2023 White Paper on Recent Issues in Bioanalysis: ISR for ADA Assays, the Rise of dPCR vs qPCR, International Reference Standards for Vaccine Assays, Anti-AAV TAb Post-Dose Assessment, NanoString Validation, ELISpot as Gold Standard (Part 3 – Recommendations on Gene Therapy, Cell Therapy, Vaccines Immunogenicity & Technologies; Biotherapeutics Immunogenicity & Risk Assessment; ADA/NAb Assay/Reporting Harmonization)

Abstract

The 17^th Workshop on Recent Issues in Bioanalysis (17^th WRIB) took place in Orlando, FL, USA on June 19–23, 2023. Over 1000 professionals representing pharma/biotech companies, CROs, and multiple regulatory agencies convened to actively discuss the most current topics of interest in bioanalysis. The 17^th WRIB included 3 Main Workshops and 7 Specialized Workshops that together spanned 1 week to allow an exhaustive and thorough coverage of all major issues in bioanalysis of biomarkers, immunogenicity, gene therapy, cell therapy and vaccines.

Moreover, in-depth workshops on “EU IVDR 2017/746 Implementation and impact for the Global Biomarker Community: How to Comply with these NEW Regulations” and on “US FDA/OSIS Remote Regulatory Assessments (RRAs)” were the special features of the 17^th edition.

As in previous years, WRIB continued to gather a wide diversity of international, industry opinion leaders and regulatory authority experts working on both small and large molecules as well as gene, cell therapies and vaccines to facilitate sharing and discussions focused on improving quality, increasing regulatory compliance, and achieving scientific excellence on bioanalytical issues.

This 2023 White Paper encompasses recommendations emerging from the extensive discussions held during the workshop and is aimed to provide the bioanalytical community with key information and practical solutions on topics and issues addressed, in an effort to enable advances in scientific excellence, improved quality and better regulatory compliance. Due to its length, the 2023 edition of this comprehensive White Paper has been divided into three parts for editorial reasons.

This publication (Part 3) covers the recommendations on Gene Therapy, Cell therapy, Vaccines and Biotherapeutics Immunogenicity. Part 1A (Mass Spectrometry Assays and Regulated Bioanalysis/BMV), P1B (Regulatory Inputs) and Part 2 (Biomarkers, IVD/CDx, LBA and Cell-Based Assays) are published in volume 16 of Bioanalysis, issues 8 and 9 (2024), respectively.

Abbreviations and Definitions

AAV:

Adeno-associated virus

Ab:

Antibody

ACE:

Affinity capture elution

ADA:

Anti-drug antibody

AI:

Artificial intelligence

Anti-id:

Anti-idiotypic

BAV:

Biomarker assay validation

BCR:

B cell receptor

BD:

Biodistribution

BLA:

Biologics license application

BMV:

Bioanalytical method validation

BsAb:

Bispecific antibody

BTM:

Blood transcription modules

CAR-T:

Chimeric antigen receptor T Cell

CDC:

Complement dependent cytotoxicity

CDx:

Companion diagnostic

cGMP:

Current good manufacturing practices

CHO:

Chinese hamster ovary

CIC:

Circulating Immune Complexes

CK:

Cellular kinetics

CLIA:

Clinical Laboratory Improvement Amendments

Companion diagnostics:

A companion diagnostic device can be an in vitro diagnostic device, testing kit or imaging tool that provides information that is essential for the safe and effective use of a corresponding therapeutic product.

CoP:

Correlate of protection

COU:

Context of Use

CP:

Cut point

CRISPR:

Clustered regularly interspaced short palindromic repeats

CRISPR-Cas9:

It stands for clustered regularly interspaced short palindromic repeats and CRISPR-associated protein 9. By delivering the Cas9 nuclease complexed with a synthetic guide RNA (gRNA) into a cell, the cell’s genome can be cut at a desired location, allowing existing genes to be edited, removed and/or new ones added.

CRO:

Contract Research Organization

CTD:

Common technical document

CV:

Coefficient of variation

DCA:

Domain competition assays

DDA:

Domain detection assays

dPCR:

Digital polymerase chain reaction

ddPCR:

Droplet digital polymerase chain reaction

DNA:

Deoxyribonucleic acid

DoE:

Design of experiments

dsDNA:

Double stranded DNA

DT:

Drug tolerance

ECLA:

Electrochemiluminescence assay

ECD:

Extra cellular domain

ELISA:

Enzyme-linked immunosorbent assay

ELISpot:

Enzyme-linked immune absorbent spot

FFP:

Fit for purpose

FIH:

First in human

FIX:

Factor IX

FPR:

False positive rates

FSC/SSC:

Forward scatter/Side scatter

GCP:

Good Clinical Practices

gDNA:

Genomic DNA

GTx:

Gene therapy

HAMA:

Human anti-mouse antibodies

hFIX:

Human Factor IX

HLA:

Human leukocyte antigen

HMW:

High molecular weight

hFVIII:

Human FVIII

IC:

Immune complex

ICF:

Informed consent form

ICS:

Intracellular cytokine staining

IDE:

Investigational device exemption

IFN-γ:

Type II interferon

IMPD:

Investigational Medicinal Product Dossier

IND:

Investigational new drug

ISI:

Integrated Summary of Immunogenicity

ISR:

Incurred sample reproducibility

IVD:

In vitro device

KOL:

Key opinion leader

LBA:

Ligand binding assay

LCM:

Life cycle management

LCMS:

Liquid chromatography mass spectrometry

LDT:

Laboratory developed test

LIMS:

Laboratory Information Management System

LLOQ:

Lower limit of quantitation

LOD:

Limit of detection

MAA:

Marketing authorization application

MAPPs:

MHC-associated peptide proteomics

mAb:

Monoclonal antibody

MHC:

Major histocompatibility complex

MIQE:

It stands for the minimum information for publication of quantitative real-time PCR experiments. MIQE guidelines describe the minimum information necessary for evaluating qPCR experiments and target the reliability of results to help ensure the integrity of the scientific literature, promote consistency between laboratories, and increase experimental transparency.

ML:

Machine learning

MOA:

Mechanism of action

mRNA:

Messenger RNA

MSR:

Minimum significant ratio

NAb:

Neutralizing antibody

NGS:

Next generation sequencing

NHP:

Non-human primate

NTC:

No template control

pAb:

Polyclonal antibody

PAI:

Pre-approval inspection

PBMC:

Peripheral blood mononuclear cell

PC:

Positive control

PCR:

Polymerase chain reaction

PD:

Pharmacodynamic

PHA:

Phytohemagglutinine

PK:

Pharmacokinetics

PMC:

Postmarketing commitment

PPV:

Positive predictive value

PRNT:

Plaque reduction neutralization test

QC:

Quality control

qPCR:

Quantitative polymerase chain reaction

RCV:

Replication competent virus

RG:

Reference genes

RIN:

RNA integrity number

RNA:

Ribonucleic acid

RNP:

Ribonucleoprotein

ROA:

Route of administration

RT:

Reverse transcription

rVLP:

Recombinant virus-like particles

SAP:

Statistical analysis plan

scFv:

Single-chain variable fragment

SEC:

Size exclusion chromatography

sgRNA:

Single guide RNA

S/N or SNR:

Signal to noise ratio

TAb:

Total antibody

TCR:

T cell receptor

TE:

Target engagement

Tfh:

T follicular helper

TI:

Transduction inhibition

Transduction inhibition:

The inhibition of the transduction of cells by serum or other matrices in a cell-based assay. The inhibition may be caused by antibodies, low molecular weight drugs, or proteins present in the sample.

TLF:

Tables, listing and figures

VCN:

Vector copy number

Vector shedding:

the dissemination of viral vector released outside the treated subject via excreta (e.g., urine and feces), and secreta (e.g., saliva, semen, sweat).

VRNT:

Virus reduction neutralization test

WRIB:

Workshop on Recent Issues in Bioanalysis

Index Part 3

Introduction

SECTION 1 – Gene Therapy, Cell Therapy and Vaccines

Hot Topics & Consolidated Questions Collected from the Global Bioanalytical Community

Discussions, Consensus and Conclusions

Recommendations

SECTION 2 – Immunogenicity of Biotherapeutics

Hot Topics & Consolidated Questions Collected from the Global Bioanalytical Community

Discussions, Consensus and Conclusions

Recommendations

References

Introduction

The 17^th Workshop on Recent Issues in Bioanalysis (17^th WRIB) took place in Orlando, FL, USA on June 19–23, 2023. Over 1000 professionals representing pharma/biotech companies, CROs, and multiple regulatory agencies convened to actively discuss the most current topics of interest in bioanalysis. The 17^th WRIB included 3 Main Workshops and 7 Specialized Workshops that together spanned 1 week to allow an exhaustive and thorough coverage of all major issues in bioanalysis of biomarkers, immunogenicity, gene therapy, cell therapy and vaccines.

Moreover, in-depth workshops on “EU IVDR 2017/746 Implementation and impact for the Global Biomarker Community: How to Comply with these NEW Regulations” and on “US FDA/OSIS Remote Regulatory Assessments (RRAs)” were the special features of the 17^th edition.

As in previous years, WRIB continued to gather a wide diversity of international, industry opinion leaders and regulatory authority experts working on both small and large molecules as well as gene, cell therapies and vaccines to facilitate sharing and discussions focused on improving quality, increasing regulatory compliance, and achieving scientific excellence on bioanalytical issues.

The active contributing chairs included:

Dr. Mitra Azadeh (Ultragenyx), Mr. Mike Baratta (Takeda), Dr. Gopa Biswas (US FDA), Dr. Katherine Block (Genentech), Mr. Mark Dysinger (Alexion), Dr. Seongeun Julia Cho (US FDA), Dr. Isabelle Cludts (UK MHRA), Ms. Kelly Coble (Boehringer Ingelheim), Dr. Vilma Decman (GSK), Dr. Steven Eck (AstraZeneca), Dr. Anna Edmison (Health Canada), Dr. Fabio Garofolo (BRI Frontage),

Dr. Swati Gupta (AbbVie), Dr. Shawna Hengel (Seattle Genetics), Ms. Sarah Hersey (BMS), Dr. Allena Ji (Chiesi), Dr. Wenying Jian (Janssen), Dr. Surinder Kaur (Genentech), Dr. Uma Kavita (Spark Therapeutics), Dr. Christopher Kochansky (Exelixis) Dr. Yi-Dong Lin (Takeda), Dr. Meena (Stoke), Dr. Johanna Mora (BMS), Dr. Rachel Palmer (Sanofi), Dr. Susan Richards (Sanofi) Dr. John Smeraglia (AstraZeneca), Dr. Ivo Sonderegger (Takeda), Dr. Yuan Song (Genentech), Dr. Hiroshi Sugimoto (Takeda), Dr. Matthew Szapacs (AbbVie), Dr. Martin Ullmann (Fresenius Kabi), Dr. Meenu Wadhwa (UK MHRA), Dr. Jian Wang (Crinetics), Dr. Russell Weiner (Takeda), Dr. Long Yuan (Biogen), Dr. Yiyue (Cynthia) Zhang (US FDA).

The participation of major and influential regulatory agencies continued to grow at the 17^th WRIB during its traditional Interactive Regulators’ sessions including presentations and panel discussions on:

• Regulated Bioanalysis/BMV and Biomarkers/CDx/BAV: Dr. Mohsen Rajabi Abhari (US FDA CDER), Mr. Abbas Bandukwala (US FDA CDER), Dr. Kimberly Benson (US FDA), Dr. Gopa Biswas (US FDA CDER), Dr. Seongeun (Julia) Cho (US FDA CDER), Dr. Xiulian Du (US FDA CDER), Dr. Anna Edmison (Health Canada), Ms. Dulcyane Fernandes (Brazil ANVISA), Dr. Brian Folian (US FDA CDER), Dr. Akiko Ishii-Watabe (Japan MHLW), Dr. Dany Ivanova (Health Canada), Dr. Sean Kassim (US FDA CDER), Dr. Olga Kholmanskikh (Belgium FAMHP), Dr. Elham Kossary (WHO), Ms. Sonja Kwadijk-de Gijsel, (Netherlands IGJ/EU EMA), Dr. Yang Lu (US FDA CDER), Ms. Tamara Pinkney (US FDA CDRH), Dr. Kara Scheibner (US FDA CDER), Dr. Yoichi Tanaka (Japan MHLW), Dr. Mary Thanh Hai (US FDA CDER), Dr. Yow-Ming Wang (US FDA CDER), Ms. Emma Whale, (UK MHRA), Dr. Joshua Xu (US FDA NCTR), Dr. Li Yang (US FDA), Dr. Yiyue (Cynthia) Zhang (US FDA CDER)

• Biotherapeutic Immunogenicity, Gene Therapy, Cell Therapy and Vaccines: Dr. Mohsen Rajabi Abhari, (US FDA CDER), Dr. Nirjal Bhattarai (US FDA CBER), Dr. Alessandra Buoninfante (EU EMA), Dr. Isabelle Cludts (UK MHRA), Dr. Heba Degheidy (US FDA CBER), Dr. Sandra Diebold (UK MHRA), Dr. Akiko Ishii-Watabe (Japan MHLW), Dr. Mohanraj Manangeeswaran (US FDA CDER), Dr. Christian Mayer (Austria AGES/EU EMA), Dr. Kimberly Maxfield (US FDA CDER), Dr. João Pedras-Vasconcelos (US FDA CDER), Dr. Kara Scheibner (US FDA CDER), Dr. Sophie Shubow (US FDA CDER), Dr. Yoichi Tanaka (Japan MHLW), Dr. Seth Thacker (US FDA CDER), Dr. Omar Tounekti (Health Canada), Dr. Daniela Verthelyi (US FDA CDER), Dr. Meenu Wadhwa (UK MHRA), Ms. Leslie Wagner (US FDA CBER), Dr. Joshua Xu (US FDA NCTR).

The 17^th WRIB included the traditional evening roundtables, which were attended by both industry key opinion leaders (KOL) and regulatory representatives. The extensive and fruitful discussions from these roundtables together with the lectures and open panel discussions amongst the presenters, regulators and attendees culminated in consensus and recommendations on items presented in this White Paper.

A total of 55 recent issues (’hot’ topics) were addressed and distilled into a series of relevant recommendations. Presented in the current White Paper is the background on each issue, exchanges, discussions, consensus and resulting recommendations.

Due to its length, the 2023 edition of this comprehensive White Paper has been divided into three parts for editorial reasons. This publication covers Part 3 recommendations.

Part 1A – Volume 16 Issue 9 Month May 2024

Mass Spectrometry, Chromatography & Sample Preparation (6 Topics)

• 1.Deuterated Drugs: Implication for Bioanalysis
• 2.Quantitation of Nanoparticles and Novel Lipids for mRNA Vaccines and Therapeutics
• 3.Clinical Tissue, Tumor and FFPE Biopsy Quantification
• 4.Hybrid Assays (IA-MS) - New Applications and Enhanced Complementarity with Conventional Technologies
• 5.Mass Spec Targeted Proteomics for Large Protein Panels: Regulatory & Technology
• 6.Advanced Mass Spec and Multiplexed Approaches for Biotherapeutics

Mass Spectrometry Assays, Latest Developments, Challenges, & Solutions (6 Topics)

• 1.High Efficiency and Quality Non-Regulated Bioanalysis
• 2.Quantitative Bioanalytical Methods for Small Molecule Covalent Inhibitors
• 3.Quantitative Mass Spectrometry for Chiral Bioanalysis
• 4.Biomarkers Novel Approaches for Method Development/Validation
• 5.Oligonucleotides Novel Approaches for Method Development/Validation
• 6.ADC Novel Approaches for Method Development/Validation

BMV & Regulated Bioanalysis Latest Developments, Challenges, & Solutions (6 Topics)

• 1.US FDA/OSIS Remote Regulatory Assessments (RRAs)
• 2.Sample Collection, Reconciliation, Chain of Custody
• 3.Patient Centric Sampling for Clinical Analysis
• 4.Special Issues in the Implementation of ICH M10
• 5.Look into the Future of BMV ICH M10
• 6.PK/PD Data Exchangeability in and out of China

Part 1B – Volume 16 Issue 9 May 2024

Input from Regulatory Agencies on Regulated Bioanalysis/BMV & Biomarkers/IVD/CDx/BAV

US FDA

• Office of Study Integrity and Surveillance (OSIS): A Day in the Life of Our Remote Regulatory Assessments

• Data Integrity Issues from Recent FDA BIMO Inspections and Remote Regulatory Assessments

• Examination on Overall Acceptable Incurred Sample Reanalysis (ISR)

• Common Deficiencies Related to Method Validation in ANDA Submissions

• Use of Biomarkers in Drug Development: A Regulatory Perspective

• CDRH CLIA Categorization Processes

• Opportunities for Collaborative Innovations Informed By Review Experience in Office of Clinical Pharmacology

• 21st Century Cures: Biomarker Qualification and Analytical Guidance

• PD Biomarker Bioanalysis - Guidance Recommendations and Review Experience

Health Canada

• ICH M10 Implementation at Health Canada: Comparative Bioavailability Studies for Chemical Drugs

• Case Studies from Health Canada Review: Comparative Bioavailability Studies for Chemical Drugs

UK MHRA

• MHRA Inspection Update

• Implementation of Data Integrity Guidance in the UK

Belgium FAMHP / EU EMA

• Integrated CDx and pharmaceutical development in EU regulatory landscape

Dutch Youth & Health Inspectorate (IGJ)

• Introduction on the Health and Youth Care Inspectorate

Brazil ANVISA

• ANVISA updates on ICH M10 Implementation

WHO

• Sharing some of the Significant Deficiencies - CRO Inspections

Input from Regulatory Agencies on Immunogenicity, Gene Therapy, Cell Therapy & Vaccines

Immunogenicity

US FDA

• Tools to Assess Immunogenicity Risk and New Computational Methods

• Updates of OCP Efforts on Evaluating Clinical Impact of Immunogenicity

• Neutralizing Anti-drug Antibody Assays for Biologics- a review of format choices in recent 351 (a) BLA approvals

• Biosimilar User Fee Act (BsUFA) III Regulatory Science

• An Update on Immunogenicity Review Committee Activities

• Recent Advances on Cellular and Gene Therapies

• Is your assay fit for purpose?

