Predicting the first-in-human (FIH) dose for AAV gene therapy

 Predicting the first-in-human (FIH) dose for AAV gene therapy from non-human primates, such as cynomolgus monkeys (cynos), involves careful extrapolation to ensure safety while aiming for therapeutic efficacy. Cynos are commonly used for AAV studies because of their physiological and immunological similarities to humans, making them an effective model for dose prediction. Here’s a structured approach to predicting the FIH dose for AAV gene therapy based on cyno data:

1. Determining the Relevant Cyno Dose

• Therapeutic Efficacy Threshold: In cyno studies, a dose that achieves therapeutic efficacy is identified first. For AAV gene therapy, this could mean achieving a target transgene expression level or a specific biomarker change in the target tissue.
• No Observed Adverse Effect Level (NOAEL): The NOAEL is the highest dose in cynos that does not produce observable adverse effects. It’s crucial to identify this dose to balance efficacy with safety in humans.
• Maximum Tolerated Dose (MTD): If toxicities are observed at doses above the therapeutic threshold, the MTD helps define the upper limit for safe dosing.

2. Selecting an Appropriate Allometric Scaling Method

To extrapolate from cynos to humans, several allometric scaling methods can be applied:

(a) Body Weight Scaling

• Method: The dose in vector genomes per kilogram (vg/kg) is scaled directly based on body weight.
• Formula: $$\text{Dose}_{\text{Human}} = \text{Dose}_{\text{Cyno}} \times \left( \frac{\text{Weight}_{\text{Human}}}{\text{Weight}_{\text{Cyno}}} \right)^{0.67}$$
• Application: This method is straightforward but may not always fully capture human pharmacokinetics or the biodistribution differences for AAVs.

(b) Body Surface Area (BSA) Scaling

• Method: BSA scaling is often preferred for gene therapy, as BSA more accurately reflects metabolic scaling across species. For AAV, it may help in approximating systemic exposure and transduction efficiency.
• Formula: $$\text{Dose}_{\text{Human}} = \text{Dose}_{\text{Cyno}} \times \left( \frac{\text{BSA}_{\text{Human}}}{\text{BSA}_{\text{Cyno}}} \right)$$
• Application: BSA scaling is generally used for biologics and is considered conservative, making it suitable for safety-focused FIH dose predictions.

(c) Physiological Scaling (PBPK Modeling)

• Method: Physiologically Based Pharmacokinetic (PBPK) modeling incorporates species-specific anatomical, physiological, and biochemical parameters to better predict AAV distribution and clearance. PBPK models often take into account liver function, immune responses, and vector biodistribution.
• Application: This method is more complex and resource-intensive but can provide a more precise prediction, especially for therapies targeting specific organs or tissues.

3. Adjusting for Immune Responses and Neutralizing Antibodies

• Pre-existing Immunity: Humans often have pre-existing immunity against AAV serotypes, which can neutralize the vector. Adjustments to the dose may be necessary to ensure enough vector reaches target tissues.
• Immune Suppression: Some protocols use immune-suppressing regimens to mitigate immune responses, which could impact dose requirements. Immunosuppressive treatments in the cyno model should be carefully considered to match human protocols if used.

4. Applying Safety Factors for FIH Dose Prediction

Given the potential immunogenicity and toxicity risks, a safety factor is applied to the scaled dose to further ensure safety in humans:

• Conservative Safety Factor (10- to 50-fold): Typically, a 10- to 50-fold safety factor is applied to the NOAEL from cynos, depending on the anticipated risk and uncertainty in dose-response translation.
• Choosing the Factor: For novel AAV serotypes, high doses, or target organs with higher toxicity risk (like the liver), a higher safety factor is recommended (20- to 50-fold). For more well-characterized AAV vectors with extensive preclinical data, a lower safety factor (10- to 20-fold) may be sufficient.

5. Example Calculation for FIH Dose Prediction

Let’s assume:

• Efficacy Dose in Cynos: 1 x 10^13 vg/kg
• NOAEL in Cynos: 5 x 10^13 vg/kg
• Body Weight of Cyno: ~5 kg
• Body Weight of Human (Average): ~70 kg
• Chosen Safety Factor: 10-fold (conservative)

Step 1: Calculate Scaled Human Dose Using Body Weight Scaling

Using the NOAEL as a reference and a scaling factor of $0.67$:

$$\text{Dose}_{\text{Human}} = 5 \times 10^{13} \, \text{vg/kg} \times \left( \frac{70}{5} \right)^{0.67} \approx 2.92 \times 10^{14} \, \text{vg/kg}$$

Step 2: Apply Safety Factor

$$\text{FIH Dose} = \frac{2.92 \times 10^{14} \, \text{vg/kg}}{10} = 2.92 \times 10^{13} \, \text{vg/kg}$$

Thus, the estimated FIH dose would be 2.92 x 10^13 vg/kg, subject to adjustment based on additional risk factors or immunosuppression protocols.

6. Considerations for FIH Trial Design

• Dose Escalation: Starting with the conservative FIH dose, a dose-escalation study may follow to identify the therapeutic window.
• Patient Selection and Monitoring: Pre-screening for anti-AAV antibodies, close monitoring of liver function, and immune responses are critical for patient safety.
• Endpoints: Assessing early biomarkers, transgene expression levels, and clinical outcomes helps establish a safe and effective dose range.

In conclusion, predicting the FIH dose from cyno data involves carefully balancing efficacy, safety, and immunogenicity considerations, using allometric scaling with an appropriate safety factor to ensure a cautious approach in the FIH trial.