How we use size‑exclusion membranes with 20 nm pores to retain viruses while allowing the monoclonal antibody to pass.
Welcome back to our deep‑dive series on monoclonal antibody manufacturing. In Step 8, we used anion exchange chromatography (AEX) in flow‑through mode to remove DNA, endotoxins, and residual viruses. The AEX pool is now highly pure, but regulators require two orthogonal viral clearance mechanisms – and AEX, while excellent for many viruses, is based on electrostatic binding. What if a virus has an atypical surface charge or mutates to reduce binding?
Viral Filtration (also called nanofiltration or virus removal filtration) provides a completely different, size‑based mechanism. It is a physical barrier: a membrane with pores small enough to retain viruses (typically 20 nm) but large enough to let the mAb (hydrodynamic diameter ~5–10 nm) pass through.
This step is the final safety guard before the drug substance is formulated. Let’s explore the technology, step‑by‑step operation, critical parameters, and validation requirements.
Overview: What Is Viral Filtration?
Principle: Viral filters are asymmetric, hydrophilic membranes (typically made of polyethersulfone, regenerated cellulose, or polyvinylidene fluoride) with a tight pore size rating (15–50 nm, most commonly 20 nm). They operate in normal flow filtration (NFF, also called dead‑end) mode, where the entire feed is pushed through the membrane.
How it works:
Viruses are large: enveloped viruses (e.g., HIV, influenza) are 80–120 nm; non‑enveloped viruses (e.g., parvovirus, reovirus) are 18–30 nm. A 20 nm pore filter retains viruses larger than 20 nm. For smaller viruses (e.g., minute virus of mice, MVM, ~18–20 nm), retention is based on a combination of size exclusion and depth entrapment.
Monoclonal antibodies (IgG, ~150 kDa) have a hydrodynamic diameter of ~5–10 nm, easily passing through 20 nm pores.
Why is this step critical?
It is orthogonal to low pH inactivation (chemical) and chromatography (binding). Physical size exclusion works regardless of virus surface properties.
It removes both enveloped and non‑enveloped viruses, including those resistant to low pH.
Regulatory guidelines (ICH Q5A) expect a dedicated virus removal step, often nanofiltration, for mammalian cell‑derived biologics.
Step 9.1: Viral Filter Selection
Choosing the right filter is critical. Key factors:
| Filter Brand | Membrane Material | Pore Size | Flow Rate | Typical Application |
|---|---|---|---|---|
| Planova 20N (Asahi Kasei) | Regenerated cellulose | 20 nm | Low to moderate | Gold standard for mAbs |
| Planova 15N | Regenerated cellulose | 15 nm | Lower | For smaller viruses (parvovirus) |
| Viresolve Pro (Millipore) | Modified PES | 20 nm | High | High throughput, modern option |
| Virosart CPV (Sartorius) | PES | 20 nm | Moderate | For parvovirus retention |
| Ultipor DV20 (Pall) | PES | 20 nm | Moderate | Legacy option |
Selection criteria:
Virus retention claim: Must retain the smallest relevant virus (e.g., MVM, 18–20 nm) to >4 log.
Compatibility: With product buffer (pH, ionic strength, excipients).
Flux and throughput: Higher is better for manufacturing economics.
Pre‑filter requirement: Most viral filters require a pre‑filter to protect them from aggregates or particulates.
For a Humira‑like process, Planova 20N or Viresolve Pro are common choices.
Step 9.2: Step‑by‑Step Operation
Viral filtration is typically performed after AEX and before UF/DF (formulation). The AEX pool is already in a suitable buffer (e.g., 20 mM Tris, pH 7.0–8.0, low salt). No buffer exchange is usually required.
Step 9.2.1: Pre‑filtration (Protection)
Purpose: Remove any aggregates, particles, or residual protein that could clog the viral filter.
Filter: 0.1 µm or 0.2 µm sterile filter, or a specific pre‑filter (e.g., Millipore Viresolve Prefilter, Sartorius Virosart Prefilter).
Why needed? Viral filters are expensive and have limited throughput. Aggregates >20 nm will block the membrane rapidly.
Step 9.2.2: Filter Assembly and Integrity Testing (Pre‑use)
Purpose: Verify that the viral filter is intact and has no defects before processing product.
Method: Diffusion test (gas flow) or pressure hold test according to manufacturer’s instructions. For Planova filters, a gold particle challenge is also used in validation, but for production, non‑destructive integrity tests are performed.
Acceptance criteria: Specific to filter type (e.g., maximum allowable gas diffusion rate at a given pressure).
Critical: A failed integrity test means the filter cannot be used – the batch would be at risk. Always have a spare filter available.
Step 9.2.3: Flush and Equilibration
Purpose: Remove storage solution (often 0.1 M NaOH or 20% ethanol) and equilibrate with product buffer.