• Recent updates on Flow Cytometry and Cell Therapy

• Benchmarking and Improving Indel Calling from Oncopanel Sequencing Data

EU EMA

• Vaccine Assay Validation: Regulatory Perspective

Health Canada

• Assessment of Qualified Laboratories for Cell and Gene Therapies: Health Canada’s approach

UK MHRA

• NIBSC Reference Reagents for Flow Cytometry Standardization

Austria AGES

• Fast-Track Development and Approval of Covid-19 Vaccines

Japan MHLW

• Points to Consider for Using ADA Screening Assay signal-to-noise ratio (S/N) as an Alternative to Titer in Immunogenicity Assessment

• Bioanalysis of AAV vector by qPCR and digital PCR

Part 2 – Volume 16 Issue 8 Month April 2024

Biomarkers, IVD, CDx Assays Development & Validation (7 Topics)

• 1.EU IVDR 2022 Implementation and impact for the Global Biomarker Community
• 2.IVD/CDx, CLIA Approved Assay and Regulatory Requirements
• 3.What is our Biomarker Assay Actually Measuring?
• 4.Free Target Assays for Drug Candidates Targeting Soluble Targets
• 5.Biomarkers Development/Validation for Vaccine Study Endpoints
• 6.Point of Care (PoC) Assays Development and FFP BAV in Clinical Trials
• 7.Fit for Purpose Biomarker Assay Validation (FFP BAV) Challenges for LBA & Mass Spec

Cell-Based Assays (4 Topics)

• 1.CBA Novel Strategies for Method Development/Validation
• 2.Quality Data Generation in High Dimensional Cytometry
• 3.Biomarkers Novel Approaches for Method Development/Validation
• 4.Cell Therapy and Vaccine Novel Approaches for Method Development/ Validation

Ligand-Binding/Enzyme Assays & Critical Reagents (6 Topics)

• 1.Emerging & Multiplexing Technologies in Bioanalysis
• 2.LBA Tissue Analysis
• 3.Advanced Labeled Critical Reagents Strategies and Hybridoma Phage Display
• 4.Advancements in Enzyme Assays
• 5.Novel Modalities Method Development/ Validation Challenges
• 6.Single Well Analysis (Singlicate) for ADA Assays

Part 3 – Volume 16 Issue 7 Month April 2024

Gene Therapy, Cell Therapy & Vaccines (15 Topics)

Technologies

• 1.The Rise of dPCR
• 2.Is ELISpot still the Gold Standard for Assessing Cellular Immunogenicity?
• 3.NanoString Assay Development/Validation
• 4.NGS Assay Development/Validation
• 5.qPCR Assays Development/Validation

Immunogenicity

• 6.Issues with International Reference Standards for Vaccine Clinical Assay
• 7.Anti-AAV TAb Post-Dose Assessment: Develop an Efficient Analysis Strategy
• 8.Further Considerations on LNP Immunogenicity
• 9.Vaccine Clinical Assays Life Cycle Management
• 10.AAV TAb/NAb Assessment
• 11.Transgene Immunogenicity Assessment
• 12.Vaccine Immunogenicity Assessment
• 13.PK & Biodistribution for Replication Competent Viral Vectors
• 14.PK & Biodistribution for Virus-vector Vaccine
• 15.PK & Biodistribution for CAR-T Cells

Immunogenicity of Biotherapeutics (9 Topics)

• 1.ISR for ADA Assays
• 2.Strategies to Distinguish between Clinical Pre-Existing Antibodies (PEA) and Treatment Emergent Antibodies (TEA)
• 3.Universal & Generic Assays Formats for Immunogenicity Assessment
• 4.Immunogenicity Assay/Reporting Harmonization
• 5.Immunogenicity Risk-based Approaches, Prediction and Mitigation: Focus on Novel Monitoring Strategies and in vitro Immunogenicity Assessment
• 6.Lessons Learned from Immunogenicity Studies/Filings: Sharing Advanced Experience Related to Immunogenicity Strategies/Approaches
• 7.Immunogenicity of Complex Biotherapeutics
• 8.Cut Point Assessment
• 9.Assay Comparability

SECTION 1 – Gene Therapy, Cell Therapy & Vaccines

Johanna Mora¹, Rachel Palmer², Leslie Wagner²¹, Bonnie Wu²³, Michael Partridge¹⁶, Meena³, Ivo Sonderegger⁴, John Smeraglia⁵, Nicoletta Bivi⁶, Naveen Dakappagari⁷, Sandra Diebold⁸, Fabio Garofolo⁹, Christine Grimaldi¹⁶, Warren Kalina¹⁰, John Kamerud¹⁰, Sumit Kar¹¹, Jean-Claude Marshall¹², Christian Mayer¹³, Andrew Melton¹⁴, Keith Merdek¹, Katrina Nolan¹⁵, Serge Picard¹, Weiping Shao⁵, Jessica Seitzer¹⁷, Yoichi Tanaka¹⁸, Omar Tounekti¹⁹, Adam Vigil²⁰, Karl Walravens²², Joshua Xu^24,25, Weifeng Xu¹⁵, Yuanxin Xu¹⁷, Lin Yang^24,25 & Liang Zhu¹²

Authors are presented in alphabetical order of their last name, with the exception of the first 8 authors who were session chairs, working dinner facilitators and/or major contributors.

The affiliations can be found at the beginning of the article.

HOT TOPICS & CONSOLIDATED QUESTIONS COLLECTED FROM THE GLOBAL BIOANALYTICAL COMMUNITY

The topics detailed below were considered as the most relevant “hot topics” based on feedback collected from the 16^th WRIB attendees. They were reviewed and consolidated by globally recognized opinion leaders before being submitted for discussion during the 17^th WRIB. The background on each issue, discussions, consensus, and conclusions are in the next section and a summary of the key recommendations is provided in the final section of this manuscript.

The Rise of dPCR

When is dPCR preferred over qPCR? What are recommendations for how to establish droplet digital PCR assay lower limit of quantitation?

Is ELISpot still the Gold Standard for Assessing Cellular Immunogenicity?

For cell therapy, would you consider best practice to analyze cellular immunogenicity only in cases where there is an unexplained change in cellular kinetics and/or an adverse event related to immunogenicity? After comparison in a single study, would you advise implementation in future studies? Should a cut point be defined for ELISPOT for GTx immunogenicity risk assessment?

NanoString Assay Development/Validation

What specific parameters beyond assay precision should be included for NanoString/genomics assay validation? How does one deal with setting acceptance criteria with hundreds or thousands of analytes/genes. Does one look at all analytes/genes, look at select analytes covering the range of expression (how many)?

NGS Assay Development/Validation

Similar topics as NanoString assay development. What is the preference for types of wet lab controls used for assay and pipeline control of NGS data when the genomic changes are diverse (e.g., indels)? How far must one go to sequence technically challenging regions before deeming that they are not technically feasible to sequence and are of low biological risk?

qPCR Assays Development/Validation

For PK purposes, and measuring mRNA as part of drug product, what should be used as the reference standard for qPCR assay? Is there any way to standardize dPCR instruments and minimize technical variability in dPCR measurements for quantification of target nucleic acids (e.g., shedding)? Is copies/cell acceptable to be used for the quantification of nucleic acids in biological samples instead of copies/ng DNA or copies/μL? – use reference gene to quantify # cells? What would be the preferred platform for quantitation of mRNA-LNP for PK purposes (bDNA, qPCR, dPCR or other)? Is it important to utilize the same methodology for TK/PK in clinical studies as in Tox studies?

Issues with International Reference Standards for Vaccine Clinical Assay

When international standards have been calibrated in one assay technology, how does this translate into IUs in other assay technology before a specific calibration in this assay is available?

Anti-AAV TAb Post-Dose Assessment: Development of an Efficient Analysis Strategy

Do we need to test post-dose AAV TAb? After administration of an AAV-based gene therapy, virtually all patients will be strongly positive for anti-AAV antibodies, what testing should one do for post-dose anti-AAV responses? Can the testing strategy be optimized for efficiency?

Further Considerations on LNP Immunogenicity

With an increased prevalence of pre-existing anti-PEG antibodies should these assays be developed and validated more like vaccine assays? What is the need and clinical relevance of assessing anti-PEG antibodies and titers in mRNA vaccines?

Vaccine Clinical Assays Life Cycle Management

Many legacy assays are being upgraded to new platforms. After development and validation, what is the extent of bridging and cross validation?

AAV TAb/NAb Assessment

What is the most reliable approach to prepare NAb negative pooled matrix for AAV NAb assays? Should we use arbitrary 50%INH or assay specific cut point for AAV NAb screening and titering assays? What is the titering strategy that could effectively reduce NAb negative pooled sample consumption and meanwhile generate reliable titer values? When should AAV NAb assessment be implemented for nonclinical studies involving small and large animals? Should ADA to transgene protein be assessed for all GLP studies? For clinical studies, is a CDx needed if there is a lack of impact of pre-existing anti-AAV on transgene product expression?

Transgene Immunogenicity Assessment

When do we need to do immunogenicity assessment and how much do we need to characterize the immunogenicity for transgenes and transgene products? What is the utility of applying in silico/in vitro assessment in the immunogenicity prediction for cell and gene therapy? For immunogenicity testing to detect anti-transgene product antibodies, do we need to follow the same conservative analysis paradigm that we use for protein therapeutics (screen, confirm, titer, NAb, 5% false positive, 1% false positive)?

Vaccine Immunogenicity Assessment

For bridging ADA assay and sample analysis, what is the industry’s thinking to harmonize ADA assay acceptance criteria and switching to singlicate assay? What is the relative value of binding assays compared to NAb assays?

PK & Biodistribution for Replication Competent Viral Vectors

What methods should be employed for replicant competent viral vectors since standard bioanalytical techniques such as qPCR are unable to distinguish the input drug from replicated virus?

PK & Biodistribution for Virus-vector Vaccines

For biodistribution studies, how often do we need to do infectious virus infectivity assay in addition to the PCR assay?

PK & Biodistribution for CAR-T Cells

What is your biggest challenge with CAR-T monitoring in clinical trials? What criteria would you use for selection of a CAR-T monitoring assay type (e.g., flow cytometry versus qPCR)?

DISCUSSIONS, CONSENSUS & CONCLUSIONS

The Rise of dPCR

Digital PCR for DNA or RNA (dPCR) has emerged as a prevalent assay platform for bioanalysis of gene and cell therapy products, with some key advantages over conventional qPCR. In particular, dPCR is widely used for tissue biodistribution (such as AAV) and vector shedding studies for gene therapy [27] and linkage of different vector sites and vector integrity. The 2020 and 2021 White Papers in Bioanalysis discussed in depth the differences between qPCR and dPCR [27,30]. The 2021 White Paper detailed validation best practices, including acceptance criteria for plate batches, LLOQ, LOD, and precision/accuracy requirements. It was recommended to test precision, accuracy, sensitivity, selectivity, LLOQ, and short-term stability for assay validation with QC samples of spiked plasmids.

The 2021 recommendations were followed by the 2022 GCC White Paper with additional consensus on dPCR development and validation from the CRO perspective [35]. There was agreement that experiments performed during dPCR validation are overall similar to qPCR. There are, however, some unique aspects in the experimental and technical design of dPCR that are fundamentally different from qPCR (e.g., no standard curve nor a different housekeeping gene). Finally, the 2022 White Paper in Bioanalysis provided additional case studies, practical considerations, and recommendations based on questions from the aforementioned validation recommendations [33]. This included a recommendation for reporting units in copies per μg of total genomic DNA for tissue biodistribution. Other guidance was provided for monitoring instrument reproducibility with run controls.

The 2023 discussions provided additional case studies comparing dPCR to qPCR for gene therapy applications and recommendations for dPCR validation. Both dPCR and qPCR can measure target gene for assessment of tissue transduction (tropism), biodistribution, mRNA expression or modulation, and vector shedding. A case study was discussed showing that dPCR performed better than qPCR in a method for measuring exon skipping of dystrophin mRNA with an antisense oligonucleotide (ASO). Exon skipping assays are especially challenging for Duchene Muscular Dystrophy (DMD), the gene that encodes the dystrophin protein, because of the low abundance of DMD mRNA [36], around 5 – 10 copies/nucleus. Previous methods for measuring DMD exon skipping used nested PCR or qRT-PCR with a pre-amplification step, but these methods favored the smaller skipped PCR product and overestimated the skipping rate by about 2-fold compared to dPCR. Digital PCR accurately calculated exon skip levels and showed increased sensitivity and improved precision [37].

Other case studies provided data on transgene and AAV quantification to better understand and interpret data. One study found an unexpected pattern of changes where protein expression increased while AAV vector genome copies of concatemer formation, as determined by dPCR, decreased over time following AAV administration [38]. It was found that tissue samples digested with EcoR1, a restriction enzyme within the inverted terminal repeat sequences (ITRs) of the AAV genome, resulted in much higher copy numbers per cell compared to undigested samples or those digested with Nru1, a restriction enzyme which is not present in the AAV genome. These results suggested the formation of concatemers after single stranded AAV DNA enters the nucleus and demonstrated dPCR could not accurately measure AAV vector genome copies of concatemers without proper digestion with a restriction enzyme within the ITRs.

Another case study was shown comparing different dPCR platforms. All dPCR platforms operate based on similar principles, however, there can be variations in their performance and capabilities affected by instrument design, consistency in droplet number, better separation from negative droplets, and lack of the “rain” profile of the droplets, and/or partitioning method. Variability between dPCR platforms can be a big concern since dPCR does not require a calibration curve for quantification.

Standardization efforts are underway to improve the comparability and reproducibility of dPCR results across different platforms and laboratories, including the development of reference materials, guidelines, and quality control measures. It was recommended to use a common reference sample or standard and to perform a comparative analysis between different platforms using the same samples and conditions. It is also important to carefully consider the strengths and weaknesses of different dPCR platforms before selecting one for a particular application. Some platforms may be better suited for specific types of samples or targets, depending on factors such as concentration, size, and complexity. Digital PCR remains a valuable tool for achieving high precision and sensitivity in gene therapy research. However, the case studies also highlighted that careful evaluation, optimization, and validation of dPCR assays are essential for reliable results, especially when assessing AAV gene therapy biodistribution.

The case studies and accompanying data led to consensus recommendations updating previous recommendations on dPCR. There was agreement that absolute quantification by dPCR is preferred over qPCR in cases where no standard curve is possible, and in instances when additional sensitivity is needed.

The discussion of LLOQ determination for dPCR was also revisited. There was consensus that previous recommendations of 50 copies/μg DNA from the GCC qPCR dPCR White Paper, 2021 and 2022 White Papers in Bioanalysis still hold. It was added that it is advisable to evaluate extraction efficiency by spiking plasmid (as a surrogate control) into whole blood, or from tissue extract for biodistribution. LLOQ determination should consider the instrument’s theoretical %CV and background gDNA amount to reach the recommended 50 copies/μg DNA sensitivity.

Is ELISpot still the Gold Standard for Assessing Cellular Immunogenicity?

