Immunogenicity Testing Guidelines

The culmination of a decade of progress

By Ronald R. Bowsher
EMD Millipore Biopharma Services

The immunogenic potential of a biotherapeutic is defined as its ability to provoke an immune response, either by humoral production of anti-drug antibodies (ADA) or through cellular-based immune responses1-3. For protein-based biotherapeutics, an immune response can range from the development of detectable but clinically insignificant ADA, to one that can impact drug safety and/or effectiveness4. Various categories of concerns and their potential clinical relevance are listed in Table 15. Because we have limited ability at this time to accurately predict the immunogenic potential of a biotherapeutic, immunogenicity testing is now integral to investigation of new biotherapeutics and follow-on biologicals.

Over the past decade, much progress has been made in gaining consensus in the overall approach for laboratory testing of ADA. The American Association of Pharmaceutical Scientists (AAPS) Ligand Binding Assay Bioanalytical Focus Group was formed in 2000, followed by formation of the Immunogenicity Working Group, under the leadership of Drs. Tony Mire-Sluis (United States Food and Drug Administration (U.S. FDA)) and Steven Swanson (Amgen). Moreover, reports of antibody-mediated pure red cell aplasia secondary to the administration of a recombinant erythropoietin had already spurred interest in neutralizing antibodies (NAb) and focused attention on immunogenicity testing7. A timeline of some key events shaping immunogenicity testing is depicted in Figure 1.

Table 1. Potential Clinical Immunogenicity Concerns as presented by S.L. Kirshner of the U.S. FDA, at the 2009 AAPS Immunogenicity Ligand Binding Assay (LBA) Training Course5. Abbreviations: PD = pharmacodynamics; PK = pharmacokinetics; CL = clearance.

In 2004, Mire-Sluis and colleagues in the AAPS Immunogenicity Working Group introduced the tiered assay approach for ADA testing (Figure 2) and defined requirements for ADA assay development, articulating the important analytical performance characteristics, providing definitions, and offering a standardized approach for computing an assay’s screening cut-point8.

In 2008, Shankar and co-workers published expanded details for ADA assays, including criteria for methods validation.9 The pre-study validation performance criteria addressed included:

1. screening cut-point

2. specificity (confirmatory) cut-point

3. sensitivity

4. selectivity/interference

5.precision

6. robustness

7. stability

8.ruggedness

The authors also made recommendations for in-study QC performance for run acceptance and presented a systematic stepwise approach for evaluating cut-point data and calculating the screening cut-point, shown in Figure 3. By emphasizing and delineating statistical methods for assay validation, the authors strove to eliminate subjectivity from, and to promote consistency of, the validation process.

Other important immunogenicity publications in the past decade focused on establishment of valid cell-based assays to detect and characterize neutralizing antibodies10, implications for regulatory agencies11,12, risk-based strategy12,13, and ADA testing in nonclinical safety studies14.

Figure 1. Timeline of key events that culminated in the draft European Medicines Agency (EMEA) guideline and FDA draft guidance, including publications that contributed to the convergence in ADA detection methodology. This timeline illustrates the involvement of both sponsors and regulatory agencies in developing a unified approach to immunogenicity testing.

Recent Draft Guidance: Overview and Highlights

Early in December 2009, the FDA published a draft guidance to industry pertaining to conduct of assay development for immunogenicity testing of therapeutic proteins15. This document followed the January 2007 publication by the EMEA Committee for Medicinal Products for Human Use (CHMP) Draft Guideline on the same topic16. These documents are complementary and quite consistent in their specific recommendations.

The new draft document provides guidance for ADA detection, confirmation, and assays for NAbs to support clinical investigation of protein therapeutics. Although the document is not yet finalized and acknowledges that ADA testing in preclinical species is not necessarily predictive of the human immune response, the guidance is a useful analytical framework that can aid in interpretation of toxicology and pharmacology data and may help reveal potential antibody-related toxicity.

The new FDA draft guidance supports an evolving assay approach with an expectation for preliminary validated assays by Phase I and full validation needed at the time of license application. The FDA Draft Guidelines recommend an analytical approach to immunogenicity testing that addresses the following key considerations:
• Sensitivity – detect clinically relevant levels of ADA
• Interference from matrix and from circulating therapeutic – ensure assay is valid for relevant clinical samples
• Physiological consequences – both NAb-related and induced hypersensitivity responses
• Risk-based application – testing strategy is case-by-case and takes into account the risk to patients of mounting an immune response to a therapeutic protein

Figure 2. Tiered approach for screening, characterization and quasi-quantitative titer assessment of serum samples for the presence of ADA. Tier 1 refers to initial antibody detection with assay sensitivity of at least 250 – 500 ng/mL for a surrogate antibody. “CP” indicates cut-point. Tier 2a tests for competitive inhibition after addition of excess therapeutic. Depending on the risk-based plan, additional studies may be appropriate to characterize the ADA in terms of the types of antibodies and their binding characteristics (Tier 2B). The final step, Tier 3, involves serial titering of positive samples to provide a quasi-quantitative estimate of the concentration of antibody in the test sample8.

