Shop in your local currency and language

Choose another country or region to see specific content for your location and shop online.

United Kingdom
Viral vectors

Optimizing lentiviral vector clarification using membrane and depth filtration strategies

Oct 29, 2026

Clarification is a critical unit operation in lentiviral vector (LVV) manufacturing that directly impacts downstream processing performance, product recovery, and process robustness.

In this study, we evaluated depth and membrane filtration technologies for the clarification of LVV harvests generated from a suspension of HEK293SF PacLV cultures. Using a range of depth filters and membranes, we assessed filtration performance based on throughput, turbidity reduction, and infectious LVV recovery.

Primary clarification studies demonstrated:

  • Coarse, high capacity prefilters efficiently removed larger particulates while maintaining high LVV recovery.
  • Seitz™ V100P depth filter improved turbidity reduction and produced a cleaner feed stream for downstream processing.
  • Secondary filtration performance was dependent on primary filtrate quality.
  • Supor™ EAV membrane consistently provided the optimal balance of LVV recovery and turbidity reduction.

We found the most robust performance across all evaluated criteria to be a two-stage clarification strategy consisting of Seitz V100P depth filtration followed by Supor EAV membrane filtration.

Introduction

Lentiviral vectors (LVVs) are widely used in cell and gene therapy due to their ability to stably deliver genetic material into both dividing and non-dividing cells. These vectors are commonly applied in ex vivo therapies such as CAR T-cell and stem cell-based treatments, where consistent vector quality and yield are critical for clinical success (1). As demand for LVV-based therapies increases, there is a growing need to develop robust and scalable manufacturing processes. Downstream processing remains a major challenge, particularly during harvest clarification. This step is required to remove cells, cell debris, and process-related impurities prior to further purification, and it can significantly influence process yield and robustness (1).

Clarification of LVV harvests is challenging due to both the properties of the vector and the composition of the feed. LVVs are relatively large, enveloped particles that are sensitive to processing conditions and can be retained or inactivated during filtration (2). In addition, suspension cell cultures produce harvests with high turbidity and variable impurity loads, which can impact filtration performance and downstream operation stability (1).

Filtration is widely used for clarification because it provides a scalable and straightforward approach to solids removal. Depth filters and membrane filters are often combined to achieve sufficient removal of impurities while maintaining acceptable product recovery. However, filter selection plays a critical role in determining process performance. Tighter filtration media can improve filtrate clarity but may increase vector retention, while more open structures can facilitate higher throughput at the expense of turbidity reduction (1,2). Development of an effective clarification strategy therefore requires evaluation of filtration performance across multiple parameters, including throughput, turbidity reduction, and LVV recovery. These parameters are interdependent and must be balanced to support downstream processing and overall process efficiency.

This study evaluates several filtration technologies for LVV clarification using suspension-derived harvest material. The objective is to compare filter performance and identify configurations that provide an optimal balance between the aforementioned critical quality attributes. The target filter train is two steps and includes appropriate clarification and bioburden filtration unit operations. The final harvest clarification quality is UF/DF ready, this is typically though a 0.45 um filter or with a turbidity less than 10 NTU.

Materials and methods

Table 1. Process and product characteristics

Product LVV harvest
Packaging cell line HEK293SF PacLV cells produced in a BioFlow system
Live cell density (cells/mL) 2.24 × 106
Media HyClone™ HyCell™ CHO cell culture medium and HyClone peak expression medium
Viability (%): Harvest 93
Aggregation (%) 23-24
Titer (TU/mL): Harvest 1.37 × 107
Feed turbidity (NTU) 153

Table 2. List of filters tested during screening

Filter Material of construction Retention rating (µm) Area tested (cm2)
Fluorodyne™ II DBL (Bioburden reduction membrane) Polyvinylidene fluoride (PVDF) 0.45 (absolute) 11.1
Supor EAV (bioburden reduction membrane) Polyethersulfone (PES) 0.2 (nominal) 2.8
HDC II J060 (polypropylene ‘tapered pore’ prefilter) Pleated polypropylene 6.0 11.1
Seitz V100P (depth filter) Cellulose, perlite,
and diatomaceous earth
2.0–4.0 22.0

Filter pre-conditioning

After rinsing with water, the filters were equilibrated with cell media at 1.5 hold-up volumes.