Cellular immunogenicity is not a common assessment for conventional biotherapeutics, but it may be important for cell therapies, as they have potential to trigger both innate and adaptive immune responses, which could impact the persistence, safety, and efficacy of cell therapies. The bioanalytical support of cellular assessments requires careful method development, qualification, and sample analysis, as well as harmonization of practices and evaluation of risks. The 2022 White Paper in Bioanalysis [27,30,33] recommended performing a risk assessment based on the product design and generating data for immunogenicity evaluation. A case-by-case discussion with regulatory agencies to get agreement on the bioanalytical methods was recommended [39]. Some of the assays available for cellular immune response assessments include ELISpot, multi-parameter (multiple cytokine) assessment by ELISpot/fluorospot, intracellular cytokine flow cytometry, whole blood stimulation ELISA, T cell receptor repertoire analysis, proliferation assays, degranulation assays, and chromium release assays [40,41].

ELISpot is currently one of the most widely implemented assays to measure either wanted T cell response (such as vaccine effect or CAR-T) or unwanted response in gene and cell therapies. Previous discussions on ELISPOT and its development, validation, and harmonization across laboratories have been discussed extensively in 2020, 2021, and 2022 White Papers in Bioanalysis [27,30,33]. These recommendations included ensuring that PBMCs are isolated within a pre-defined time from whole blood collection by well trained personnel to ensure cell viability and functionality, automated reading for spot counts, and that laboratories performing ELISpot assays participate in external QA assessments, when possible. This has led to the development of proficiency panels for assay qualification and calibration/alignment for inter-laboratory assay performance. Utilization of common positive controls including pan positive controls (such as CEF (HLA Class I Control, a mixture of 32 peptides derived from the human Cytomegalovirus (CMV)), Epstein-Barr Virus (EBV), Influenza Virus (Flu), phytohemagglutinin (PHA), concanavalin A, phorbol 12-myristate 13-acetate in combination with ionomycin (PMA/Io), staphylococcal enterotoxin B (SEB) or anti-CD3/CD28 antibodies), and negative controls. Finally, the agreed parameters for validation include analytical sensitivity (AS)/lower limit of detection (LLOD), upper limit of quantitation (ULOQ), lower limit of quantitation (LLOQ), assay specificity, intra and inter-assay precision, dilutional linearity, and QC specifications [42].

The GCC also supported prior recommendations for ELISpot assay validation presented in Maecker et al., Janetzki et al., Piccoli et al. and Corsaro et al. [27,43–46]. Additional recommendations were provided for SOPs (e.g., PBMC counting method, plate reading requirements), critical reagents, QCs, and a minimum of 10 donors for precision testing.

The availability of these shared data and recommendations has further grown ELISpot’s prevalence. However, disadvantages of the platform include PBMC isolation requiring high volumes of blood from patients, sample transport logistics, PBMC viability and functionality after thaw depending on sample processing and cryopreservation steps, and limited multiplexing capability to differentiate response from T cell subpopulations. A new case study was reviewed demonstrating these challenges with wide variability of stimulations over time. This led to renewed efforts to develop new best practices for the assay at a new laboratory with a validation plan to detail the purity and concentration of peptide stimulants. Camera settings, applied during sample well imaging, were also modified for more consistency. However, challenges remained with sample quality and viability, with dead cells affecting response.

This led to a new discussion on whether ELISpot is still the gold standard or whether other methods to assess T cell immunity should be considered. In addition to alternate methods highlighted above, a case study was shared of TruCulture^® as an orthogonal approach. The TruCulture^® system is a syringe-based device designed for point of care use [47]. The tube contains 1 mL of cell culture medium which may include a variety of immunological stimulants. One ml of whole blood is drawn in and mixed with the medium. Control tubes with no stimulants are used to assess background levels. The tubes are incubated at 37°C in a dry heat-block for 24–48 hrs. Cells are separated by a valve separator component. The supernatant (free of cells) is stored at -80 C until analysis. The evaluation of TruCulture^® as an alternative involved selecting peptides, controls, and optimizing the incubation time. The case study demonstrated comparable data, for a positive control, between TruCulture^® and ELISpot, while requiring a lower sample volume, and reducing the sample preparation issues.

These case studies led to a discussion on overall recommendations for cell therapy immunogenicity, including ELISpot and other platforms. Further for cell therapy, the discussion focused on whether or not it is best practice to analyze cellular immunogenicity only in cases where there is an unexplained change in cellular kinetics and/or an adverse event related to immunogenicity. There was an agreement on the approach to collect samples but only run assays if impact is observed (safety, PK, or other cellular kinetic change that cannot be explained by humoral ADAs). This strategy should be written in the protocol. For allogenic therapy, more data is needed. Risk does change, as minor HLAs in the product can be immunogenic. For the use of alternate or orthogonal methods for immunogenicity assessment, it is reasonable to compare new methods in one study to ELISpot, then implement. An alternative approach is to use intracellular cytokine flow cytometry in combination with cell surface markers to fully evaluate T cell responses.

Another discussion topic for ELISpot was if a cut point should be defined for GTx immunogenicity risk assessment. There was wide agreement that cut points are not advised as background variability is too high. This topic has also been discussed previously in ELISpot GCC White Paper [27,45].

NanoString Assay Development/Validation

The Nanostring nCounter analysis system is a new technology that can measure the expression of hundreds of genes in a multiplex fashion without the need for amplification or other enzymatic steps that can potentially introduce bias. The system has a simple and efficient workflow that requires minimal manual effort and produces highly reproducible data. This platform was first introduced for discussion in 2022 with initial recommendations on utility and validation [33]. It was recommended, regardless of platform, that the use of gene expression data as a surrogate endpoint needs to be evaluated case by case in discussion with Regulatory Agencies in the context of data showing correlation with clinical outcomes.

The 2023 discussion provided additional recommendations for NanoString assay development and validation with examples of implementation. The main discussion topic was which specific parameters beyond assay precision should be included for assay validation. There is no regulatory guidance yet beyond diagnostic assay white papers.

For gene expression assays such as those on the NanoString platform, it was determined that although RNA integrity is one parameter that could be impacted by instability over time, other attributes could also be affected and impact gene expression analysis. Thus, despite the availability of good technologies for determining QC metrics for mRNA, choosing just one such as RNA integrity (RIN) assessment is not sufficient for determining whether samples are of sufficient quality and formal stability in the bioanalytical assay itself should still be assessed.

Recommendations were also provided for setting acceptance criteria with hundreds or thousands of analytes/genes. For expression panels, there was an agreement that criteria should be FFP depending on the stage of the assay. For panels that are more exploratory, it was recommended to focus on assessing performance of a subset of transcripts stratified by complexity and expression level. For smaller panels with potential CDx applications, it is best practice to evaluate performance of all genes in the panel. Specific acceptance criteria for validating a NanoString gene expression assay were also suggested for the first time as shown in Table 1. Prior to performing this validation, during development the following key factors should be assessed and fixed: total amount of input RNA, hybridization time, and statistical approaches for generating final gene expression results (raw counts vs positive control normalized counts) and for determining assay cutoffs for calling results >LLOQ (i.e.: mean plus some multiple of the standard deviation for the internal set of negative controls).

Table 1. Acceptance Criteria of Validation Parameters for NanoString.

Parameter	Acceptance Criteria
Validation of Key Assay Parameters (RNA input, hybridization time, data analysis approaches for generating expression results)
Precision	• 4 samples (target populations representing study disease matrices) analyzed in triplicate in each of 3 runs (n = 9/sample) • For at least 75% of genes in each sample, 7/9 replicates must agree as < or >LLOQ. • For at least 75% of genes in each sample with at least 2/3 replicates (intra-run) or 7/9 replicates (inter-run) >LLOQ, %CV (positive control normalized counts) across replicates should be <20%
Linearity/ Acceptable Sample Input & Maximum allowable counts to ensure lack of signal saturation, chip overcrowding, and/or codeset exhaustion	• One sample analyzed in triplicate in each of three runs (n = 9) across range of starting RNA input (i.e. 15,30,45, and 60 ng) • %CV Acceptance criteria for Precision should be met at each input concentration • Mean positive control normalized counts for all replicates at each input will be adjusted/normalized relative to one input such as the SOP defined input ((i.e., counts from 15 ng multiplied by 2, counts from 60 ng divided by 2 if using example inputs above with SOP defined value of 30 ng)   • Across all inputs, %CV for adjusted mean counts should be <20% for at least 75% of genes >LLOQ.   • At each input, for at least 75% of genes >LLOQ the % relative error (%RE) of adjusted mean counts versus the mean adjusted count across all inputs should be <20% • Maximum total raw counts for a sample = sum of counts from all genes at the highest RNA input that allows above linearity criteria to be achieved (median of all replicates; n = 9) • Maximum raw counts per gene = mean of counts from all replicates (n = 9) of the gene with the highest raw counts at the highest RNA input meeting above linearity acceptance
Suggested Stability   • Sample Freeze and Thaw   • Short term sample (RNA) on ice   • Codesets at ambient temperature   • Loaded cartridges at 4°C   • Hybridization reactions at 4°C   • Loaded hybridization reactions on prep station Long term RNA sample at <-60°C	• 2 samples analyzed in triplicate at each condition • Each sample at each condition must pass precision acceptance criteria • For at least 75% of genes in each sample, positive control normalized counts at test conditions must be within 30% of baseline (precision results used when possible)
Robustness   • Hybridization time   • RNA sample integrity (RIN)	• 2 samples (preferably representing study disease matrices) analyzed in triplicate at each hybridization time • Multiple sonications of one sample with varying integrity used for RIN assessment • Each sample at each condition must pass precision acceptance criteria • For at least 75% of genes in each sample, positive control normalized counts at test conditions must be within 30% of those at 42 hr hybridization time (from precision) or those from the same sample with the highest RIN

As a proof of concept, these validation criteria were tested in a case study. For precision, 95.1% to 97.6% of genes >LLOQ had %CV<20%. 81.0% to 87.0% of genes were in agreement with respect to being < or >LLOQ in at least 7 of 9 replicates. The % recovery from the mean of mean dilution adjusted positive control normalized counts across all input RNA concentrations was tested with 96.8% to 100.0% of genes for which a minimum of 2 input RNA were >LLOQ had %RE within 80–120% versus the mean value across all inputs. The assay also met stability and robustness criteria set as proposed.

Overall, the discussions reinforced the prior recommendations with additional data that gene expression data can potentially support a surrogate endpoint showing correlation with clinical outcomes. Criteria should be FFP depending on the stage of the gene expression NanoString assay.

NGS Assay Development/Validation

Evaluating efficacy and safety is a crucial step for developing gene and cell therapies, such as products that edit genomes in vivo and ex vivo. These therapies can cause genomic changes, either desired (efficacy) or undesired (safety) that can be characterized by methods based on next-generation sequencing (NGS) technologies. While many other technologies exist, NGS is growing in use. Bioanalytical assessment and regulatory considerations were discussed in the 2022 White Paper in Bioanalysis [33]. There was agreement that NGS is an important platform for detecting on and off-target effects, but more data and experience is needed. The key considerations for NGS analytical development and validation include establishing a quality framework (reference samples, QC samples, proficiency testing, and reproducible data analysis pipelines) and consistency using automation. Validation of the software for data analysis was also agreed to be critical.

FFP validation of NGS was discussed in 2023 supporting prior recommendations but with additional supporting data. The case studies showed that important QC measures for NGS include minimum and ideal DNA/RNA yield, fragment and library size, and comparison to orthogonal methods. Key assay parameters for validation are accuracy and precision (inter- and intra-run). Operational considerations were also described for functional redundancy including backup systems when systems go offline, manual methods, and other risk mitigation strategies.

Additional discussion also addressed the use of appropriate analysis methods and pipelines for consistent analysis and data reproducibility. The use of wet laboratory controls and qualification data sets, such as GIAB reference standards, synthetic DNA oligos, and plasmids, to create data sets for NGS pipeline control was evaluated [48]. Published data sets and known positive controls are also available to check the reproducibility over time and establish QC parameters for pipeline qualification and assay sensitivity. Moreover, it was recommended to use best practices for software engineering and cloud computing for pipeline qualification. Software engineering tools help manage versions, documentations, unit tests, and systematic validations for data management and computational pipelines. Furthermore, databases should have the ability to be queried to track information, documentations, samples, and avoid clerical errors.

Based on previous case studies, recommendations were provided for preference for types of wet laboratory controls used for assay and pipeline control of NGS data when the genomic changes are diverse (e.g., indels). There was an agreement that using known positive samples with dilution is best approach for mimicking samples. Synthetic DNA such as gBlocks mixtures are more challenging to use as spike-ins; one can use them to qualify positive controls. It was also discussed that how far one must go to sequence technically challenging regions before deeming that they are not technically feasible to sequence and are of low biological risk. It was recommended that this may be informed by using various prediction methods from in silico, in vitro assays (e.g., siteseq), and animal studies. If it is low frequency, sponsors should proceed to provide justification that site has low biological risk. If it is high frequency, sponsors should try multiple ways to sequence, but may still be able to justify the site has low biological risk.

qPCR Assays Development/Validation

In 2020 and 2021, the White Papers in Bioanalysis and GCC qPCR and dPCR White Paper gave detailed advice and guidelines for validating qPCR and dPCR methods for gene and cell therapy, and vaccines [27,30,35]. More case studies were presented in 2022, to support these globally accepted recommendations [33]. Specific questions were addressed for reporting of data (e.g., definition and units of LOD for biodistribution and vaccine endpoints). Other topics included the need to characterize and explain/justify background signals in qPCR for AAV DNA template and gRNA, using orthogonal assessments of contamination. Additional recommendations were made to monitor instrument reproducibility for dPCR using run acceptance by controls with known high and low copies in addition to consultation with vendors.

In 2023, additional questions were addressed for qPCR and dPCR for gene and cell therapy. mRNA-LNP is an exciting platform for delivery of therapeutic proteins. Unlike some other delivery components used in gene therapies, there is a need for PK evaluation, with mRNA as the analyte. bDNA, qPCR and dPCR are some of the potential assay methodologies for PK assessments of mRNA. There was a discussion that the translatability between methods and the availability at CROs are among the considerations in choice of assay platforms. The clear recommendation was that it is the sponsor’s choice on what platform is FFP with no regulatory preference. Bioanalytical platform switching is not recommended between pre-clinical and clinical studies. A change of platforms can be justified with scientific rationale and data, asking regulators in advance as part of pre-IND briefing book and obtaining an alignment in writing.

In addition to PK applications of qPCR, successful PCR designs for vaccine diagnostics were also discussed. Diagnostic PCR-based tests are widely used for clinical trials that need accurate and precise detection of the target analyte or pathogen in vaccine trials. The intended use of the diagnostic test affects the balance between clinical sensitivity and specificity. High specificity is needed to avoid false positives for efficacy-based studies, while high sensitivity is preferred for clinical diagnostics and epidemiological studies. The intended use and the gene targets of interest should be considered carefully in early development. Analytical specificity can be achieved with gene targets that do not change over time and are inclusive of the target pathogen and exclusive of near neighbor pathogens. Once targets are chosen, validating highly specific PCR-based tests requires establishing cut-offs that depend on either benchmarked assays or large sample sets of healthy subjects with similar demographics as the clinical study population. Additional case studies were shared for assessing matrix effects and performing lifecycle management. Strategies shown included testing Ct variability in different matrices that may be encountered using the most common matrix as the benchmark. Lifecycle management strategies included using proficiency panels to flag reagent issues and monitoring trends in assay QCs and proficiency panel members throughout the assay’s life.

In addition to new supporting data in case studies, discussions led to updated recommendations on specific aspects of qPCR/dPCR assay validation. For measuring mRNA as part of drug product PK analysis (e.g., LNP encapsulated mRNA), a question raised was what should be used as the reference standard for a qPCR assay. There was an agreement that the standard should be as close as possible to the drug product (e.g., drug product diluted in buffer). Sponsors can show data on stability of drug product, acknowledging that the stability of mRNA in vivo could be poor. One important caveat is that because the analyte is RNA, it needs to be reverse transcribed, and the assumption is RT efficiency is 100%. There was a recognition of the challenges in reproducibility of this assay because kit lots, storage conditions, analysts, and equipment may differ considerably from lab to lab.

Another topic was whether there is any way to standardize dPCR instruments and minimize technical variability in dPCR measurements for quantification of target nucleic acids (e.g., shedding). This topic was previously discussed in the 2022 WRIB White Paper and GCC dPCR White Paper [27,30,33]. The expert panel recognized that there is no reference standard for a dPCR assay, as it is based on partition and statistical calculation. Output could be different from different instruments. Further, extraction efficiency could impact the assay. The recommendation was to perform blinded proficiency test with CROs with known positives and negatives, and characterize the differences. Plasmid can be used for this test. It was suggested to maintain the same instruments for the test and throughout the study.