Process:
Flush with WFI (water for injection) to remove storage solution.
Flush with product buffer (the same buffer as the AEX pool) until pH and conductivity match the feed.
Volume: Typically 10–20 L/m² of filter area.
Step 9.2.4: Viral Filtration (Product Processing)
Purpose: Pass the AEX pool through the viral filter and collect the filtrate.
Operating parameters:
| Parameter | Typical Range | Notes |
|---|---|---|
| Transmembrane pressure (TMP) | 0.5–2.0 bar (7–30 psi) | Lower TMP gives better throughput; high TMP causes fouling and potential virus breakthrough |
| Flow rate (flux) | 50–200 L/m²/h (LMH) | Start at lower flux to reduce fouling |
| Temperature | 15–25°C | Room temperature; cold increases viscosity and reduces flow |
| Pressure mode | Constant pressure or constant flow | Constant pressure is typical for dead‑end filtration |
| Load density | Up to 500–1,000 L/m² (depending on feed quality) | Lower if feed has aggregates |
Process:
Pump the pre‑filtered AEX pool through the viral filter at controlled pressure.
Collect the filtrate in a sterile, chilled hold tank (2–8°C).
Monitor pressure and flow rate continuously. When pressure approaches the maximum recommended (e.g., 2.5 bar), stop and either switch to a new filter or terminate the batch.
Typical throughput: For a clean mAb solution after AEX, throughput of 200–500 L/m² is common. Poorer feed (more aggregates) may reduce to 50–100 L/m².
Step 9.2.5: Post‑Use Integrity Testing
Purpose: Confirm that the filter remained intact throughout processing. This is a regulatory requirement.
Method: Same as pre‑use integrity test (diffusion or pressure hold). The test must be performed after product recovery (the filter may be washed with buffer to remove product before testing).
Acceptance criteria: Same as pre‑use. If the post‑use test fails, the batch may need to be investigated or rejected (depending on root cause).
Step 9.2.6: Filter Disposal
Viral filters are single‑use. After integrity testing, they are discarded as biohazardous waste (they have captured viruses from the batch).
Step 9.3: Critical Process Parameters (CPPs)
| Parameter | Typical Range | Impact on Virus Retention | Impact on Throughput |
|---|---|---|---|
| Transmembrane pressure (TMP) | 0.5–2.0 bar | High TMP (>2.5 bar) can force viruses through (risk) | Higher TMP increases initial flux but accelerates fouling |
| Flow rate (flux) | 50–200 LMH | High flux may reduce virus retention (size exclusion kinetics) | Higher flux increases throughput but may cause premature fouling |
| Temperature | 15–25°C | Low temperature increases viscosity, may reduce retention of small viruses | Lower temperature reduces flux |
| Feed quality (aggregates) | <0.5% aggregates by SEC | Aggregates >20 nm can block pores but do not affect virus retention directly | High aggregates drastically reduce throughput |
| pH | 6.0–8.0 | Extreme pH may alter virus size or membrane charge; within this range, minimal effect | Similar |
Worst‑case for validation: Highest TMP, highest flux, lowest temperature (highest viscosity), and highest aggregate level.
Step 9.4: Equipment and Scale Considerations
Filter holders:
Small scale (research): Stainless steel disc holders (e.g., Millipore XX44, Sartorius 16249).
Pilot scale (10–100 L): Single‑use capsules (e.g., Millipore Viresolve Pro, Sartorius Virosart).
Commercial scale (1,000–20,000 L): Multiple parallel capsules or larger format cartridges (e.g., Planova 20N cassettes, each up to 4 m²). A 20,000 L batch might require 20–40 m² of filter area.
Filtration skid:
Pump (low‑shear, e.g., peristaltic or diaphragm) with pressure control (PID loop).
Pressure sensors (inlet, outlet, and sometimes permeate side).
Flow meter.
Automated integrity tester (e.g., Millipore Integrity Tester, Sartorius Sartocheck).
Data logging to SCADA.
Facility: Viral filtration is typically performed in a classified cleanroom (Grade C or ISO 7) with the product in a closed system. The filter and downstream tank must be sterile.
Step 9.5: Viral Retention Validation
Before a viral filter can be used in manufacturing, it must be validated with a scaled‑down model to demonstrate ≥4 log reduction for relevant viruses, especially the smallest virus of concern (typically MVM, minute virus of mice, ~18–20 nm).
Validation study design (contract lab, e.g., WuXi, Charles River, Merck):
Select model viruses:
MVM (non‑enveloped, 18–20 nm) – worst‑case due to size.
X‑MuLV (enveloped, 80–120 nm) – easily retained, but required.
Reovirus (non‑enveloped, 60–80 nm) – optional.
