Fixed-bed bioreactors have become the dominant platform for manufacturing viral vectors from adherent cells at clinical and commercial scale. By packing structured carriers into a perfused bed, these systems deliver surface areas of 66-500 m2 in a single vessel, replacing hundreds of multi-layer cell stacks and the manual labour that comes with them. For gene therapy manufacturers scaling adherent HEK293 or Vero cell processes, choosing between the Pall/Cytiva iCELLis, the Univercells Scale-X, and microcarrier-based stirred tanks is one of the most consequential platform decisions in process development.
This guide compares the two leading fixed-bed bioreactor platforms, explains scale-up methodology, reviews published viral vector yield data, and provides a decision framework for choosing between fixed-bed, microcarrier, and suspension-adapted approaches.
How Fixed-Bed Bioreactors Work
A fixed-bed bioreactor is a single-use vessel containing a packed bed of structured macrocarriers through which culture medium is continuously perfused. Unlike stirred-tank reactors where cells grow on suspended microcarrier beads, the carriers in a fixed-bed remain stationary while medium circulates through the bed, creating a low-shear, high-surface-area environment for adherent cell growth.
The operating principle consists of three elements working together:
- Packed carrier bed: Woven polyester (PET) or polypropylene structured carriers are compressed into a cylindrical bed. The fibrous matrix provides 100-500 m2/m3 of specific surface area. Cells attach to the carrier fibres and grow within the three-dimensional structure.
- Medium recirculation: A peristaltic or centrifugal pump drives medium from a reservoir through the bed and back. The reservoir sits below the bed (in most designs), allowing gravity-assisted drainage during harvest. Perfusion rates typically range from 1-5 vessel volumes per day (vvd).
- Gas exchange: Oxygen transfer occurs primarily at the medium surface in the reservoir (headspace aeration) or through a sparger in the reservoir. The recirculating medium carries dissolved oxygen into the bed. CO2 stripping also occurs in the reservoir headspace.
The key advantage of this geometry is that oxygen and nutrient gradients exist only across the bed height, not across the vessel diameter. This means scale-up by increasing the vessel cross-section preserves the same gradient profile that cells experience at bench scale. Every cell in a 500 m2 bed sees the same environment as in a 0.53 m2 bench unit, provided bed height and superficial velocity are held constant.
The iCELLis Platform: Nano and 500+
The iCELLis (originally Pall, now Cytiva) is the most widely adopted fixed-bed bioreactor for viral vector and vaccine manufacturing. Two models cover the full development-to-manufacturing range: the iCELLis Nano for process development and the iCELLis 500+ for clinical and commercial production.
iCELLis Nano
The Nano is a benchtop-scale fixed-bed bioreactor with interchangeable compacted beds offering 0.53, 1.07, 2.67 or 4.0 m2 of surface area. Two bed heights are available: 2 cm (low-density) and 10 cm (high-density). The Nano uses a dedicated controller and sits on a standard bench, making it the process development workhorse for adherent cell processes.
iCELLis 500+
The 500+ scales the same carrier and bed-height geometry to production volumes. Available surface areas are 66, 133, 200, 333 and 500 m2. Like the Nano, both 2 cm and 10 cm bed heights are offered. The 500+ vessel connects to a standard bioreactor controller (DeltaV, MFCS, etc.) for GMP-compliant operation with electronic batch recording.
| Parameter | Nano (2 cm bed) | Nano (10 cm bed) | 500+ (2 cm bed) | 500+ (10 cm bed) |
|---|---|---|---|---|
| Surface area (m2) | 0.53 - 4.0 | 0.53 - 4.0 | 66 - 500 | 66 - 500 |
| Carrier material | Woven PET macrocarrier (Fibra-Cel-derived) | |||
| Bed height | 2 cm | 10 cm | 2 cm | 10 cm |
| Medium volume (L) | 0.2 - 1.5 | 0.5 - 2.5 | 25 - 180 | 50 - 350 |
| Perfusion rate (vvd) | 1 - 5 (typical: 2 - 3) | |||
| Cell density achievable | 3 - 30 x 106 cells/cm2 (carrier-dependent) | |||
| Online sensors | pH, DO, temp in reservoir | pH, DO, temp, optional biomass | ||
| Primary use | Process development | Clinical / commercial GMP | ||
A critical consideration is the choice between 2 cm and 10 cm bed heights. The 2 cm bed provides uniform nutrient and oxygen delivery with lower risk of gradients, making it suitable for oxygen-demanding processes like high-titre lentivirus production. The 10 cm bed packs 5x more surface area into the same footprint but creates steeper dissolved-oxygen gradients from top to bottom of the bed. For HEK293 cells producing AAV (moderate oxygen demand), the 10 cm bed is common. For Vero cells in vaccine production, 2 cm beds are often preferred.