Finally, the topic of reportable units for dPCR was revisited asking if copies/cell are acceptable to be used for the quantification of nucleic acids in biological samples instead of copies/ng DNA or copies/μL for dPCR. It was recommended to refer to the 2021 White Paper and GCC White Paper, and other guidance such as FDA biodistribution guidance which recommend using /ng and /μL [49]. Using copies/cell is complex but could be used if justified scientifically. It is important to follow existing guidance if lacking such justification. Reference genes would need to be identified and should be part of validation (precision and accuracy) and linearity.

Issues with International Reference Standards for Vaccine Clinical Assay

Vaccine immunogenicity assays are often calibrated with international standard sera to report data in international units or quantities. This standardization of units is essential when protection after vaccination is linked to antibody concentration [50]. International standards also help to standardize data reporting and reduce variation between laboratories.

Collaboration between the vaccine industry and the laboratories that produce the international standard sera is important. It is beneficial for vaccine manufacturer laboratories to participate in standardization exercises to measure the antibody titer of the standard sera. The scientific literature has widely documented the successes of using international standard sera [51–53].

However, several factors influence the consistency of the antibody concentration determination when using a vaccine immunogenicity assay calibrated with a standard serum. The assay reagents and procedure, intended use, and the nature of the standard serum are important points to consider. A critical question prior to generation of a standard serum is whether the pooled serum is from post-vaccine samples, post-natural (bacterial or viral) infections, or both. This is because antibodies in serum against natural antigenic epitopes vs. specific antigens (via mRNA or protein) may elicit various polyclonal antibodies in the serum to be used for laboratories with different vaccine modalities. A standard serum that has been made against an antigen might harmonize antibody concentration determination between different laboratories using samples from individuals that have received a homologous vaccine antigen while it increases differences between laboratories when testing samples from individuals vaccinated with a heterologous antigen. The impact of using a standard serum made from a pool of individual sera or the serum of a single donor should also be considered.

Case studies were discussed that illustrated the use of standard sera, standardizing assay procedures and reagents to harmonize vaccine immunogenicity assessment. The first case study demonstrated the impact of a change in standardization strategy. In this example, two assays were used to quantify Diphtheria (Di) antibodies elicited by vaccination: an ELISA and a Di neutralization assay. The Di neutralization assay was calibrated against the Equine Diphtheria Antitoxin standard. The ELISA was originally calibrated using an internal secondary human standard serum that had been calibrated using the Di neutralization assay as a reference. The ELISA was re-calibrated against the primary Equine standard serum. The 1st International Standard for Diphtheria Antitoxin Human (NIBSC 10/262) is intended for use as a reference preparation in assays to determine the concentration of anti-diphtheria antibody in human serum samples. The results of an international collaborative study suggested that this standard is suitable for use as a reference preparation in toxin neutralization tests and in vitro immunoassays (including ELISA). The standard has a diphtheria antitoxin potency of 2 International Units per ampoule (2 IU/ampoule). Potency was determined relative to the International Standard for Diphtheria Antitoxin Equine, DI (NIBSC code 09/204) (15 & 16).

Importantly, the change of standard had no impact on the neutralization assay calibration but had an impact on the ELISA assay calibration. When comparing results to historical data generated using a previous reference that was calibrated using a different strategy, any change in a standardization strategy may have an impact on antibody concentration results. This change in values has the potential to impact the design of vaccine clinical trials. Standardization strategies should be performed in a collaborative study including vaccine manufacturers. Human standards are preferable as they can be used in a broader range of assay technologies and most closely represent clinical samples.

A second case study demonstrated the importance of reagents. In this study, International Standards for serum antibodies to HPV type 16 & HPV type 18 were used to report antibody concentration in IUs. Both VLP-16 lots showed similar signal of the H16.V5 mAb. This study showed that the use of mAbs can facilitate calibration and are useful tools to monitor the quality of the coated VLP-16 in ELISA.

The final case study for this topic discussed the use of consensus protocols and SOPs and common reagents to reduce inter-lab variation for all 4 influenza subtypes tested compared to in-house testing. In-house testing with a matched human serum standard was as good as, if not better than, consensus testing. It also showed that biological standards provide a more practical solution but come with the issues of matching seasonal influenza virus antigenic drift. In the absence of available serum standards for seasonal influenza serology, it was recommended to test using consensus protocols.

Overall, the lessons learned were that vaccine immunogenicity assay standardization is key to support harmonization of vaccine immunogenicity assessment. The main tools to reach VIA standardization are international standardization exercises including governmental and non-governmental organizations (sponsors), reference laboratories, and laboratories from manufacturers. Partners who participate in standardization exercises can contribute to the supply of reagents, reagent qualification tools, human biological samples, designing the experiments and interpretating the data. Availability of standard sera (primary and secondary) is the primary tool to standardize antibody concentration determination through the calibration of the assays. Standardizing assay procedures, critical reagents and critical reagent qualification are also important to further improve standardization of antibody concentration determination. Change in a standardization strategy might have an impact on vaccine development programs and should be performed in collaboration between the different partners. The impact of vaccine composition and technology on vaccine immunogenicity assessment must be evaluated as the mechanistic correlate of protection might be different.

These case studies led to discussion of how to translate results into arbitrary unitage (such as IU) for an assay technology when international standards have been calibrated in another assay technology before a specific calibration in this assay is available. There was an agreement that international reference standards are needed to report data in International Units when there is an accepted CoP and to compare data between trials/products. IUs are defined by the assay used and by the standard applied. It is not adequate to compare results of two different assay technologies because they leverage the same International Reference Standard. Specifically, the antibody response post-vaccination is polyclonal and complex. Therefore, different assays may capture different components of the immune response. To compare results of two different labs, the same reference standard and two concordant assays are needed – this is particularly important for collaborative working-groups. However, concordance between two assays should be demonstrated with a common sample panel and a statistical model. Of note, different vaccines might induce different responses and mechanistic correlates of protection. This must be considered when comparing vaccine immunogenicity data.

Anti-AAV TAb Post-Dose Assessment: Development of an Efficient Analysis Strategy

AAV viral capsids are immunogenic in humans, and due to environmental exposure, there is a high prevalence of pre-existing TAb (and NAb) that are cross-reactive among AAV serotypes. Post-dose AAV TAb in patients who received AAV gene therapies has shown very high seroconversion rates and titers.

Challenges have been encountered using the traditional tier-based approach to screen, confirm and titer the drug-boosted AAV TAb response, such as requirement of large amount of capsid (for confirmatory test to demonstrate specific response to AAV) and resources to titer (each sample needs to go through titrations reaching a cut point).

Current post-dose sample analysis has operational challenges and high laboratory test burden. This includes multi-tier tests following regulatory immunogenicity guidance documents for anti-drug antibodies (ADA) to screen, confirm and titer the response. Large quantity of capsid (reagent or drug product) is needed to confirm the screening results of positive samples. A high titer TAb sample may need to be re-confirmed multiple times with additional serial dilutions for titer determination, resulting in significant increase of test. Data may not provide added information for safety-efficacy assessment, given all dosed patients are positive with high titers. Finally, blood collection, processing, and shipment has posed a burden to patients, investigators, and healthcare systems, with limited benefit on risk-efficacy assessment.

Alternative approaches for efficient and practical testing were proposed and discussed among industry, CROs and regulators to generate AAV TAb data that is informative as part of risk-benefit assessment in patients. Single, IV-dosed AAV gene therapy was presented as the focus in case studies. Locally administered AAV for ocular, muscle, or CNS therapies were not part of the discussion.

Current data has shown that AAV capsids are highly immunogenic as a foreign viral antigen and pre-existing antibodies are prevalent (∼45–65%) in humans [54]. Therefore, exclusion of individuals with positive or high titer pre-existing TAb (or NAb) for cut point assessment is needed. Post-dose positivity is 100% in subjects (regardless of pre-dose TAb status) with very high titers. AAV TAb may be cross-reactive among various AAV serotypes due to high protein sequence homology. In addition, NHP AAV TAb showed similar characteristics as human data.

In one key study, AAV5 TAb showed high titer response in all treated patients [55]. AAV5 capsid elicited robust TAb responses and titers were reaching ∼9 million All patients were positive at week 8 (earliest post dose time point in the study and demonstrated a sustained response with no reductions in titer observed up to 3 years post dose, and this persistence appears to be independent of any apparent dose relationship. Ultimately, the AAV5 TAb testing effort had limited clinical utility. It was not related to transient ALT elevations for liver function. The NAb (TI, transduction inhibition) correlation to efficacy was unclear.

Therefore, with consistently observed high seroconversion rates and high titer post-dose AAV total antibody (TAb to any serotype) responses from intravenous dosed gene therapies, the discussion topic centered around the need, or lack thereof, to test post-dose AAV TAb. There was agreement that the need is program specific for sponsors to assess risk. There is no one-size-fits-all approach with data even though no clear signals of risk are seen with current experience. Once sponsors have supporting data, it was recommended to ask regulators.

If there is a need for AAV TAb assessment, recommendations were provided for how to perform efficient testing. Options are to perform sparse sampling and archive samples for evaluation when patients have clinically relevant safety events suggesting AAV TAb as a potential mechanism. Another option is to perform the screening assay alone, classifying the results as negative or positive (+, ++, +++), with no need for confirmatory and no titer assay. The last option is to screen the samples with no confirmatory test but include a titer assessment. This involves a pre-dilution step, such as 1/100, and a larger serial dilution factor, such as 1/5 or 1/10, with upper dilution limit. The top titer is reported as >100,000. Alternatively, the laboratory has the option to presume the sample TAb positivity, bypassing the screening and confirmatory testing and proceeding directly to titer. In the cases of high titer, the same sample titration strategy as mentioned in the last option can be implemented. It is sufficient to report as the top titer without further testing. Additionally, samples should be banked for more comprehensive testing in the future.

Further Considerations on LNP Immunogenicity

Lipid nanoparticles (LNP) are nanoscale carriers of RNA molecules that can be used for RNA interference, mRNA delivery, mRNA-based vaccines and gene therapies. These RNA molecules are encapsulated in LNP that have various lipid components, including PEG-ylated lipids. The 2021 White Paper in Bioanalysis [30] addressed the application of LNP for gene-based therapeutics (GTx) and the challenges related to bioanalysis and immunogenicity of gene editing components, such as LNP and transgene. The 2022 White Paper in Bioanalysis provided additional case studies and strategies for assessment based on this data [33]. There was an agreement that patients who received biologics conjugated with PEG have shown anti-PEG antibodies which continue to increase in prevalence. Using low molecular weight PEG in LNP can lower the risk of immunogenicity in contrast to higher MW PEGs in pegylated biologics. In addition, stable PEG is in general more immunogenic than biodegradable PEG (by desire) with much shorter half-life in plasma. Before testing in the clinic, it was recommended to evaluate the risk and monitor the effects in a suitable preclinical model. In early-stage clinical studies, samples can be stored and may be analyzed if there is a clinical safety signal that may indicate immunogenicity.

The case studies discussed in 2023 highlighted the challenges of anti-PEG antibody methods. Anti-PEG antibodies are common and pre-existing in humans (reported from studies up to 76%) [56]. Identification of samples from treatment naïve subjects for cut point assessments is challenging. Titer methods to determine minimum significant response (MSR) limits to treatment boosted responses may be needed for cut point assessments. Further, testing may not follow conventional tiered approach (e.g., direct testing in titer assay, sample pretreatments). It was shown that it is important to monitor for treatment boosted responses and correlation to clinical events. Characterization of anti-PEG antibodies such as isotyping may aid in understanding causes of hypersensitivity or other responses to drug formulation products, but are not required.

With an increased prevalence of pre-existing anti-PEG antibodies, it was discussed if these assays should be developed and validated more like vaccine assays. Further, it was asked if there is clinical relevance of assessing anti-PEG antibodies and titers in mRNA vaccines. LNP are based on PEG-lipids and are an essential component of mRNA vaccines. More investigation and data are needed to evaluate clinical impact (safety or efficacy) of anti-PEGs antibodies from a large enough vaccinated human population who has received COVID-19 mRNA vaccines [57–59]. Therefore, the need for these assays and their format may need to be reconsidered. A risk-assessment should guide decision of the need of the assay and the assay format. No risk on safety and efficacy may indicate that no anti-PEG testing is required. Therefore, with current available safety data, using the tiered approach for testing anti-PEG of LNP vaccines may be reconsidered. Importantly however, if a new vaccine platform is being tested leveraging LNP, it is crucial to start with risk assessment that will guide the assay development and validation strategy.

Vaccine Clinical Assays Life Cycle Management

Many bioanalytical assays for vaccines need to be maintained for the whole product life cycle, from clinical trials to market approval to post-marketing studies, such as co-administration studies with new vaccines in clinical development. Therefore, these assays need to be well characterized, controlled and their performance monitored for many years or decades. Some of the challenges in vaccine life cycle management are long-term support of instrument platforms, re-supply of critical reagents and consumables (e.g., cells for neutralizing antibodies, capture and detection antibodies for ELISA assays, ECL plates), with consistent quality, and maintenance strategy in years without active clinical testing. It is possible for assays to outlive the instrument technology or businesses that support the chosen platform technology, therefore, as new technology emerges, legacy assays should be updated and comply with new regulatory guidance. If vendors discontinue the production of the above components for the methods, new assays may be warranted without the ability to demonstrate comparability between the existing method and the new method. Archived samples as “proficiency” samples are critical to demonstrate new assay intended use for cross validation.

Recommendations for monitoring of immune assays for vaccines were first covered in the 2021 White Paper for Bioanalysis [33]. Prior recommendations included immunogenicity assays, functional activity using multiplexing methods or new technologies to improve efficiency and reduce sample volumes. When using novel technologies, regulatory agencies may offer consultations and scientific advice and sponsors are encouraged to take advantage of the service early in development. In essence, health authorities want manufacturers to use a scientifically justified rationale to select assay methods that ensure the assay is suitable for its intended use, rather than using a “gold standard” assay that may not be the best scientific choice.

These recommendations evolved in 2023 with additional case studies of assay modernization. First, a new Hepatitis A Virus Electrochemiluminescence (HAV ECL) assay was developed to replace the old Hepatitis A Virus Enzyme Immunoassay (HAV EIA), which was technically challenging and expensive. The new assay used a design of experiment (DOE) method to optimize the assay conditions and reduce the sample and reagent volumes. The new assay met all the validation criteria and was approved by the FDA without comment for the detection of total anti-Hepatitis A antibodies in human serum in subjects from epidemiology studies or vaccine clinical trials.

Next, a diagnostic ECL kit that measures IgG antibodies against Haemophilus influenzae type b capsular polysaccharide in human serum was selected to replace a legacy radioimmunoassay. The kit, Vacczyme^™ Human Anti-Haemophilus influenzae Type b Enzyme Immunoassay, was validated bioanalytically with an LLOQ of 0.15 μg/mL. Acceptable performance was demonstrated around the 0.15 μg/mL and 1 μg/mL for short-term and long-term seropositivity and seroconversion levels.

By changing mindsets to modernize clinical assays through innovation, bioanalytical scientists can continue to support life-saving vaccines that enable human health. The extent of bridging and cross validation for new platforms was discussed. There was an agreement that if a new assay is developed to substitute a previous one, a full validation with a concordance experiment should be performed. If the changes between old and new assays are minor, bridging may be sufficient. Furthermore, the extent of bridging depends on differences between the old and new assay and the purpose of assay (e.g., primary endpoint vs. exploratory endpoint). It is not uncommon that old and new assays are not concordant. In these cases, an attempt to provide a scientific explanation of the causes should be proposed, as well as sound justification of the new assay for the intended use. For validation experiments, the entire range of the assay should be tested. Studies to establish a correlate of protection (CoP) require careful design and are often technically challenging. Consulting your regulatory authority early in clinical development should be a priority to ensure the appropriate study design.

On the other hand, assays can be offline for many years before the assay needs to be used again. Re-validation or partial validation may be required if many critical reagents and controls have expired. In some cases, sample results can be compared to historical data. A legacy proficiency panel and demonstration of assay control over the long term may be used to compare back to original assay performance.