Scale‑down filter: Use the same filter membrane and same manufacturing conditions (same flux, TMP, temperature, buffer) on a small disc (e.g., 25 mm diameter).
Spike the virus into the AEX pool (or a surrogate pool with similar composition).
Filter at worst‑case conditions (highest TMP, highest flux, lowest temperature, and at the end of the run when filter is partially fouled – this is critical).
Collect filtrate and measure infectivity (plaque assay or TCID₅₀).
Calculate LRV = log₁₀(virus input) – log₁₀(virus in filtrate).
Acceptance criteria:
≥4 log for MVM (smallest virus)
≥4 log for X‑MuLV (typically >6 log)
The filter manufacturer often provides a “virus retention claim” based on their validation, but you must still perform process‑specific validation with your product and buffer.
Important: The validation must include fouled filter conditions (i.e., near the end of the batch) because fouling can affect virus retention. Some viruses may be pushed through if the membrane is stressed.
Step 9.6: Impact on Product Quality
Viral filtration is generally gentle. However, potential issues:
| Issue | Cause | Mitigation |
|---|---|---|
| Low mAb recovery | Non‑specific binding to membrane; filter saturation | Use low‑binding membrane (e.g., Planova, Viresolve Pro); pre‑condition with buffer or placebo |
| Aggregate increase | Shear or air/liquid interface during filtration | Use low‑shear pump; avoid air entrainment |
| Slow filtration (low flux) | Aggregates or particles blocking pores | Improve pre‑filtration (0.1 µm); optimize feed quality (reduce aggregates) |
| Membrane rupture | Excessive TMP (>3 bar) | Set pressure alarm at 2.5 bar; use pressure relief |
Recovery target: >95% mAb recovery through viral filtration (pre‑filter + viral filter combined).
Step 9.7: What Can Go Wrong? (Troubleshooting)
| Problem | Possible Cause | Remedy |
|---|---|---|
| Pre‑use integrity test fails | Filter damaged during handling; improper installation | Discard filter; use a new one; check installation procedure |
| Rapid pressure rise (low throughput) | High aggregate level; feed not pre‑filtered adequately | Reduce load aggregate (optimize CEX/AEX); add 0.1 µm pre‑filter; lower TMP |
| Post‑use integrity test fails | Filter ruptured during processing; TMP too high | Investigate batch – may need to reject or re‑filter; increase pressure monitoring |
| Low virus retention in validation | Worst‑case conditions too aggressive; filter not validated for that virus | Re‑validate with lower TMP; choose a different filter (e.g., Planova 15N for smaller viruses) |
| Low mAb recovery (<90%) | Membrane binding; product precipitation on filter | Pre‑condition filter with buffer containing polysorbate; check product stability |
| Batch takes too long (hours) | Flux too low; filter area insufficient | Increase filter area (parallel filters); operate at higher TMP (within limits) |
Step 9.8: Integration with Downstream Steps
The viral filtrate is now virally safe (validated clearance of ≥4 log for relevant viruses). It is typically called Viral Filtered Pool.
Hold conditions: 2–8°C for up to 24–48 hours before final concentration and buffer exchange (UF/DF). The next step is Step 10: Ultrafiltration/Diafiltration (UF/DF) , where the mAb is concentrated to its final drug substance concentration (e.g., 50–100 mg/mL) and formulated into the final buffer (e.g., citrate, polysorbate, trehalose).
Step 9.9: Alternatives and Emerging Approaches
Smaller pore filters (15 nm, e.g., Planova 15N): Used for products where parvovirus clearance is critical and the mAb can pass through 15 nm pores. Not necessary for most IgG1 mAbs (20 nm is sufficient).
Two‑stage viral filtration: Using a 50 nm pre‑filter to remove larger aggregates, then a 20 nm final filter. Improves throughput but adds cost.
Continuous viral filtration: In a continuous manufacturing train, viral filters are used in parallel with automatic switching when pressure rises. Still emerging.
For a standard mAb like Humira, a single 20 nm viral filter (Planova or Viresolve) with appropriate pre‑filtration is the proven, regulatory‑accepted approach.
What Comes Next?
The viral filtered pool is dilute (typically 2–10 g/L mAb) and still in the AEX buffer. The final step before bulk drug substance is Step 10: Ultrafiltration/Diafiltration (UF/DF) – concentrating the mAb to its final formulation strength (e.g., 50–100 mg/mL) and exchanging the buffer into the stable, injectable formulation.
Questions for our readers:
Which viral filter do you use for your mAb process (Planova, Viresolve, Virosart)?
Have you experienced low throughput due to aggregates? How did you solve it?
Share your experiences below. Next week: Step 10 – UF/DF for concentration and formulation.
The text above is owned by the site above referred.
Here is only a small part of the article, for more please follow the link
Also see:
Manostaxx – Industrial Management Consulting
