ADC Bioanalysis Knowledge Base

EMA_MABEL_FIH

7 min read

EMA Guideline on Strategies for First-in-Human Trials / MABEL (2018 Rev.1)

概述 Overview

Document: Guideline on Strategies to Identify and Mitigate Risks for First-in-Human and Early Clinical Trials with Investigational Medicinal Products (Revision 1)
Reference: EMEA/CHMP/SWP/28367/07 Rev.1
Agency: EMA (CHMP / Safety Working Party)
Effective date: 1 February 2018
Status: Final (active)
Source: https://www.ema.europa.eu/en/strategies-identify-mitigate-risks-first-human-early-clinical-trials-investigational-medicinal-products-scientific-guideline
PDF: https://www.ema.europa.eu/en/documents/scientific-guideline/guideline-strategies-identify-and-mitigate-risks-first-human-and-early-clinical-trials-investigational-medicinal-products-revision-1_en.pdf

This guideline was developed following the TGN1412 monoclonal antibody trial catastrophe (2006) in which healthy volunteers experienced life-threatening cytokine storm at doses below the NOAEL. It introduced the MABEL concept as a mandatory consideration for high-risk biologics entering first-in-human (FIH) trials.

核心要点 Key Points

1. Risk Stratification — When Is MABEL Required?

Not all investigational products require MABEL-based dose selection. The guideline defines a risk classification approach:

High-risk products (MABEL mandatory):

  • Immunomodulatory biologics with pleiotropic immune stimulatory effects (e.g., CD28 superagonists, cytokine agonists)
  • Products with novel mechanisms of action with uncertain in vitro-to-in vivo translatability
  • Products where nonclinical models are unreliable predictors of human toxicity

Standard-risk products (NOAEL/HED sufficient):

  • Cytotoxic molecules (including ADC payloads) where the mechanism is well understood
  • Products with established pharmacological class with predictable toxicity profiles

ADC classification: ADCs generally fall in the standard-risk category for the FIH starting dose calculation (NOAEL/10 approach) because the cytotoxic payload has a well-understood mechanism. However, the antibody component of a bispecific ADC or an ADC with an immunomodulatory antibody backbone (e.g., anti-CD3 bispecific, checkpoint-activating antibody) may require MABEL consideration.

Practical guidance: Consult with EMA at Scientific Advice (Type B meeting equivalent) if the ADC:

  • Uses a T-cell-engaging or immune-activating antibody backbone
  • Has a novel payload with uncertain potency-to-dose relationship
  • Targets an antigen expressed on circulating immune cells (CD3, CD28, CD40L)

2. MABEL Calculation — Methods

Definition: MABEL = anticipated dose leading to a minimal biological effect level in humans. It is derived from all available pharmacological data:

Method 1 — Receptor occupancy (RO)-based:

  1. Determine RO-response relationship from in vitro binding studies (Kd from SPR or ELISA)
  2. Target minimum pharmacologically active RO (typically 1–10% for agonists; defined differently for antagonists)
  3. Project plasma concentration needed to achieve target RO in humans = MABEL concentration
  4. Convert to dose using projected human Vd and CL (from allometric scaling)
MABEL concentration = Kd × (target RO%) / (100% - target RO%)
MABEL dose = MABEL concentration × Vd_human

Method 2 — In vitro potency-based:

  1. Determine minimum concentration producing measurable effect in the most relevant in vitro assay (e.g., minimum ADC concentration causing >10% cell kill in antigen-positive cell line)
  2. Use the in vitro EC₁₀ or IC₁₀ as MABEL concentration
  3. Apply a 100-fold safety factor to convert to clinical starting dose (because in vitro does not fully predict in vivo)

Method 3 — In vivo animal MABEL:

  1. Identify minimum effective dose (MED) in the most sensitive relevant animal model (xenograft efficacy or PD biomarker study)
  2. Convert MED to human equivalent dose by allometric scaling
  3. Apply appropriate safety factor (typically 10–100× for MABEL-derived starting dose)

For ADCs: MABEL is typically derived from the minimum TGI dose in xenograft studies, converted to human via allometric scaling. This often gives a starting dose similar to or slightly below the NOAEL/10 approach.

Starting dose selection rule:

Starting dose = LOWER of:
  (NOAEL in most sensitive species) × (conversion factor to HED) ÷ (safety factor, typically 10)
  OR
  MABEL-derived dose ÷ (appropriate safety factor)

3. Dose Escalation Principles for FIH

The guideline describes key principles for managing escalation risk:

Sentinel dosing: In Phase 1 trials with high-risk products, dose the first patient(s) with an observation period (≥24 h or until next infusion) before dosing the rest of the cohort. Most FIH ADC trials use sentinel dosing for cohort 1 (1–2 sentinels, observed for 24–48 h before remaining patients in the cohort).