AAV TAb/NAb Assessment

Gene therapies using viral vectors need assays to measure immune responses to the vector and the transgene protein. These assays include TAb or NAb assays, which monitor the capacity of a biological matrix to bind to the vector or block the vector from transducing target cells. Pre-existing antibodies to viral vectors can compromise the safety and efficacy of gene therapies. The incidence of pre-existing antibodies varies by capsid serotype, geographic region, and age. They may or may not affect the clinical outcome depending on the dose and titer. Pre-existing antibodies may prevent repeat dosing with the same or similar vector. Sponsors may exclude patients with high levels of pre-existing TAb or NAb [60].

Previous White Papers in Bioanalysis have compared the advantages and disadvantages of TAb and NAb assays [30,33]. Several resources have also provided recommendations for ADA or NAb assay development and validation [61,62]. In 2023, additional case studies were presented, proposing a fit-for-purpose sample testing strategy for anti-AAV ADA and NAb assessment.

A case study for TAb assays reported the development of an anti-AAV9 antibody assay using the ACE format that resulted in acceptable sensitivity, drug tolerance and precision. High pre-existing reactivity was detected in most of the serum samples tested. Using only the samples with low pre-existing ADA (PEA) allowed the estimation of an acceptable Tier 1 cut point (136 ECLU). Serum with high PEA levels was a challenge for cut point estimation but could also be challenging for use in the tittering assay. It is unclear at the moment how high PEA would affect the determination of treatment-emergent ADA.

The following updated recommendations were discussed for AAV NAb assessment. The most reliable approach to prepare NAb negative pooled matrix for AAV NAb assays is screening hundreds of samples in the NAb assay and select those generating assay response lower than 20% inhibition for the preparation of pooled matrix sample. The case study also raised the question of using arbitrary 50% inhibition or assay specific cut point for AAV NAb screening and tittering assays. The recommendation was to consider establishing a statistically set cut point instead of using 50% inhibition as the therapy has a high immunogenicity risk that requires a more sensitive cut point than the generic cut point.

Another topic was the tittering strategy that could effectively reduce NAb negative pooled sample consumption and meanwhile generate reliable titer values. Instead of using neat NAb negative pool sample, the test samples can be serially diluted in assay diluent. %INH is computed by comparing the assay signal to that derived from NAb negative pool sample with the same dilution. The final question was when AAV NAb assessment should be implemented for nonclinical studies and whether antibody to transgene protein should be assessed for all GLP studies. For GLP tox studies, the NAb assays is not typically implemented. It was recommended to use total antibody binding assay which would be adequate to help interpret the PK results. Regarding ADA to transgene protein, although ADA assay for transgene protein was often required by the agency, the expert panel considered that there was not much value in preclinical studies and the determination should be risk based.

Transgene Immunogenicity Assessment

Administration of a gene therapy may also trigger an immune response to the expressed transgene protein product. Moreover, many patients in gene therapy trials have received, or are currently receiving, enzyme replacement therapy (ERT as exogenous recombinant protein similar to the transgene product). Therefore, there is a potential need to address testing for pre-existing antibodies against the transgene product, choosing the right assays (TAb/NAb etc.), and deciding if a companion diagnostic is needed.

The need for transgene immunogenicity, risk assessment strategies, and understanding the advantages/disadvantages of different assay platforms have been discussed extensively in the 2019–2022 White Papers for Bioanalysis [27,30,33,63]. The continued recommendation is that the need to assess transgene immunogenicity is evaluated on a case-by-case basis dependent on risk and data observed for safety, PK, and efficacy, taking product and patient factors into consideration. Data shows even low titer antibodies can affect transgene expression in some cases but have no effect in other cases [55,64]. The TAb assay is often preferred for its advantages in simplicity, sensitivity, throughput, and clinical relevance, as it can correlate with immunotoxicities. The need for a NAb assay that is specific for the transgene protein is evaluated on a case-by-case basis and should be focused on the mechanism and site of action. For example, for a transgene expressing an enzyme that is targeted to a tissue or only has activity in a tissue, systemic neutralizing activity may not be relevant, whereas a cellular uptake NAb assessment may be informative. ELISpot assays may also be relevant as it was recommended to assess both cellular and humoral responses to transgene if a safety or efficacy signal is anticipated. Location of expression or site of administration may also be a critical component of the risk assessment. Systemic immunogenicity may not be relevant for therapies administered in the eye, ear, or immune privileged sites (e.g., CNS). In addition, for transgene products expressed intracellularly, systemic immunogenicity may also not be informative.

To reinforce these recommendations, case studies were discussed reviewing gene therapy products already approved or under development and the transgene strategies those products employed [55,64]. For Zolgensma^™, with the survival of motor neuron transgene product, no humoral responses were detected. Luxturna^™, with a transgene expressing retinal pigment epithelium-specific 65 kDa protein, minimal data was reported. Finally, Roctavian^™, expressing factor VIII protein transgene employed a bridging ADA assay and found low levels of binding responses and no inhibitors, while for Hemegenix^™, expressing the Padua variant of factor IX, no inhibitors were observed (binding data was not performed and/or reported) [54,63]

Future discussion is needed to share experiences with the common challenges of transgene products including whether to assess inhibition of targeting or tissue uptake, performing domain mapping of anti-transgene responses, availability of reagents, and platform (immunoassay, FACS, IHC?).

New tools, such as artificial intelligence/ machine learning (AI/ML), for transgene protein immunogenicity assay development were also discussed. Anti-transgene protein antibody assays should be carefully designed by selecting the proper critical reagents, including surrogate expressed proteins. Expressed surrogate proteins from different expression systems can show different immunogenicity profiles of the transgene protein in human samples, which could over- or under-detect ADAs of transgene proteins in clinical studies. In this case, proteins expressed from mammalian cells are preferred over that from E.coli or insect cells. As assay critical reagents, the surrogate protein can also impact the detection of ADA, depending on whether the protein composed of subunits is intactly expressed or formed by chemically cross-linked subunits. Pre-existing autoantibodies that are often present in disease samples can impact the statistical determination of cut points, which will significantly affect the performance of the assay, including assay sensitivity and drug tolerance level of transgene proteins. Outlier factor should be carefully chosen to set up proper cut points for immunogenicity assessment for clinical studies. To address these statistical challenges with cut points and drug tolerance, experience with AI/ML modeling tools to guide development were shared. A model was shown to be able to predict drug tolerance of the ADA screening assay based on the signal to background (S/N) for a selected cut point. The predicted drug tolerance from the model closely matched with the observed drug tolerance from the screening assay developed suggesting AI/ML has great potential to assist assay development.

These case studies led to renewed discussion on transgene assay strategy and assay development challenges. Regarding the question of whether there is a need to perform immunogenicity assessment and deep characterization for transgenes, the expert panel reinforced the prior recommendation of performing immunogenicity risk assessment taking into account factors such as but not limited to dose, route of administration, safety, or efficacy risks etc. It is recommended to collect samples and perform analysis based on scientific rational and immunogenicity risk assessment. Justification can be made to limit or avoid sample analysis dependent on risk assessment and overall value.

The role of CDx for pre-existing anti-AAV or transgene product expression, and also post-dose testing for anti-AAV was discussed. There was agreement that CDx assays for pre-existing anti-AAV would be required if used for enrollment. However, the need should not be generalized in reference to capsid type and population and evaluated case by case depending on aforementioned factors. For monitoring transgene product expression, flow cytometry is preferred as PCR could detect dead cells. Post-dose immunogenicity will follow the principles as discussed earlier in this White Paper.

Another discussion topic was whether for immunogenicity testing to detect anti-transgene product antibodies, it is needed to follow the same conservative analysis paradigm that we use for protein therapeutics (screen, confirm, titer, NAb, 5% false positive, 1% false positive). There were questions regarding the value in continuing to utilize the current approach for low-risk transgene products, but the risk assessment should guide the approach.

Finally, the utility of applying in silico/in vitro assessment in the immunogenicity prediction for cell and gene therapy was discussed. Many companies continue to evaluate the use of in silico and in vitro tools to predict immunogenicity. However, there is no in silico tools to predict non-protein T cell responses. While there may be value in using in silico and in vitro tools for mitigating immunogenicity risk and ranking molecules it was agreed that it has low value as a clinical immunogenicity predictive tool and traditional risk assessment methods still have priority.

Vaccine Immunogenicity Assessment

Vaccine clinical trials often use immunogenicity assays to measure primary endpoints for vaccine approval. These assays are also used to measure biomarkers (e.g., pre-vaccination serostatus (a susceptibility/risk biomarker), or immune marker of protection (a pharmacodynamic/response biomarker)). The development, qualification, and validation of these assays along with harmonization were discussed in 2020 to 2022 White Papers in Bioanalysis [27,30,33]. In 2020, it was recommended to use a phased approach to clinical vaccine assay development, separated into 3 distinct phases: 1) assay setup (establish assay format and run parameters), 2) qualification (determine assay performance), and 3) validation (confirm performance in “real life” conditions with pre-defined acceptance criteria). Specific guidance was given for each phase. Another prior recommendation was that to understand the complex immune response, one should use a systems biology approach and combine data from different tools (RNA expression, humoral and cellular immune responses) in the early stages of vaccine development, when the mechanism of action is important.

A specific case study for vaccines was shared supporting prior recommendations. This study involved characterization of immune response to an mRNA pan-respiratory (Influenza, RSV, and SARS-CoV-2) vaccine candidate. The immunogenicity strategy for this candidate included antigen binding assays and viral NAb for SARS-CoV-2 and RSV. Influenza proteins were characterized for immunogenicity with hemagglutination inhibition assays. The fit for purpose assay qualification for these assays included intra- and inter-assay precision, parallelism (for binding assays), linearity, dilution linearity, specificity, robustness, sample stability, and concordance.

Specific questions were addressed to add to prior recommendations. First, singlicate assays and acceptance criteria harmonization were discussed for bridging ADA validation and sample analysis. There were differing opinions on these topics amongst the panel, however there was agreement that more data would be beneficial. There was agreement that there was value in the use of titration assays to evaluate and that utility of singlet analysis should be demonstrated as part of assay validation.

The expert panel also shared experience with binding assays compared to NAb assays. There was extensive discussion on types of assays and format and the panel felt it important to demonstrate scientific value of either assay. With respect to CDx assays, the panel considered total antibody was more valuable to support clinical studies.

PK & Biodistribution for Replication Competent Viral Vectors

Oncolytic virus (OV) therapy is an effective and emerging treatment for some cancers and has been approved for certain indications. Most OVs can replicate and may have different pharmacokinetics than conventional drugs, as they can increase in concentration after replication and have multiple mechanisms of action [65]. This has been observed in both animal and human studies. Preclinical studies to understand the dose response relationship are an important part of the development process for a new product.

Regulatory guidance specifies that the assay should be capable of specifically detecting vector sequences in non-clinical and clinical tissues. Assays should be sensitive, specific, and reproducible and the assay range should take into account the dose-ranging studies in clinical trials [66]. This would suggest that a qRT-PCR method to detect the viral genome would be an appropriate analyte for PK. However, standard bioanalytical techniques are unable to distinguish the input drug from replicated virus and may hinder better interpretation of PK study bioanalytical results.

A case study was highlighted combining two novel approaches to characterize PK and biodistribution (BD) profile after systemic administration of an oncolytic virus (OV) in healthy mice, and to distinguish the PK profile of the input drug from replicated virus. First, to decouple input drug from secondary replication, the sponsor developed and characterized a uniquely identifiable replication-incompetent tool virus that retained important critical quality attributes of the drug. This tool distinguishes replication in blood and tissues from input virus. This approach was used in tumor bearing mice, the animal model of disease, to distinguish the contribution of virus replication which may come from the target tissue (tumor compartment). In addition, this tool was used to determine the contribution of replication to the PK and BD profiles. Second, to discriminate the genomic and antigenomic viral RNA strands contributing to replication dynamics in tissues, an in-situ hybridization method was developed using strand-specific probes to assess spatiotemporal distribution in tissues. This latter approach demonstrated that distribution, transcription, and replication localized to tissue-resident macrophages, indicating their role in the unique PK of replication competent viral vectors. Ultimately, the study resulted in a refined PK/BD profile for a replication competent OV, new proposed PK parameters, and deeper understanding of OV PK/BD, using unique approaches that could be applied to other replicating vectors.

Together this approach provided a refined understanding of the non-traditional PK profile observed with many OV therapies and may guide clinical trial design by improving interpretation of clinical bioanalytical results. Further experience with case studies of replication competent virus biodistribution and discussion was recommended for 2024.

PK & Biodistribution for Virus-vector Vaccines

In addition to immunogenicity, extensively discussed in previous White Papers, biodistribution studies are required to determine the distribution and the persistence of virus vector vaccines to target and nontarget tissues following direct in vivo administration in animals. This analysis is important to assess the overall safety profile of the product. Often, more than 40 different tissues/organs are required to be monitored for biodistribution.

Detecting the vaccine product that is based on a virus-vector that can replicate and infects tissues/organs is challenging for three reasons. First, the tissues/organs have different structures and components, so the method to break them down and release the viruses while keeping the activity of the virus needs to be optimized for each different organ. Next, there is no validated protocol to ensure the method is optimized due to a lack of a true control: not all organs from treated animals have the virus and even if they do, there is no way one can tell how many copies of the virus are in the tissue. Currently, the best option for a control tissue is to add the virus to normal tissue blocks before breaking them down, so essentially the viruses are suspended in the tissue blocks. But this does not reflect true human clinical biodistribution, as in the true sample, viruses are either deep in the tissue or even intracellular. Lastly, the tissue homogenates added to the infectivity assay are often toxic to the cells used in the assay leading to a lower estimation of the viral load

Case studies were shown attempting to address these challenges. For the challenges of different organ structures, currently there is no good solution, although there is increased interest in organoids combining with GFP-expressing virus. As for the tissue homogenate toxicity to cells, it was found that the easiest solution is to wash away the homogenate after 2 hours of incubation. This gives enough time for the infectious virus to infect the cells with minimal toxicity to the cells. There was agreement that more case studies and experience will aid future recommendations.

In the interim, recommendations were provided for the question of how often to do infectious virus infectivity assays in addition to the PCR assays. There was an agreement that guidance does refer to the need for infectivity assay for replication competent products. However, there was recognition of the challenges highlighted (e.g., poor controls, high variability, tissue homogenate toxicity to the cells in the assay).

PK & Biodistribution for CAR T Cells

Nearly 1,000 cell and gene therapy clinical trials are currently underway investigating novel therapeutic mechanisms. Reliable quantitation and localization/biodistribution of these state-of-the-art CAR-T products is paramount for successful regulatory submission. Key analytical tools (e.g., flow cytometry, PCR and multiplexed IF) employed during clinical monitoring and biodistribution of the first approved CAR-T therapy were critically reviewed with supportive patient case studies and demonstration of the challenges of these assays [67–69].

Due to the significant need for these assays, they have been continually discussed in White Papers in Bioanalysis from 2018 to present [19,27,30,32,33,63]. The initial recommendation was that flow cytometry is the primary platform used to monitor the expansion of circulating CAR-T cells and also the cellular kinetics of CAR-T cells. The 2021 White Paper compared direct measurement with flow cytometry and indirect measurement with qPCR (quantifying copy numbers). In recent years, dPCR has also been increasingly employed due to the advantages of absolute quantification (no calibrators) and 600% less input DNA for a 4-plex (600 ng vs 3200 ng). It was recommended that the choice of method should reflect the consideration of the research question at hand as well as the pros/cons of each method as exemplified by Figure 1 below.

Figure 1. Considerations for choice of flow cytometry or qPCR for CAR-T Monitoring.

In 2023, in addition to reviewing the pros and cons of flow cytometry, qPCR, and dPCR for CAR-T monitoring, case studies were discussed employing newer methods such as fluorescence ISH-IHC in a clinical workflow using patient biopsies for CAR-T biodistribution. Multiplexed ISH and IHC can demonstrate tissue homing and correlation with tumor burden via a CAR-T tumor interaction score.

Based on additional data from new platforms, the panel discussion supported prior discussion and recommendations regarding the challenges with CAR-T monitoring in clinical trials and the criteria for selection of a CAR-T monitoring assay type (e.g., flow cytometry versus qPCR and other biopsy-based platforms). Experiences and challenges were shared and recognized by regulators (complex assay development, biopsies difficult to obtain unless signal seen, reporting standards). The regulatory requirement is to collect biopsies for analysis in case of adverse events (e.g., secondary malignancy, or safety) or for correlation with efficacy. There was a consensus that biopsies beyond this are for research interest and up to sponsor’s judgement recognizing biopsies can be difficult to obtain. The prior recommendation was supported that for monitoring, flow cytometry is preferred compared to qPCR to avoid potential confounding to results from DNA released by dead or apoptosing CAR-T cells.

RECOMMENDATIONS

Below is a summary of the recommendations made during the 16^th WRIB.