Consistent with previously published white papers and the EMEA draft guidance, the FDA recommends a multi-tiered assay approach as shown in Figure 2, with the following additional criteria:

1. Tier 1 screening assay should:
• Exhibit 5% false positive rate to maximize detection of true positive samples
• Be able to detect all immunoglobulin classes and IgG subclasses.
• Ensure that labeling of the detection reagent should not obscure important binding epitopes
• Involve careful selection of the assay buffer and blocking reagents used to prevent nonspecific binding
• Rule out matrix interference via selectivity experiments, in which different lots of matrix are analyzed after spiking with zero, low, and high concentrations of surrogate antibody
• When diluting samples to minimize matrix interference, evaluate minimal required dilution (MRD) from a panel of 10 or more samples
• Should not dilute samples more than 1:100 so as not to compromise sensitivity

2. Tier 1 screening assay should be validated for sensitivity, specificity, and precision:
• Sensitivity
– Assess sensitivity using a preparation of purified antibodies
– Sensitivity = interpolated concentration at the predetermined cut-point response
– Determine sensitivity in test sample matrix at MRD
– Required sensitivity = 250 – 500 ng/mL
• Specificity
– Critical for interpretation of immunogenicity results
– If therapeutic is related to an endogenous protein, assess antibody cross-reactivity with both proteins
– If therapeutic is in a family of homologous proteins, assess antibody cross-reactivity with all family members
• Precision
– Measure intra-day precision: six replicates per plate
– Measure inter-day precision: three replicates per day for three days
– Measure inter-operator variability when appropriate

3. Tier 3 quasi-quantitative result should be reported as a titer value, as opposed to a result in mass units, following interpolation against a standard curve.

4. The assay cut-point needs to be determined systematically using a statistically valid approach (as in Figure 3), removing outliers to lessen analytical variance. To estimate the cut-point, 50–100 presumed negative samples should be screened multiple times across multiple assays, using a balanced statistical design. The cut-point may need to be re-established for different patient populations.

5. For the preparation of in-study cut-point assay controls, the FDA guidelines make the following recommendations:
• Purify surrogate ADA from animal sera and spike into the human sample matrix
• For direct-binding ADA, use a primate surrogate ADA to prepare QC samples
• Establish a negative control for pre-study validation and in-study sample analysis
• The positive controls should reflect the human immune response
• Prepare QC samples at low, mid and high values in the ADA assay
• The low QC is set statistically so its failure rate is approximately 1%

Figure 3. Systematic stepwise approach for data-driven determination of a screening cut-point, as presented by Shankar and colleagues9. Key elements in the statistical analysis are data normality, outlier evaluation, need for transformation, and evaluation of run means and variances9. Outliers can result from nonspecific interactions or preexisting antibodies.

The new draft guidance provides some recommendations for collection of sera from patients for detection and characterization of ADA. First, pre-exposure samples should be obtained from all patients to provide baseline information. For detection of IgM, collect samples seven to 14 days post-exposure. In contrast, serum samples should be collected four to six weeks post-exposure for detection of IgG ADA. Secondly, test samples need to be collected when there will be minimal interference from the therapeutic protein. If drug-free samples cannot be obtained during the study, consider sampling at approximately five half-lives.

In its recommendation for assessments of NAbs, the FDA strongly recommends cell-based bioassays. Cell-based assays are believed to be more reflective of in vivo immunogenicity than competitive ligand-binding assays (LBAs), despite the higher variability and limited quantitative ranges. If a therapeutic protein possesses multiple domains, it may be appropriate to consider several NAb assays. Tier 2 confirmatory assays, involving competition or immunodepletion, are critical for NAb assays. The NAb cut-point should be determined statistically in a systematic manner similar to the Tier 1 screening assay. In addition, the sensitivity of NAb assays are estimated in a manner similar to screening assays and should be reported in terms of mass units (ng/mL).

