) inhibits O-GlcNAc-selective N-acetyl-beta-d-glucosaminidase (O-GlcNAcase), the enzyme that removes O-GlcNAc from proteins. The active site of the enzyme was recently proposed to include aspartates 174, 175, and 177, with STZ inhibition via a transition state analog. We explored the effect of STZ on the tryptic peptide digest pattern of O-GlcNAcase. LC/MS/MS analysis demonstrated that STZ modified two areas of the enzyme. One peptide, LGCFEIAK (894-901), in a C-terminal region previously proposed to possess O-GlcNAcase activity, was methylated by STZ. Another peptide, EYEIEFIYIASPGLDITFSNPK (128-149), was detected only after treatment with STZ and was in an N-terminal region, overlapping a glutamate-rich area containing an adjacent phenylalanine residue. No covalent modification of this peptide could be demonstrated. Detection of this peptide after treatment with STZ was accompanied by the simultaneous inability to detect the nearby peptide KLDQVSQFGCR (157-167), which contains a cysteine residue recently shown to be essential for enzymatic activity. To determine which of the first two peptides might also be important for O-GlcNAcase activity, site-specific mutagenesis was performed. Mutation of the N-terminal phenylalanine and serine residues resulted in almost complete inhibition of activity. In contrast, mutation of conserved C-terminal glycine and cysteine residues caused little inhibition of enzymatic activity. Together, these data extend the region of the active site N-terminally and give independent evidence to support the idea that STZ inhibits O-GlcNAcase through formation of a transition state analog that resides in the active site of the enzyme and in doing so alters its conformation and ensuing tryptic digest pattern.

Differential reactivity of anti-primate and anti-human secondary antibodies against human and monkey immunoglobulins: implications for determining the sensitivity of immunogenicity assays.

Development of immunogenicity assays for assessment of human antibodies to therapeutic proteins requires a quantitative determination of assay sensitivity. In the absence of true human positive controls, this is usually accomplished by utilizing affinity-purified antibodies from non-human primates or monoclonal antibodies. In the former case, it is generally considered that non-human primate antibodies will be recognized equally to human antibodies by secondary anti-human immunoglobulin reagents used in immunogenicity assays. We present results here demonstrating that this is not the case. In reality, anti-human immunoglobulin secondary antibodies do not recognize primate immunoglobulins as well as human immunoglobulins. As a result, the use of affinity purified primate antibodies to determine the sensitivity of an immunogenicity assay will likely result in the true sensitivity of the assay being underestimated.

Effect of double antigen bridging immunoassay format on antigen coating concentration dependence and implications for designing immunogenicity assays for monoclonal antibodies.

The double antigen bridging immunoassay has been used extensively for detection of immunogenicity responses to therapeutic monoclonal antibodies. We have analyzed parameters affecting performance of this type of immunoassay including microtiter plate antigen coating concentration, enzyme-labeled antigen conjugate dilution and assay format (one-step versus two-step). We present results demonstrating that the format of the assay has a significant impact on the optimal parameters to maximize assay performance. A one-step assay format achieves maximal sensitivity across a broad range of coating concentrations and at a lower concentration of conjugate than that in a two-step format. In contrast, a two-step format requires very low coating concentrations and higher conjugate concentrations to achieve maximal sensitivity and suffers from significantly reduced sensitivity at higher coating concentrations. Together, these findings indicate that a one-step assay format can greatly reduce the effect of coating concentration variation on assay performance.

Glucose stimulates the association of Crk with p130Cas in pancreatic beta cells.

OBJECTIVES: Previously, we demonstrated glucose-induced beta-cell tyrosine phosphorylation of p130Cas, a protein containing 15 YXXP repeats that can become tyrosine phosphorylated and bind Src-homology 2 (SH2)-containing proteins. In light of the importance of p130Cas in other cell types, we determined which beta-cell proteins exhibited glucose-induced association with p130Cas. METHODS: beta cells were stimulated with glucose and/or the muscarinic agonist carbachol to determine which SH2-containing adapter proteins underwent glucose-induced association with p130Cas. RESULTS: The SH2-containing adapter protein Crk underwent glucose-induced association with p130Cas, while other SH2-containing proteins such as grb2, PI3 kinase, Shp-2, paxillin, and pyk2 did not. Glucose-induced Crk-p130Cas association was rapid and sustained and was maximal with the combination of glucose and carbachol, paralleling insulin secretion. There was no increased tyrosine phosphorylation of Crk itself. The expression of Crk in isolated rat islets was also demonstrated. CONCLUSION: beta cells contain the SH2-containing adapter protein Crk, which undergoes glucose-induced association with p130Cas.

Glucose stimulates the tyrosine phosphorylation of Crk-associated substrate in pancreatic beta-cells.

Several years ago, we demonstrated that glucose induced tyrosine phosphorylation of a 125-kDa protein (p125) in pancreatic beta-cells (Konrad, R. J., Dean, R. M., Young, R. A., Bilings, P. C., and Wolf, B. A. (1996) J. Biol. Chem. 271, 24179-24186). Glucose induced p125 tyrosine phosphorylation in beta-TC3 insulinoma cells, beta-HC9 cells, and in freshly isolated rat islets, whereas increased tyrosine phosphorylation was not observed with other fuel secretagogues. Initial efforts to identify p125 were unsuccessful, so a new approach was taken. The protein was purified from betaTC6,F7 cells via an immunodepletion method. After electrophoresis and colloidal Coomassie Blue staining, the area of the gel corresponding to p125 was excised and subjected to tryptic digestion. Afterward, mass spectrometry was performed and the presence of Crk-associated substrate (Cas) was detected. Commercially available antibodies against Cas were obtained and tested directly in beta-cells, confirming glucose-induced tyrosine phosphorylation of Cas. Further experiments demonstrated that in beta-cells the glucose-induced increase in Cas tyrosine phosphorylation occurs immediately and is not accompanied by increased focal adhesion kinase tyrosine phosphorylation. Finally, it is also demonstrated via Western blotting that Cas is present in normal isolated rat islets. Together, these results show that the identity of the previously described p125 beta-cell protein is Cas and that Cas undergoes rapid glucose-induced tyrosine phosphorylation in beta-cells.