The Rise of dPCR

• dPCR is preferred over qPCR when absolute quantification is not needed or when no standard curve is possible. Comparisons with qPCR continue to show advantages of dPCR for sensitivity, accuracy, and precision.

• Recent White Papers have discussed approaches for establishing the LLOQ of a dPCR assay (recommended 50 copies/μg)

  • ○
    Extraction efficiency should also be evaluated by spiking plasmid into whole blood or by performing tissue extraction for biodistribution.

Is ELISpot still the Gold Standard for Assessing Cellular Immunogenicity?

• ELISpot is among the most prevalent platforms for cellular immunogenicity assessments, but has disadvantages such as difficulties with PBMC isolation, high sample volume, sample transport logistics, challenges for cell viability and functionality, and limited multiplexing.

  • ○
    Alternate platforms exist for analysis and sample preparation. Explore and qualify alternate platforms if challenges experienced with ELISpot.
• Recommendations were provided for when to analyze cellular immunogenicity.

  • ○
    For autologous cell therapies, there was a consensus that samples should always be collected but immunogenicity assays can be run if there is a need (i.e., safety, PK, or other cellular kinetic change).
  • ○
    For allogenic cell therapies the analysis should be done.

• It is advisable to perform ELISpot assay comparisons to alternate platforms (e.g., TruCulture^®) in a single study before further implementation.

• It is not advisable to define a cut point for ELISpot for immunogenicity as background variability is too high.

NanoString Assay Development/Validation

• Since there is no regulatory guidance, the validation parameters for NanoString assay validation can follow diagnostic assay white papers.

  • ○
    RNA degradation is important to assess for stability both by RNA QC metrics and formal stability assessment.

• Acceptance criteria and number of genes from the panel to evaluate should be determined by FFP principles depending on stage of assay. Exploratory use can set criteria with a subset of genes but assays with potential CDx use should evaluate all genes on the panel.

NGS Assay Development/Validation

• Recommendations for validation are similar to NanoString assay validation.

• Acceptance criteria can be set by picking a representative panel of genes and determining sensitivity and a certain read depth.

• For detection of diverse genomic alterations (e.g., indels) the preferred wet lab control is a positive sample with dilution. Synthetic controls are more challenging.

• The need to sequence technically challenging regions depends on biological risk and in silico/in vitro predicted frequency.

  • ○
    Low risk and low frequency alterations can be omitted by providing a justification.
  • ○
    Should attempt to sequence high frequency regions multiple ways.

qPCR Assays Development/Validation

• For PK measurement of an mRNA, the reference standard should be as close as possible to the drug product (drug product itself of LNP encapsulated mRNA).

  • ○
    It is important to assess stability of the standard as well.
• Updates were provided for platform choice for PK (bDNA, qPCR, dPCR).

  • ○
    No regulatory preference but preferred not to switch across species. If switching, cross validation or comparability study may be required.
• Updated recommendations were provided for dPCR instrument standardization and technical variability especially in the absence of reference standards.

  • ○
    Perform blinded proficiency test with CROs with known positives and negatives and characterize the differences. Plasmid can be used for this test.
  • ○
    Try to maintain same instruments throughout the study.
• The recommendations for units for quantification was re-visited (e.g., copies/ng DNA, copies/μL, copies/cell).

  • ○
    Copies/ng and /μL were previously recommended. Copies/cell can be used if justified scientifically with a reference gene validated for precision, accuracy, and linearity.

Issues with International Reference Standards for Vaccine Clinical Assay

• Vaccine immunogenicity assay standardization is key to support harmonization of vaccine immunogenicity assessment.

• The main tools to reach vaccines immunogenicity assays standardization are international standardization exercises including governmental and non-governmental organizations (sponsors), reference laboratories, laboratories from the industry.

• Availability of standard sera (primary and secondary) is the primary tool to standardize antibody concentration determination through the calibration of the assays.

• Standardizing assay processes, critical reagents and critical reagent qualification are also very important to further improve standardization of antibody concentration determination.

• Change in a standardization strategy might have an impact on vaccine development programs and should be performed in collaboration between the different partners.

• Arbitrary unitage, such as IU, is defined by the assay used and by the standard applied. It may not be adequate to compare results of two different assays even with the same standard. To compare the results between two different laboratories, the same reference standard and two concordant assays are needed.

Anti-AAV TAb Post-Dose Assessment: Development of an Efficient Analysis Strategy

• AAV TAb post-dose impact assessment on PK, PD and safety has limitations since most of the dosed subjects have high titer antibodies and testing has high operational burden with limited impact of data on risk assessment.

• The need to test post-dose AAV TAb is based on risk and data specific to a program, even though low risk is seen with the current data.

• To perform testing efficiently, it was recommended to directly measure titer on a subset of samples and bank the remaining samples.

Further Considerations on LNP Immunogenicity

• Data shows that prevalence of pre-existing anti-PEG antibodies is increasing, and these assays are reconsidered to be developed and validated more like vaccine assays.

• More data is needed to determine if there is any clinical impact (safety or efficacy) of anti-PEG antibodies. Therefore, the need for these assays and their format may need to be reconsidered and the alignment of the strategy with regulatory agencies is important.

  • ○
    A risk assessment should guide decision of the need of the assay and the assay format.
  • ○
    If a new vaccine platform leveraging LNP is used, start with risk assessment that will guide the assay strategy.

Vaccine Clinical Assays Life Cycle Management

• Case studies continue to show the need and benefits of migrating legacy vaccine immunogenicity assays to modern and novel platforms.

• Prior recommendations evolved to cover the bridging and cross validation needed during migration to a new platform. The extent of bridging depends upon the assays intended use, such as a primary or secondary endpoint.

• If an entire assay is substituted, then full validation with concordance testing is needed.

  • ○
    Samples should be tested that cover the entire range of the assay.

• If changes are minor, bridging may be sufficient.

• If an assay has been offline for an extended period of time, a retirement plan should be set up, and re-validation may be needed.

AAV TAb/NAb Assessment

• Important considerations for developing AAV NAb assays were discussed.

  • ○
    Screening 100s of samples is the most reliable approach to prepare a NAb negative pooled matrix.
  • ○
    For AAV NAb screening and titering assays, it was recommended to establish a statistically set cut point.
  • ○
    To reduce NAb negative pooled sample consumption while generating reliable titer values, there was consensus to use medium to high dilution to bring samples to a manageable titer level then serially dilute 1:3.
  • ○
    For non-clinical studies (e.g., GLP tox studies, pre-clinical ADA to transgene), the utility of NAb assays is limited. Use a risk-based approach.

• There was an agreement that CDx assays for anti-AAV may be required if used for enrollment.

Transgene Immunogenicity Assessment

• Prior recommendations were reinforced that immunogenicity risk assessment should be performed to determine the need for transgene immunogenicity assessment.

  • ○
    Take account of factors such as tissue expression profile (e.g., intracellular vs systemic), dose, route and location of administration (e.g., eye, ear vs IV/SC), safety, or efficacy risks etc.
  • ○
    Samples should be banked for analysis based on scientific rationale and immunogenicity risk assessment. Justification can be made to limit or avoid sample analysis.

• The development of anti-transgene protein antibody assays need to take into consideration the critical reagents (surrogate expressed proteins, positive controls), and pre-existing anti-transgene antibodies or autoantibodies, which can impact the assay performance as well as the determination of cut points.

• Many companies continue to evaluate the use of in silico and in vitro tools to predict immunogenicity, but predictive value is uncertain. The value of continuing the current testing approach (screen, confirm, titer NAb, 5% false positive, 1% false positive) was questioned, especially for low risk transgene products. Some approved products followed the protein therapeutics testing paradigm while others have not.

Vaccine Immunogenicity Assessment

• Prior recommendations for vaccine assay development, qualification, and validation were reinforced for the 2020, 2021, and 2022 White Papers in Bioanalysis.

• There was extensive discussion on types of assays and format (e.g., bridging vs. NAb assay) and the panel felt it important to demonstrate scientific value of either assay.

  • ○
    For CDx assays, total antibody was recommended to be more valuable in the clinic.
• No consensus was reached on whether singlicate assays are acceptable for vaccine immunogenicity assessment.

  • ○
    Greater data would be beneficial, particularly for titration assays.

PK & Biodistribution for Replication Competent Viral Vectors

• Traditional bioanalytical methods such as qPCR are unable to distinguish the input drug from replicated virus and may hinder better interpretation of PK study bioanalytical results.

• Replication-incompetent tool viruses can be used to support classical PK analysis of replication-competent viruses.

• ISH-based methods can be utilized to characterize the tissue distribution, location, and replication patterns of viral vector-based drugs at the cellular level.

PK & Biodistribution for Virus-vector Vaccines

• Guidance refers to the need for infectivity assays for replication competent products including virus vector vaccines. The panel and regulators recognize the challenges from infectivity assays (e.g., poor controls, high variability, tissue homogenization challenges).

  • ○
    Attempt the best possible method with justification.

• More discussion and data are needed for further recommendations.

PK & Biodistribution for CAR-T Cells

• While flow cytometry and qPCR continue to be the predominant platforms for measurement of CAR-T cells, dPCR, and multiplex IHC ISH using biopsies are also used.

• The regulatory requirement is to assess biopsies when adverse events occur (e.g., secondary malignancy site, efficacy, or safety).

  • ○
    Biopsies beyond this are for research interest and up to sponsor’s judgement.

• The prior recommendation was supported that for monitoring, flow cytometry is preferred compared to qPCR.

SECTION 2 – Immunogenicity of Biotherapeutics

Daniela Verthelyi²¹, Robert J. Kubiak³⁰, Kelly Coble²⁰, Swati Gupta²⁶, Mohsen Rajabi Abhari²¹, Susan Richards², Yuan Song²⁷, Martin Ullmann²⁸, Boris Calderon⁶, Isabelle Cludts⁸, Fabio Garofolo⁹, George R Gunn²⁹, Shalini Gupta³⁵, Akiko Ishii-Watabe¹⁸, Mohanraj Manangeeswaran²¹, Christian Mayer¹³, Kimberly Maxfield²¹, Fred McCush¹⁰, Christine O’Day³¹, Kate Peng²⁷, Johann Poetzl³², Michele Rasamoelisolo³², Ola M. Saad²⁷, Kara Scheibner²¹, Sophie Shubow²¹, Sam Song³⁴ & Seth Thacker²¹

Authors are presented in alphabetical order of their last name, with the exception of the first 8 authors who were session chairs, working dinner facilitators or major contributors.

The affiliations can be found at the beginning of the article.

HOT TOPICS & CONSOLIDATED QUESTIONS COLLECTED FROM THE GLOBAL BIOANALYTICAL COMMUNITY

The topics detailed below were considered as the most relevant “hot topics” based on feedback collected from the 16^th WRIB attendees. They were reviewed and consolidated by globally recognized opinion leaders before being submitted for discussion during the 17^th WRIB. The background on each issue, discussions, consensus, and conclusions are in the next section and a summary of the key recommendations is provided in the final section of this manuscript.

ISR for ADA Assays: Reanalysis or Reproducibility?

Could monitoring responses from actual study samples provide a more realistic assessment of in-study performance of immunogenicity assays than using quality controls? Can these results be considered as ISR?

Strategies to Distinguish between Clinical Pre-Existing Antibodies (PEA) & Treatment Emergent Antibodies (TEA)

Is differentiation between pre-existing ADA (PE-ADA) and treatment emergent ADA (TE-ADA) in patients with PE-ADA at baseline routinely done for submission trials regardless of impact of ADA on safety/efficacy/exposure?

Universal & Generic Assays Formats for Immunogenicity Assessment

What components/tiers (e.g., cut point and outlier analysis) do we need to execute for submissions? What is needed for validation of a non-clinical ADA assay? Could a Universal Positive Control be used in both non-clinical and clinical immunogenicity assays?

Immunogenicity Assay/Reporting Harmonization

When is ADA assay cross validation necessary? How to demonstrate suitability of S/N approach vs. titer? Is S/N more sensitive and accurate compared to titer to determine the magnitude of ADA responses? How to handle the transition from titer to S/N? Since ICH M10 for PK has been completed, can we expect ADA assay validation harmonization next?

Immunogenicity Risk-based Approaches, Prediction & Mitigation: Focus on Novel Monitoring Strategies & in vitro Immunogenicity Assessment

There are still gaps in the ability to predict the likelihood of an immune response and clinical consequences due to immune response, what can we do better?

Lessons Learned from Immunogenicity Studies/Filings: Sharing Advanced Experience Related to Immunogenicity Strategies/Approaches

Assay adequacy to address clinically relevant ADA. Inadequate data interpretation by clinicians. ADA assay standardization would be ideal but is it possible in reality? Options to overcome Rheumatoid Factor (RF) interference in ADA assays (leading to false ADA results – either positive or negative – in RA patients).

Immunogenicity of Complex Biotherapeutics

How clinically meaningful is it to identify ADA binding to the respective domains by performing assays for domain specificity for registration studies? Should the bioanalysis of ADA domain binding specificity be driven by the potential clinical risk to patients? Could the immunogenicity bioanalysis strategy for registration studies focus on binding ADA to intact drug (positive/negative and titer) and only disease relevant functional domains?

Cut Point Assessment

How should the FPR be calculated? How should population specific cut points in small studies (e.g., rare diseases, pediatrics) be set?

Assay Comparability

Sponsors often transfer immunogenicity assays to CROs. How should reproducibility of sample results obtained in two different laboratories with the same assay be evaluated? What criteria should be used to ensure acceptable reproducibility?

DISCUSSIONS, CONSENSUS & CONCLUSIONS

ISR for ADA Assays: Reanalysis or Reproducibility?

To monitor how an immunogenicity assay performs during a study, positive and negative quality control samples are tested in each run. The positive quality control samples are prepared by adding surrogate anti-drug antibodies to a negative control, which is usually a healthy serum pool, and used as suitability controls. However, this may not indicate how effectively the assay can detect human anti-drug antibodies in samples from treated patients [29]. This underscores the importance of incurred sample reproducibility (ISR) for evaluation of in-study performance of ADA assays.

While re-analysis of in-study samples has been extensively discussed for other bioanalytical assays e.g., incurred sample reproducibility (ISR) for biomarkers and PK [29], it has not yet been discussed for ADA assays. ISR may be a tool for in-study monitoring of performance of immunogenicity assays, but such ISR would be quite different from quantitative assays. Differences in ISR approach for quantitative and qualitative assays are shown in Table 2. To overcome these differences, a workflow for tiered ADA testing was proposed, which could be considered ISR. During standard tiered immunogenicity analysis, approximately 5% of all samples and 100% of all screened ADA-positive samples are tested twice under identical conditions i.e., in the screening tier (T1) and in the confirmatory tier without drug (T2). In T2, these samples should have responses above the screening cut point and approximately the same signal strength as in the screening assay. A comparison between T1 and T2 responses could give more information about the assay performance than quality controls. In this workflow, the focus is on reproducibility of the screening result in terms of S/N (i.e., response in the absence of drug) [70,71]. Whenever possible, signal intensity, expressed as signal/noise (S/N)), can be used to characterize ADA levels replacing discrete titer values with continuous S/N values [72]. It is proposed to calculate the probability of repeating a positive/negative classification thus replacing categorical positive/negative values with continuous probability values calculated from the S/N values. It is also possible to roughly assess reproducibility of relative sample rankings by ADA levels (e.g., high > medium > low), which arguably is more important for assessment of ADA impact on PK, efficacy, and safety than simple reproducibility of ADA classifications measured as a percent of agreement between two tests.

Table 2. Differences between ISR for quantitative assays and for ADA.

Quantitative Assay	Immunogenicity Assay
• One test and one result per sample	• Multiple results from sequential tiers
• Easy to select samples with concentrations close to Cmax and from elimination phase	• Majority of samples are likely to be ADA-negative
• Lower limit of quantitation selected to characterize elimination phase	• Unclear what ADA level may result in clinical impact; ADA incidence and titers depend on immunogenicity of the drug and sensitivity of the assay
• Exposure is time bound	• ADA may continue to circulate after drug is no longer detectable
• Results expressed as continuous values (e.g., ng/mL)	• Results expressed as categorical values (positive vs negative) and discrete titers (e.g., <10, 10, 20, 40, etc.)
• Analyte concentrations used directly in PK, exposure-response, and safety analyses	• ADA is assessed in context of impact on PK, PD, efficacy, and safety

Case studies were shown demonstrating this approach. It was shown that for intra-assay CV = 20%, probability of reproducing the same positive classification increases from 50% at the cut point to >99% for S/N >2 × cut point. For 95% of samples, signal intensity between tiers can change from approximately 0.6- to 1.7-times the screening signal. Cohen’s weighted kappa statistic was proposed to evaluate the agreement between the screening and confirmatory tiers, where signal intensity or probability of repeating the same classification are used as ordinal variables. Weight matrices are used to assign different weights to disagreements between the screening and confirmatory results in accordance with the seriousness of the disagreements. Based on data from case studies, Cohen’s weighted kappa <0.8 indicates problematic assays performance. Cohen’s weighted kappa statistic could serve as a basis for evaluating results of incurred sample re-analysis as well as cross validation of immunogenicity assays across different laboratories. Overall, the ability to repeat signal and classification in consecutive testing tiers provides a measure of assay performance complementary to performance of assay controls.