Clarification

To avoid dilution, the first 70% of the hold-up volume was discarded to waste, we then began to collect filtrate. Throughout, we recorded filtration run time, inlet pressure, filtrate volume/mass collected, and turbidity (instantaneous and pool) at set intervals, monitoring closely for breakthrough. For gene therapy and viral applications, the end-point of clarification is critical. Excessive plugging is to be avoided. The end points for this study were a terminal differential pressure of 10 psid or when sufficient data was generated to establish capacity.

Buffer chase

In any depth filtration process displacing the feed stock can increase product yield, however for gene therapy product, yield can be significantly increased with a post-filtration buffer rinse step. Therefore, after filtration:

  • Filter lines were switched into cell culture media and the harvest was displaced into the product pool.
  • The filter and lines were further rinsed with 1.5 hold-up volumes of cell media at the filtration flow rate and collected in aliquots.

Performing bench-scale membrane filterability trials

Filtrates from the primary filters were used as feed material for the secondary (bioburden reduction) filter tests. No recovery chases were performed for these tests as the lines were pumped dry and the volume before and after the test was assumed to be the same.

Results

Results from the primary clarification of LVV harvest are shown in Figure 1. Previous screening (data not shown) of 10 µm and 4.5 µm HDC II series filters identified the J060 (6 µm) filter as having the optimal combination of throughput and turbidity reduction for a similar LVV harvest. This screening also identified the V100P (2 to 4 µm) depth filter as an effective primary filter so both these filters were tested in duplicate. The V100P (2 to 4 µm) reduced turbidity by over 98.5% with recovery at 80% and 90% for the two filters tested. The J060 had good turbidity reduction at ~ 94% and good recovery at 81% and 82% for the two filters tested. Overall, the V100P filter improved turbidity reduction and recovery but had lower capacity.

Primary clarification of LVV harvest comparing pressure drop and throughput for Seitz V100P and HDC II J060 filters. Filtration stopped near 10 psid or when the product pool was depleted

Fig 1. Primary clarification of LVV harvest. Filtration was stopped when pressure approached 10 psid or the product pool was depleted.

The filtrate produced by the HDC II J060 and Seitz V100P filters was evaluated through secondary filters shown in Figure 2. As indicated by turbidity reduction the V100P filtrate did not foul any filters tested while the J060 filtrate led to fouling of the Fluorodyne II DBL filter with a throughput of 200 L/m2 at 8 psid. The cleaner V100P filtrate had higher recoveries after secondary filtration for both DBL and EAV filters (see Table 3).

Secondary filtration of clarified LVV harvest comparing pressure drop and throughput for DBL after J060, DBL after V100P, and EAV after V100P

Fig 2. Secondary filtration of clarified LVV harvest.

Table 3 summarizes all the data gathered in this experiment. The top candidates for primary clarification are V100P and J060. Of these, the J060 has the highest capacity while the V100P filter has the most turbidity reduction and slightly improved recovery. The V100P filtrate quality may be good enough so that a secondary filter may not be required prior to tangential flow filtration (TFF). The top filter for secondary filtration is the EAV as it had good turbidity reduction and the highest recoveries.