Outsourcing Assay Development

Formation of host antibodies to therapeutic proteins can produce a range of clinical sequelae, from no effect to very harmful ones. This spectrum of immunogenic responses, combined with our inability to reliably predict the human immune response via experimentation or computation, necessitates the development of effective, sensitive, and reproducible immunoassays for testing the immunogenicity of new therapeutics. The design of such assays can be challenging, especially for new, innovative products. In addition, the suitability of the assays may need to be refined during the product development lifecycle to ensure their continued suitability. Implementation of an analytical strategy to support immunogenicity testing is outlined in a ‘risk-based plan’ that weighs the risk to patients of mounting an immune response to a therapeutic protein.

For all of these reasons, it may be advantageous for drug developers to partner with an experienced organization. Such a partner should have extensive experience and knowledge in developing immunogenicity programs. They should use a tiered method of immunogenicity testing (Figure 2) and possess expertise in the detection, characterization, and quantification of ADA. Commonly used immunoassay platforms to detect ADA include ELISA, EIA, RIPA, SPR (Biacore™), ECL, DELFIA® and cell-based assays. An efficiently managed service organization should provide careful in-study quality control data with patient results, and engage in good and timely communications with the sponsor.

References

1.Schellekens H. (2002) Immunogenicity of therapeutic proteins: clinical implications and future prospects. Clin Therapeutics 24(11): 1720-40.
2.Koren E et al. (2002) Immune responses to therapeutic proteins in humans – clinical significance, assessments, and prediction. Curr Pharm Biotechnol 3:349-60
3.Schellekens H. (2003) The immunogenicity of biopharmaceuticals. Neurology 61: S11-12.
4.Woodcock J. (2007) Testimony U.S. House of Representatives – Assessing the impact of a safe and equitable biosimilar policy in the US. http://www.hhs.gov/ asl/testify/2007/05/t20070502h.html
5.Kirshner SL. (2009) AAPS Immunogenicity LBA Training Course, Seattle WA (June 2009).
6.Miller KJ, Bowsher RR, Celniker A, et al. (2001). Workshop on bio-
analytical methods validation for macromolecules: summary report. Pharmac Res 18(9): 1373-83.
7.Casadevall N, Nataf J, Viron B, et al. (2002) Pure red-cell aplasia and antierythropoietin antibodies in patients treated with recombinant erythropoietin. NEJM 346(7): 469-75.
8.Mire-Sluis AR, Barrett Y-C, Devanarayan V, et al. (2004) Recommendations for the design and optimization of immunoassays used in the detection of host antibodies against biotechnology products. J Immunol Methods 289: 1-16.
9.Shankar G, Devanarayan V, Amaradvadi L, et al. (2008) Recommendations for the validation of immunoassays used for detection of host antibodies against biotechnology products. J Pharm Biomed Anal 48: 1267-81.
10.Gupta S, Indelicato SR, Jethwa V, et al. (2007) Recommendations for the design, optimization and qualification of cell-based assays used for the detection of antibody responses elicited to biological therapeutics. J Immunol Methods 333(102): 1-9.
11.Shankar G, Shores E, Wagner C, and Mire-Sluis AR. (2006) Scientific and regulatory considerations on the immunogenicity of biologics. Trends Biotechnol 24(6): 274-80.
12.Chamberlain P and Mire-Sluis AR. (2003) An overview of scientific and regulatory issues for the immunogenicity of biological products. Dev Biol (Basel) 112: 3-11.
13.Koren E, Smith HW, Shores E, et al. (2008) Recommendations on risk-based strategies for detection and characterization of antibodies against biotechnology products. J Immunol Methods 333(1-2): 1-18.
14.Ponce R, Abad L, Amaravadi L, et al. (2009) Immunogenicity of biologically-derived therapeutics: assessment and interpretation of nonclinical safety studies. Regulatory Tox Pharmacol 54: 164-82.
15.U.S. Dept. Health and Human Services, FDA (CDER, CBER). Guidance for Industry – assay development for immunogenicity testing of therapeutic proteins (Dec. 2009).
16.European Medicines Agency (EMEA), Committee for Medicinal Products for Human Use. (2007) Guideline on immunogenicity assessment of biotechnology-derived therapeutic proteins.

DELFIA is a registered trademark of Perkin Elmer, Inc. Biacore is a trademark of GE Healthcare Companies.

Ronald R. Bowsher, Ph.D. is chief scientist of EMD Millipore Biopharma Services. Dr. Bowsher is also a co-founder of the Ligand Binding Assay Bioanalytical Focus Group within AAPS, a group that dedicates its efforts to promoting harmonization and education for the development, validation and application of ligand binding assay methods for the bioanalysis of biotherapeutics, anti-drug antibodies and novel biomarkers. He can be reached at ron_bowsher@millipore.com