These case studies and proposed framework led to discussion of whether monitoring responses from clinical study samples analysis provides a more realistic assessment of in-study performance of immunogenicity assays than evaluating the data from system suitability control samples, and whether these results can be considered as ISR. There was agreement on the value of this approach and there was also agreement that “reproducibility” rather than “reanalysis” referring to “ISR” for ADA assays was more appropriate. Therefore, repeatability of screening/confirmatory results could be considered for assay monitoring along with other tools such as Levy-Jennings plots. However, use of quality control samples (positive and negative) is still necessary for assessing ADA assay performance in the medium and long terms. Indeed, the proposal to use repeatability of screening/confirmatory results as performance indicator is a very short-term follow-up as screening and confirmatory assays are usually performed within a few days of each other. Another consideration is using a panel of clinical samples to analyze over time to ensure that there is no assay drift. It could also show that patients that are ADA borderline positives do not fluctuate widely between ADA positive and negative through subsequent cycles. Further data and discussion of this potential novel approach was recommended for future White Papers.

Strategies to Distinguish between Clinical Pre-Existing Antibodies (PEA) & Treatment Emergent Antibodies (TEA) for Multidomain Drugs

The design of antibodies has improved in recent years, and current antibody drugs have less side effects since they are engineered with binding characteristics that minimize off-target interactions. This makes development of multi-domain drugs more attractive, as they have more than one target and may work better than combining different drugs. There is a lot of experience with making and testing single-domain drugs, but less on how to test ADA for multi-domain drugs. For the clinical testing of ADA, there are some challenges for multi-domain drugs, such as choosing appropriate positive controls, deciding how to measure the specificity for each domain, and finding a way to distinguish between ADA that are present prior to treatment with drug (PEA) and the ADA that develop following treatment with drug (TEA).

Previous WRIB discussions on multidomain therapeutics led to the recommendation that customized bioanalytical strategies and methodologies should be planned that are based on the target biology and mechanism of action, appropriate for the phase of development, and considering key reagent availability [30]. FDA guidance also recommends that when examining immune responses to bispecific antibodies, it may be appropriate to develop multiple assays to measure specific responses to the different domains [73].

Case studies of multi-domain biotherapeutics containing scFv were provided that describe the reagents and tools along with the methods used to identify the targeted epitopes of pre-existing ADA and domain specificity characterization in clinical ADA samples. As part of the case studies, the development of an orthogonal ADA method as a strategy to differentiate pre-existing ADA from treatment emergent ADA in clinical samples from pivotal trials was shared. In Phase 1 and non-pivotal trials, the case study used a standard tiered approach (screen, confirm and titer) for identification of PEA patients. In pivotal Phase 3 trials, the case study included the addition of NAb, domain specificity and TEA assessment to distinguish TEA in post-dose positive samples from PEA positive patients. It is generally regarded that understanding domain specificity and PEA may be more helpful early in development, however, at that stage key reagents are not always available.

The bioanalytical challenges for PEA were shared including confirming the presence of PEA when there is target interference, and to identify domain specificity epitopes targeted by the PEA. Target interference was assessed using the ADA bridging method to rule out false positive signals detected in the assay due to potential dimerization of targets in matrix. Immunodepletion was used to evaluate if the positive responses observed in baseline serum from individuals were due to IgG. The method involved diluting samples (baseline individual matrix) 1:1 in binding buffer, adding them to IgG depletion spin plate, collecting flow through after incubation, and finally testing them in a bridging ADA assay.

After showing that PEA specifically targeted a specific domain (Domain B) of the multidomain therapeutic the TEA assay for the Phase 3 study was developed with a critical tool reagent modified in Domain B such that the PE-ADA epitope would be removed. This assay was used in a tiered approach with samples from PEA positive patients (pre-dose samples that screen and confirm positive for ADA to drug) undergoing a further characterization for TEA.

The strategies for PEA and TEA for multi-domain therapeutics and other biologics were also discussed. Differentiation between PEA and TEA in patients with PEA at baseline is routinely done for submission of relevant/pivotal trials. There was agreement that if the magnitude of the PEA response would mask a clinically relevant treatment emergent response, a modified assay using a bioengineered surrogate drug should be discussed in the immunogenicity risk assessment and depending on the conclusion, the modified method may be developed and implemented.

Universal & Generic Assays Formats for Immunogenicity Assessment

While it is widely understood that immunogenicity in preclinical models is not predictive of human immunogenicity, ADA assays in preclinical studies are utilized for the main purpose of confirming exposure and safety assessment (where toxicity may be observed, where efficacy may be impacted, or optimization of lead candidates). However, ADA assays take a long time to set up. Preclinical studies often use a drug with multiple backbones in a single study which requires specific assays for each backbone or species. Therefore, several steps have been taken to streamline nonclinical immunogenicity assays with generic formats [74].

Case studies were discussed to demonstrate these efforts including the development of a generic ADA (gADA) assay using the Gyros platform. This assay was robust for cynomolgus monkey, rat, and mouse samples, and used in preclinical studies to confirm adequate drug exposure. The gADA assay is inherently drug tolerant as anti-human antibodies are used to simultaneously capture the dosed (and spiked) drug and its associated ADAs. Anti-species antibodies are then added to detected bound ADAs. This format permits drug tolerance levels of about 1 mg/mL. In addition, the assay is somewhat insensitive to backbone changes although modifications are ongoing to improve selectivity in cynomolgus monkey.

The gADA assay was able to confirm a lack of ADA and brought focus on the problem of an interferent causing problems in the development of a total antibody assay. In another study, gADA was able to shed light on an antibody-drug conjugate (ADC) assay. Since this is a total ADC assay performed by LC MS/MS, interferents are not typically problematic. However, a very peculiar PK profile was observed and determined through the gADA assay, to be caused at least in part, by immune complexes to the payload. In this case, a specific ADC assay (capture: anti-ID backbone, detect: anti-payload) was then able to measure “free” drug and adapted for GLP studies. It was hypothesized that this assay is adaptable to a universal generic ADA assay for nonclinical studies. In addition to confirming exposure, this approach has also been useful in a case study for an overall PK package to deliver optimal results and inform on the best approach to use for the clinical development.

Additional case studies were shown employing human glycan specific monoclonal antibody – universal reagents that bind human IgM, IgG1, IgG2, IgG3 and IgG4 wildtype Fc (with a glycan) and do not cross-react with antibodies with a specific Fc mutation in the therapeutic candidates [75].

These case studies led to a larger discussion regarding validation of these assays (gADA and all non-clinical assays). The panel considered what components/tiers (e.g., cut point and outlier analysis) need to be executed for a submission dossier and what is needed for validation of a non-clinical ADA assay. It was agreed that non-clinical ADA validation is different from clinical ADA validation. An abbreviated approach may be acceptable as the intent of the non-clinical ADA assessment is predominately to support unexpected exposures or toxicity findings.

For gADA assays specifically, universal positive control antibodies are raised to conserved regions of the therapeutic proteins. It was considered whether the use of universal positive controls can be used in both clinical and non-clinical immunogenicity assays. There was agreement that the use of a universal positive control may be a viable alternative for non-clinical ADA assays. The use of universal positive control for clinical studies is still being investigated.

Immunogenicity Assay/Reporting Harmonization

Multiple topics for ADA assay reporting harmonization were re-visited from prior years including Signal to Noise (S/N) vs Titer and NAb assay validation.

Immunogenicity in clinical studies is generally reported qualitatively (ADA positive/negative), with additional quasi-quantitative characterization data (e.g., titer). Titer is a common method to measure ADA responses, however, has low throughput, and higher costs for immunogenic biologics. Recently, the continuous variable Signal to Noise (S/N) ratio was suggested as a less labor-intense output with which to measure the magnitude of immunogenicity response [72]. Health Authorities seem open to consider this parameter, initially in addition to titer on a case-by-case basis and recommend sponsors discuss with regulators during the development program [33].

In the 2022 White Paper in Bioanalysis, FDA presented their view on this parameter including limitations from a clinical pharmacology perspective as well as highlighting that suitability of S/N ratio to replace titer analysis needs to be demonstrated in method validation and that using either S/N or titer leads to a similar conclusion about the strength of the immunogenicity response in evaluating the impact of immunogenicity on PK (and PD, if applicable) assessed in clinical studies [34]. From the regulatory perspective, S/N approach needs more thorough comparisons with the current titer approach using different biological products to establish evidence-based support for its future use. Until then, the titer data is preferred for assessing the clinical risk of immunogenicity.

The recommendation was to demonstrate that using either S/N or titer leads to a similar conclusion about evaluating the impact of ADA on PK using data from clinical development program. To demonstrate this, it is important to choose a sufficiently sensitive dataset and conduct correlation analyses for PK vs. S/N, PK vs titer, and S/N vs titer. In addition, justification for choice of S/N to determine magnitude of ADAs should be provided in eCTD 5.3.1.4 (BA Reports), 2.7.1 (Summary of BA Methods) and 5.3.5.3 (ISI) along with discussion of S/N with Agency during program development including S/N development data. It was also highlighted that the Agency may still request titer analysis during BLA review and that samples should be banked for further investigation. Terms of use should be reflected in the ICF. Limitations were also highlighted such as whether S/N can determine if a subject had treatment boosted ADA positivity or if drug levels impact the accuracy of S/N parameter, along with other issues still to be addressed such as reporting of S/N on labels and acceptance by prescribing physicians.

The 2023 discussion included additional case studies using S/N. The main differences between S/N ratio and titer were described [76]. While S/N ratio is a continuous variable, titer results are categorical and based on the applied dilution scheme (1:2, 1:3, etc.) and therefore might be less precise. A strong correlation (R = 0.92) between S/N ratio from screening assay and corresponding titer results was shown in a clinical biosimilar PK study. In a multidose study, S/N overlaying of individual patient’s immunogenicity profiles (similar to PK profiles) allowed a sensitive analysis on similarity between biosimilar and reference. Based on the continuous nature of the ADA S/N ratios, a new parameter to assess immunogenicity on an individual subject level over time, i.e. ADA S/N AUC was used. This allowed a sensitive evaluation of the impact of immunogenicity on PK, i.e. ADA S/N AUC versus PK AUC, on a single subject level. It was demonstrated that ADA S/N AUC was superior to titer AUC with regards to the meaningfulness of ADA responses. In another case study of low immunogenic compounds, titer was inferior to S/N generated results as titer could not report or distinguish between low immunogenic response while S/N delivered precision results for all samples. In addition, MSR provides a rigorous approach for setting S/N ADA treatment boosted criterion addressing one previous concern. Case studies also showed limitations of the S/N approach however, such as wrong interpretation and estimated response magnitude when there was presence of a hook effect in the assay.

Based on these case studies, additional recommendations were provided reinforcing the 2022 recommendations. To show suitability of S/N, it was added that S/N should be correlated with titer and validated during first clinical studies. Regulatory Agencies felt it is still necessary to gather more data and publications. Regarding the unanswered question of practitioner acceptance of S/N, it was reiterated that labelling & medical communication are affected. Significant and long-lasting educational efforts are necessary to transition to S/N. However, it was also recognized that this challenge can be overcome since titer is rarely reported on a label and only in qualitative terms if it is mentioned.

Another topic addressed for harmonization was ADA and NAb assay validation harmonization. Shankar et al. and other publications pioneered a cross-industry collaboration with regulators to suggest performance and statistical criteria for ADA assay validations, making them suitable for clinical research [77–80]. Since then, with more experience, ADA methods have improved showing better drug tolerance and target tolerance, and much improved sensitivity. Regulatory expectations have also solidified guidance for ADA [81,82]. The 2019–2022 White Papers in Bioanalysis have also undertaken the effort for harmonization with interpretation of 2019 FDA guidance, examining ADA assays for complex modalities, and continued conversation of ADA method use for complex situations [27,30,63]. In 2022, an ADA validation, testing and reporting harmonization paper was published with over 40 contributors drawn from industry and the FDA [83]. The paper shared recommendations on numerous ADA method performance characteristics, reporting tables, experimental designs, and statistical tools.

Case studies showed that sponsors are implementing and undertaking efforts to harmonize internally and across the industry. For example, for determining cut points, 50 subjects x 2 analysts x 3 days were assessed by combining screening and confirmation as a first pass, saving time and the number of plates. To test positive and negative result consistency, 8 concentrations of drug and target to cover a broad range (test up to failure) were tested along with drug tolerance assessment with and without sample pre-treatment. Suggestions were given to maximize data content wherever possible by optimizing plate maps and to save time by examining performance in pre-validation prior to starting validation.

There was discussion of a potential international effort to harmonize ADA assay validation since ICH M10 PK assay validation guideline is now completed. The value of a harmonized guidance was appreciated; however, this has not been pursued to date.

Immunogenicity Risk-based Approaches, Prediction & Mitigation: Focus on Novel Monitoring Strategies & in vitro Immunogenicity Assessment

The immune response to protein, gene and cell therapies can have unintended clinical consequences. Based on this, strategies to reduce and manage the risk are applied for successful clinical development. Different in vitro tools are being used early in biologics development to evaluate the risk of developing unwanted immune responses. Regulatory bodies in general do not require non-clinical in vitro or in vivo immunogenicity studies but acknowledge the use of emerging technologies on target selection and drug design (new in silico, in vitro and in vivo models). They further acknowledge that suitable assays can be used to inform the comparative immunogenicity risk of follow-on products. These include in silico analysis (MHCII binding, human similarity, manufacturability), in vitro methods (MAPPs, cytokine release, PBMC activation and proliferation, platelet clumping), and human HLA typing.

These emerging technologies were discussed in detail in the 2021 and 2022 White Papers in Bioanalysis [30,33]. Recommendations were made that these methods are not consistently used yet but are being considered by both industry and regulators. The more robust and reliable these methods can be made, the greater reliability there will also be in the associated predictions made for immunogenicity risk. Organoid methods are particularly interesting and need more data. Adoption of in vitro immunogenicity assessment to inform valuable biosimilar development is of interest, more specifically for interchangeability. FDA is focusing research resources on this topic as part of the BsUFA III [84].

As additional data was desired, additional case studies with these tools were discussed in 2023. One case study evaluated the immune response to cross-linked extracellular matrix product and its residuals. This led to planning for an adequate monitoring strategy which considered feasibility data from the retrospective analysis of the clinical study samples, isotyping, testing of the normal healthy serum to evaluate the antibody prevalence, titer distribution and cross-reactivity to tissue.

Another case study explored using epitope mapping, T cell response assessment, and primate immunization models to investigate the cause of rare retinal vasculitis/occlusion events. The third case study centered on the release of the FDA Guidance for Industry on ANDAs [85] for certain highly purified synthetic peptide drug products that refer to listed drugs of rDNA origin [85]. The studies explored the use of a DC:T cell co-culture assay to evaluate the CD4+ T cell proliferation by flow cytometry for immunogenicity assessments. During product development, three product-related impurities were found. The impurities have amino acids modifications with possibility of introducing neo-epitopes. In silico analysis of the impurities did not suggest the presence of new T cell epitopes in the impurities when compared to the unmodified peptide. To confirm the in silico results, all 3 impurities were purified and tested using a DC:T cell assay. Many lessons were learned from the assessment including essential parameters i.e., the population sample type and size, negative and positive controls, the primary and utility of the secondary endpoints, and the reproducibility of the assay. For data analysis, criteria for cell viability, and responses to the negative and positive controls must be defined a priori. Finally, the statistical method must be carefully selected and justified to interpret the results. In this case, the CD4 T cell system provided reliable prediction to the immunogenicity risk assessment platform.