Table 3. Summary of filtration results

Feed material Filter train Volume
(mL)
Flux
(LMH)
Throughput
(L/m 2)
Max. pressure drop
(ΔP)
Pool turbidity
(NTU)
Step yield
(%)
Harvest V100P 453 155 197.3 9.4 2.1 80
Harvest V100P 502.4 148.9 228.4 10.0 1.9 90
Harvest J060 327.0 315.6 294.6 10.2 9.1 82
Harvest J060 56.0 324.5 313.2 9.2 8.7 81
J060 filtrate DBL 223.5 219.7 201.4 7.8 3.3 53
V100P filtrate DBL 108.8 196.0 98.0 0.2 1.8 77
V100P filtrate EAV 129.0 412.6 > 460.7 0.6 2.0 92

Discussion

Our study evaluated multiple filtration strategies for LVV clarification and we demonstrated that filter selection significantly influences throughput, turbidity reduction, and LVV recovery. Findings that are consistent with previous reports highlighting that an insufficient filtration strategy for clarification is a key factor in reduction of yield and product quality for LVV downstream processing (1).

Primary clarification results showed clear differences between filter types. HDC II J-series filters provided the highest throughput and maintained LVV recovery close to quantitative levels. This performance is consistent with more open filtration structures that minimize product retention while removing larger particulates. However, such configurations often result in higher residual turbidity, which can impact downstream unit operations (2).

In contrast, we found that the Seitz V100P depth filter achieved substantially lower filtrate turbidity across both study rounds. This increased clarity led to reduced capacity due to fouling. However, we observed that cleaner filtrate from the V100P resulted in 25% better LVV recovery across the DBL secondary filter.

The impact of primary clarification on secondary filtration performance was evident. Filtrates generated using Seitz V100P resulted in more consistent downstream filtration behavior, indicating that improved removal of fine particles reduces membrane fouling and enhances process robustness.

Among the secondary filters evaluated, the Supor EAV membrane provided the highest LVV recovery while maintaining low turbidity. In contrast, the Fluorodyne II DBL membrane showed lower and more variable recovery, suggesting increased interaction between the LVV particles and the membrane surface. Such interactions have been previously reported for membrane-based separations and can contribute to product loss during downstream processing (1).

Filtrates generated from HDC II J-series filters showed increased fouling during secondary filtration, particularly when processed with Fluorodyne II DBL membrane. This resulted in reduced throughput and recovery, demonstrating that insufficient primary clarification can negatively affect subsequent unit operations. These findings highlight the importance of aligning primary and secondary filtration steps to minimize process inefficiencies.

Taken together, these results indicate that clarification should be considered as a multi-stage process rather than a single operation. High capacity prefilters enable efficient bulk solids removal, while depth filters and membrane filters are required to achieve the level of clarity necessary for robust downstream processing.

The combination of a Seitz V100P depth filter followed by Supor EAV membrane provided the most consistent performance. This configuration enabled effective turbidity reduction while maintaining high LVV recovery and acceptable throughput. However, the HDC II J060 filter was also an effective primary filter.

Conclusion

This study demonstrates that LVV clarification performance depends on both filter selection and process configuration. Primary filtration using open structures supports high throughput but may not provide sufficient turbidity reduction for downstream processing, whereas depth filtration improves filtrate quality at the expense of capacity.

We found that a two-stage clarification strategy combining Seitz V100P depth filter with Supor EAV membrane provides a strong balance of throughput, turbidity reduction, and LVV recovery. This approach improves downstream processability, reduces fouling, and minimizes product loss.

The importance of integrated filter selection during process development is evident in this study, showing that a practical framework for optimizing clarification strategies can enhance impurity removal while maintaining high LVV recovery, supporting scalable and efficient LVV manufacturing.

References
  • Moreira AS, Cavaco DG, Faria TQ, Alves PM, Carrondo MJT, Peixoto C. Advances in Lentivirus Purification. Biotechnol J. 2021 Jan;16(1):e2000019. https://doi.org/10.1002/biot.202000019
  • Tona RM, Shah R, Middaugh K, Steve J, Marques J, Roszell BR, Jung C. Process intensification for lentiviral vector manufacturing using tangential flow depth filtration. Mol Ther Methods Clin Dev. 2023 Mar 3;29:93-107. https://doi.org/10.1016/j.omtm.2023.02.017

CY61591

Develop and refine your lentivirus production
Discover end-to-end support with scalable single-use solutions for cell and gene therapy workflows

Did you find what you were looking for?