Dose escalation scheme:

  • Modified Fibonacci sequence (100%, 50%, 33%, 25%, 25% increases) for cytotoxic ADCs
  • More aggressive escalation (doubling) acceptable at sub-pharmacological dose levels below MED
  • Accelerated titration: single patient per cohort with pharmacokinetic bridging acceptable if low-risk based on nonclinical data

Stopping rules: Pre-specified DLT criteria and stopping rules must be in the protocol:

  • Grade ≥3 non-hematologic toxicity (excluding nausea/vomiting/alopecia)
  • Grade ≥4 hematologic toxicity
  • Any Grade 2 toxicity that does not resolve to Grade 0–1 within 14 days
  • Specific safety signals for ADC payload class (e.g., Grade ≥2 neuropathy for MMAE-ADCs; Grade ≥2 ILD for DXd-ADCs)

4. EMA-Specific Requirements vs. FDA

AspectEMA (this guideline)FDA (2005 FIH guidance)
Risk classificationFormal high-risk vs. standard-risk classificationLess formal; MABEL discussed but not mandated as separate concept
MABELRequired for high-risk; consider for othersRecommended but not a separate mandatory calculation
Sentinel dosingStrongly recommended for high-riskCase-by-case
Biomarker planRequired as part of FIH protocolRecommended
Dose escalation designAdaptive designs encouragedSame

EMA Reflection Paper on ADCs (Draft 2020 / Final pending)

Note: EMA published a draft Reflection Paper specifically on ADCs (EMA/CHMP reference, ~2020) addressing:

  • CMC/quality considerations for ADC manufacturing (DAR characterization, linker stability testing)
  • Nonclinical: dual S9 + S6(R1) framework; surrogate ADC studies
  • Clinical: analyte selection (aligns with FDA ADC ClinPharm 2024 — TAb + cAb + free payload); E-R expectations
  • Bioanalysis: LBA + hybrid LBA-MS acceptable; critical reagent requirements

Status as of knowledge cutoff: Draft; comments were received 2021. Final publication was expected 2024–2025. Check EMA website for latest status before citing in regulatory submissions.

常见问题和挑战，具体案例和解决方案

Challenge 1: ADC Starting Dose — NOAEL/10 Gives a Sub-Efficacious Dose. Can We Start Higher?

Problem: NOAEL in monkey is 6 mg/kg (Q3W). NOAEL/10 = 0.6 mg/kg starting dose. Non-clinical xenograft data suggests the minimum efficacious dose is 3 mg/kg. Starting at 0.6 mg/kg means the first 4–5 cohorts will be below any pharmacologically active dose.

EMA/FDA position: Starting dose must prioritize patient safety over speed. However, both agencies encourage PK/PD-guided accelerated escalation in sub-pharmacological dose ranges:

Solution:

  1. Start at NOAEL/10 (0.6 mg/kg) with accelerated escalation (single patient per cohort, doubling) until the first TKI dose (projected therapeutic range based on allometric scaling of xenograft TGI threshold)
  2. Switch to modified Fibonacci (50% → 33% → 25% increments) once in the pharmacologically active range
  3. Include PK sampling at each cohort to confirm exposure is tracking allometric predictions — if PK matches, evidence of pharmacological activity begins at projected dose

Case: T-DXd Phase 1 (DS8201-A-J101) — started at 0.8 mg/kg; accelerated to 4 mg/kg over 5 cohorts; then conventional escalation to 8 mg/kg. Selected 6.4 mg/kg (now 5.4 mg/kg per labeling update based on DESTINY trials) and 8 mg/kg for expansion cohorts.

Challenge 2: No Clinical PK Data from a Related ADC — How to Project Human PK for MABEL?

Problem: This is a first-in-class ADC with a novel antibody and novel payload. No approved ADC uses the same linker-payload combination. How do we project human PK and Vd for MABEL dose calculation?

Solution:

  1. Antibody component: Use IgG1 population parameters (CL ~3–5 mL/day/kg, Vd ~50–80 mL/kg) as prior; refine with monkey PK from TK study using allometric scaling (BW scaling exponent 0.75 for CL, 1.0 for Vd)
  2. Payload component: Use known payload physicochemical properties (MW, logP, protein binding) + in vitro clearance (microsomal stability) to project human CL using IVIVE (in vitro to in vivo extrapolation)
  3. Conduct a pre-IND monkey PK/TK study to measure actual monkey PK → scale to human
  4. Apply 3× uncertainty factor to projected human MABEL concentration to account for inter-species translation uncertainty

Graph View

Backlinks