The last case study presented used major histocompatibility complex (MHC) associated peptide proteomics (MAPPs) as an ex vivo method to assess the immunogenicity risk of biotherapeutics and identify potential T-cell epitopes within the biotherapeutic sequence. While MAPPs was initially reported for the identification of MHC-I associated peptides, MHC-II presentation immunopeptidomics methodologies are more commonly being adopted now in industry specifically for the assessment of biotherapeutic drug-derived epitopes. MAPPs detection using mass spectrometry was also discussed in the 2022 White Paper in Bioanalysis [34]. In order to confidently implement this methodology in drug development the method was adapted and optimized for increased efficiency and reliability compared to previous techniques using a semi-automated workflow [86]. The method used biotin/SA-capture antibody, anti-HLA-DR mAb, and adalimumab-treated DCs. Using automated immunoaffinity capture, the assay enriches HLA-DR peptide complexes out of cell lysate, subsequently releasing bound peptides and analyzing for identification by nano-LC-MS/MS. The method improved reproducibility as demonstrated across multiple days and preparations, reduced cell requirement, and distinguished biotherapeutics with high and low immunogenicity. A high frequency of peptide and cluster presentation was observed for moderate to high immunogenic molecules while minimal presentation was observed for low immunogenic molecules.

These case studies led to further discussion on the use of immunogenicity prediction tools and remaining gaps. There was an agreement that assay sensitivity and suitability controls are critical regardless of the platform used. Harmonization of standards and benchmarks is needed. There is interest in using these assays to screen leads. Predicted immunogenicity for monoclonal antibodies is being standardized, however predicting immunogenicity for other modalities needs more work. Early assessment of biotransformation in the development process is also needed to allow time for optimizing in vitro-assays. HLA type to represent world population was recommended for incorporation but recognized to be challenging to maintain in the assay. It was finally recommended for candidate selection to screen using at least 10 donors and then analyze it in the bigger pool of donors when candidate progresses developmental stages.

Lessons Learned from Immunogenicity Studies/Filings: Sharing Advanced Experience Related to Immunogenicity Strategies/Approaches

Regulatory filings constitute a critical step in the drug development pathway towards market approval after pivotal trials of investigational drugs show positive clinical outcomes. For sponsors, the proactive preparation of clear and complete data dossiers, as well as timely addressing Health Authority (HA) questions during the review process are essential to accelerate this process and deliver drugs to patients more quickly.

One case study shared experience of the strategy whether to include missing ADA or not. In this study, there was missing ADA data from Chinese patients in a global trial planned to be filed in the US/EU and in China. Multiple options were weighed, and it was decided to file concurrently with US/EU. If global filings are acceptable with partial ADA data, the sponsor would file in China without ADA results from Chinese patients with a mitigation plan to prepare ADA analyses for China subpopulation as soon as the missing data are available. The drug was approved in this indication with an addendum CSR submitted (with China subpopulation ADA data). Submission to China HA CDE came later in same month with subsequent approval for marketing (including in China).

Another case study shared experience with measuring immune responses against mAb or Fc-fusion proteins which represents a challenge when rheumatoid factor (RF) is present in serum [87,88]. RF crosslinking with assay reagents may trigger falsely elevated signals in ADA assays. The ACE method had the greatest RF interference impact, followed by the ACE-Bridge in RA patient samples (resulting in higher false-positive signals). A novel Fab ACE-Bridge assay with monovalent recombinant Fab reagent eliminated RF interference. The lesson learned was that the use of monovalent-recombinant Fab in ADA assays can be translated into other therapeutic antibodies as a viable approach to minimize RF interference that is likely to be encountered in RA patient samples.

Case studies were shared for bispecific antibodies (bsAbs) immunogenicity assessment. Bispecific Abs contain binding motifs from 2 distinct antibodies and require more complex engineering and manufacturing processes compared with traditional monoclonal antibody therapeutics [89–91]. This may increase the potential for inducing an immune response. More rigorous assessments should be considered to adequately characterize the immunogenicity of bsAbs. Immunogenicity results should be assessed on the totality of data. Data regarding immunogenicity of bsAbs from clinical trials are limited but are actively being collected across the industry, which will continue to benefit the bsAb development.

A final discussion topic for lessons learned from immunogenicity assessments or filings was understanding if ADA assays are adequate to address clinically relevant ADA. There was agreement that ADA incidence should not be compared between different molecules, sometimes even for the same molecule but in different indications, and between biosimilars. Continued education is necessary to emphasize this in different formats such as publications to encourage companies to develop highly sensitive assays to capture ADAs for immunogenicity evaluation. It is crucial to balance the need for highly sensitive assays and reporting ADA incidence which is clinically meaningful. The consequence of failing to strike the right balance is losing the clinician’s interest in data that has no relevance for them.

Immunogenicity of Complex Biotherapeutics

Biotherapeutics are often improved by bioengineering methods that change their properties, such as their half-life and how efficiently they are absorbed. However, structural modifications to proteins and peptides, such as sequence changes, addition of linkers or cross-linking, can create new epitopes or disrupt existing ones, potentially impacting antibody formation. Additionally, biotherapeutics with extended half-life are often multi-domain molecules, which adds complexity and challenges to ADA assays, including the need for different assay formats and the development of multiple assays for different components, as well as challenges in developing reagents.

This topic was discussed in the 2022 White Paper in Bioanalysis [33]. Recommendations included that depending on the risk, reporting total ADA for multi-domain therapeutics may be sufficient if there is no clinical impact. Consultation with regulators is necessary before submission. A NAb assay that measures the effect of the drug on its target was agreed to usually be enough. Separate assays for each domain are not usually needed.

Additional case studies of ADA assessment for half-life extension biotherapeutics were shared. These studies showed that it is important to ensure adequate sampling timeframe for ADA. When reporting data, it helps to put data in context of study duration. Specifically for immunogenicity, it was beneficial to consider not only the number of doses given but also overall drug exposure. A multi-tiered approach was also demonstrated, and adverse event driven immune-related assays and sample collection were incorporated.

There was again discussion on whether it is clinically meaningful to identify ADA binding to the respective domains by performing assays for domain specificity for registration studies and if the bioanalysis of ADA domain binding specificity should be driven by the potential clinical risk to patients. Furthermore, it was asked if the immunogenicity bioanalysis strategy for registration studies could focus on binding ADA to intact drug (positive/negative and titer) and only disease relevant functional domains. With respect to health care providers, the primary concern is the immunogenicity to the drug and any relevant safety signals. If such signals are observed in early studies, then it was recommended to consider evaluating reactivity to domains. Sponsors need to understand the immunogenicity profile of their multi-domain drug which may necessitate additional assessments in early clinical phase studies and may be performed in later phase studies based on risk assessment and safety signals.

Cut Point Assessment

The cut points of ADA assays are important for the validation process, as they set the threshold to decide if a sample is negative or positive. The cut point determination has been long discussed since the 2018 White Paper in Bioanalysis [19,24,27,30,33]. Both the analytical and biological variations of the assay are relevant for setting the CP. It was previously recommended that when the means and variances of the assay runs are similar, fixed cut points can be used. When the means change between different factors (e.g., analysts, plates or assay runs), but the variances stay similar, floating cut points should be used. Other prior recommendations included alternative approaches for cut point determination such as using assay variability (minimum/precision cut point), achieving acceptable assay sensitivity (e.g., ≤100 ng/mL), and using the assay specificity curve and visualizing/gating the data. Discussion with regulatory agency was recommended for implementing alternative approaches.

In 2023, new data was shared based on recent observations that commercially available drug naïve sera often could not represent the study sample population. The study demonstrated it is necessary to confirm whether the validation CP is suitable for the population being studied in the study phase.

It was highlighted that within current industry practices, there are 5 different variations on the formula used to calculate baseline sample FPR (false positive rate) during cut point determination. The major difference among those formulas is where to exclude the confirmed positive samples in the calculation formula (from numerator only or from both numerator and denominator). In general, confirmed positive samples are excluded from both the numerator and the denominator, as was also reported in the white paper on harmonization practices for the bioanalytical immunogenicity study report [90]. However, if the sample set used for the calculation has only a very small number of confirmed baseline positive samples, using the full sample set for the denominator may be an option as it does not meaningfully alter the FPR. The specific formula used should be noted in the bioanalytical report. Significant changes from the FPR for the screening assay established during pre-study validation indicate that validation cut point is no longer appropriate and in-study cut points should be determined.

It should be noted that the during in study analysis of baseline samples, the FPR is not calculated this way since (unlike cut point determination) confirmation assay is not performed on all samples. Hence, when determining the appropriateness of the validation cut point to in-study sample analysis, the baseline screening positivity rate is typically the only metric available for comparison.

Another topic discussed was that in certain settings, such as rare diseases and pediatric patients, it is challenging to establish CP using solely the baseline samples. To establish a population-specific CP in such studies, it was recommended to refer to previous White Papers [92]. Specifically, for validation cut point (VCP) vs SSCP, it was recommended to use visual tools such as histograms and boxplots to assess the similarity of the two set of the data distributions. This should be used with formal statistical assessment of the differences of means (ANOVA) and variances (Levene’s test). In such cases, a SCP using study baseline samples supplemented with commercially available population-specific subject samples with similar characteristics, as determined by statistical analysis can be appropriate.

In addition, it was recommended to include general language in the informed consent form (ICF) to collect blood sample for “immunogenicity assessment and assay validation purposes” to allow for further characterization, including setting-up of in-study cut points, as needed. It may also be useful to consider using samples from subject/patient screening for assay validation purposes. In such case samples from subjects/patients, who meet study inclusion criteria, can be used before baseline sample become available. The advantage is the completion of the CP assessment and potentially resetting it earlier for studies with challenging timelines.

Assay Comparability

Sponsors often transfer immunogenicity assays to CROs. Regulators may find it difficult to evaluate the immunogenicity data for the intended use(s) of the product, because of the complexity of the immunogenicity assays. Therefore, a concerted effort was made in the 2021 White Paper in Bioanalysis to propose a workflow for demonstrating assay comparability and cross validation in collaboration with regulators [27]. Regulators recommended not to perform assay comparability using platform agnostic methods and recommended to not compare across different programs. Generation of an integrated dataset was recommended with a bridging studies of prior ADA positive samples. Importantly, statistical methods such as the Cohen’s kappa test are needed to show strength of agreement. Titer methods also should show 80–90% match using minimum significant ratio (MSR) statistical assessment.

Additional case studies were shown supporting these prior recommendations. This study employed the ISR for ADA strategy (repeatability) described earlier in this section. It showed that S/N ratio and probability of repeating ADA classification can be used for inter-lab comparability of negative samples which typically account for majority of responses in a clinical study. Cohen’s weighted kappa was used to provide useful statistics for objective comparison of results between different laboratories.

Following discussion of this additional case study, prior recommendations from 2021 and 2022 were supported. An additional recommendation was made that a panel of representative samples should be used to determine the reproducibility of the lab and should assess ADA status (positive/negative) and magnitude of response for ADA positive samples in the form of titer assessment or S/N ratios. In addition, consulting Clinical & Laboratory Standards (CLSI) guidelines, which also focuses on diagnostics, may be informative in assessing potential criteria to conclude successful assay transfer/cross validation.

RECOMMENDATIONS

Below is a summary of the recommendations made during the 17^th WRIB:

ISR for ADA Assays: Reanalysis or Reproducibility?

• During tiered testing some samples are tested at least twice – these data can be used to assess incurred sample “reproducibility”.

  • ○
    A workflow and statistical measures were proposed for tiered testing of actual study ADA positive samples for assay monitoring complementary to assay controls.
  • ○
    This strategy can be used to evaluate reproducibility of measurements within the same laboratory (ISR) or across different laboratories (cross validation).

• There was agreement that repeatability of screening/confirmatory results could be considered, and that “reproducibility” is more appropriate for immunogenicity assays since no additional re-testing of sample is advocated.

• An additional proposal was to use a panel of clinical samples monitored over time to ensure that there is no assay drift.

Strategies to Distinguish between Clinical Pre-Existing Antibodies (PEA) & Treatment Emergent Antibodies (TEA)

• For multi-domain therapeutics, additional tiers of characterization testing, including domain specificity and differentiation between PEA and TEA, may be required beyond standard immunogenicity testing.

• Specialized reagents and tools may be needed to develop the additional tiers of testing.

• If the magnitude of the PEA response would mask a treatment emergent response, then a modified assay should be developed using a bioengineered surrogate drug to differentiate PEA from TEA.

Universal & Generic Assays Formats for Immunogenicity Assessment

• Preclinical studies often use multiple drug constructs in a single study–- several steps have been taken to streamline nonclinical immunogenicity assays with generic assays.

• Several generic ADA assays have been developed with universal reagents that are inherently drug tolerant.

• Non-clinical ADA validation is different from clinical ADA validation,

  • ○
    An abbreviated approach may be acceptable.

• In generic non-clinical assays, the use of Universal Positive Controls may be a viable alternative. Their use in clinical studies requires more data to be shared before further recommendations are made.

Immunogenicity Assay/Reporting Harmonization

• S/N should be correlated with titer and validated during first clinical studies.

• It still is necessary to gather more data and publications on use of S/N.

• Based on the continuous nature of the S/N ratio parameter compared to categorical results obtained by titer assays, S/N ratio enables new ways to evaluate immunogenicity.

• S/N ratio can be superior to titer analysis when assessing the impact of ADAs on clinical parameters such as PK.

• Significant and long-lasting educational efforts for labeling and health care provider acceptance will be necessary to transition to S/N.

• During ADA validation, examining drug and target interference over a broad range (e.g. test to failure) is informative and frequently reduces need for further regulatory inquiry.

• Design validation experiments to maximize data content and reduce overall time when possible.

• It is still too early to pursue immunogenicity guidance in the mold of ICH M10.

Immunogenicity Risk-based Approaches, Prediction & Mitigation: Focus on Novel Monitoring Strategies & in vitro Immunogenicity Assessment

• In vitro immunogenicity risk assessment methods are invaluable tools to predict immunogenicity and can be incorporated into drug development for selection of biological candidates.

  • •
    In the absence of method standardization and regulatory recommendations, the tools require characterization to ensure robust performance and reliability.
  • •
    There is a need for standards and controls and benchmarks not just within a company but across industry.

• Both DP and DS can be used for assessment of the risk of the API. The assessment of API variants can be done individually, and their relative risk compared with the API.

• When assessing the product for innate immune response modulating impurities the assessment should be done using the drug product but the impact of formulation on the assay should be considered and mitigated.

• Biotransformation should be assessed early to allow time to optimize in vitro assays.

• When using in vitro assays to assess immunogenicity risk of a new molecular entity it is recommended to initially screen at least 10 donors, as variability is expected between blood donors. A larger pool of donors may be useful to support candidate selection.

• When using in vitro assays to assess the immunogenicity risk of product related impurities in comparative immunogenicity risk assessments, the MHC distribution of the cells should represent those in the target population. Generally, data from 30–40 donors is expected.

Lessons Learned from Immunogenicity Studies/Filings: Sharing Advanced Experience Related to Immunogenicity Strategies/Approaches

• ○ADA incidence should not be compared between different molecules.
• ○Clinicians should be educated on the fact that ADA incidence is not always correlated with clinical impact.
• ○Education of clinicians will be needed if S/N approaches are used in place of titers.
• ○The use of monovalent-recombinant Fab in ADA assays can be translated into other therapeutic antibodies as a viable approach to minimize rheumatoid factor interference.

Immunogenicity of Complex Biotherapeutics

• Sponsors need to understand the immunogenicity profile of their multi-domain drug which may include assessment in early clinical phase studies.

• In general, it is expected that the reactivity to the different domains of complex products (e.g., bispecific Abs) will be assessed.

• For pivotal clinical studies, the performance of assays to understand the immunogenicity of individual domains is expected but there could be regulatory attenuation of guidance on an as-needed basis based on risk assessment and safety signals.

Cut Point Assessment

• There are several approaches for baseline sample FPR calculation formulas, but the general recommendation is to remove confirmed positives from the calculation, however if the number of confirmed positives is small, can include the full sample set in the denominator.

• In cases when it is challenging to establish a cut point with baseline samples due to sample availability, it may be possible to supplement with samples from subjects with similar characteristics [92].

Assay Comparability

• Prior 2021 and 2022 recommendations were reinforced to show ADA assay comparability by performing bridging using prior ADA samples.

  • ○
    A panel of representative samples should be used to determine the reproducibility of the lab and should assess ADA status (positive/negative) and magnitude of response for ADA positive samples.
  • ○
    Statistical methods such as Cohen’s kappa test and minimum significant ratio may be useful.

• In addition, Clinical & Laboratory Standards (CLSI) guidelines, which focuses on diagnostics, may be informative in assessing potential criteria to assess assay comparability.

Financial disclosure

The authors have no financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

Competing interests disclosure

The authors have no competing interests or relevant affiliations with any organization or entity with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, stock ownership or options and expert testimony.

Writing disclosure

No writing assistance was utilized in the production of this manuscript.