Scale-X: Univercells Modular Approach
The Scale-X platform (Univercells Technologies) is a newer entrant that takes a different engineering approach to fixed-bed bioreactor design. Instead of a single large packed bed, Scale-X uses structured carriers in a bioreactor with integrated modules, allowing a modular scale-up path.
The Scale-X family spans four models: the nexo (0.5 m2) for screening, the hydro (2.4 m2) for process development, the carbo (10-30 m2) for clinical manufacturing, and the nitro (200-600 m2) for commercial production. All share the same structured carrier and bed geometry, enabling linear scale-up from bench to GMP.
Scale-X hydro
The hydro is the process development workhorse, offering 2.4 m2 of growth surface in a 2,700 mL working volume. It uses structured carriers in a fixed bed with controlled perfusion and is fully instrumented with pH, DO, and temperature probes. The hydro connects to standard automation platforms and is the unit most frequently compared to the iCELLis Nano in published benchmarking studies.
Scale-X nitro
The nitro is the commercial manufacturing unit, providing 200-600 m2 of surface area. It maintains the same carrier type and bed geometry as the hydro, enabling direct process transfer. The nitro integrates with Univercells' NevoLine Upstream platform for automated GMP manufacturing within a 3 m2 cleanroom footprint.
| Parameter | Scale-X nexo | Scale-X hydro | Scale-X carbo | Scale-X nitro |
|---|---|---|---|---|
| Surface area (m2) | 0.5 | 2.4 | 10 - 30 | 200 - 600 |
| Carrier material | Structured polypropylene / polyester fixed bed | |||
| Working volume (mL) | 75 | 2,700 | 1,800 - 4,200 | TBD |
| Online sensors | pH, DO, temp | pH, DO, temp | pH, DO, temp | Full suite |
| Primary use | Screening | PD | Clinical | Commercial GMP |
In a head-to-head benchmarking study, Leinonen et al. (2020) found that the Scale-X hydro produced approximately twice the lentiviral titre (9.8 x 105 TU/cm2) compared to the iCELLis Nano (4.7 x 105 TU/cm2) despite having only 60% of the surface area. For adenovirus, the hydro achieved 1.11 x 1011 VP/mL versus 8.53 x 1010 VP/mL in the iCELLis Nano. The platform also integrates with the NevoLine automated manufacturing system, which Univercells claims can reduce drug substance COG by 18-61% depending on the product.
Corning Ascent FBR is a newer fixed-bed platform (launched 2024) using vertically stacked woven PET mesh discs. It offers surface areas from 1 to 100 m2 (scalable to 1,000 m2) and claims viable cell harvest capability, which addresses one of the main limitations of the iCELLis and Scale-X platforms. The Ascent separates the media conditioning vessel from the bioreactor and measures dissolved oxygen at both bed inlet and outlet, enabling growth modelling without cell sampling.
Platform Comparison: iCELLis vs Scale-X vs Microcarrier STR
Choosing between fixed-bed and microcarrier stirred-tank platforms requires balancing surface area per footprint, scale-up predictability, harvest simplicity, and monitoring capability. The radar chart below compares the three main approaches across six critical engineering dimensions.
| Attribute | Fixed-bed (iCELLis / Scale-X) | Microcarrier STR | Suspension-adapted STR |
|---|---|---|---|
| Cell lines supported | Adherent only (HEK293, Vero, MRC-5, MDCK) | Adherent (bead-to-bead transfer required) | Suspension-adapted HEK293, CHO, HEK293SF |
| Max surface area per vessel | 500-600 m2 | ~15-40 m2 (200 L STR at 3-10 g/L beads) | N/A (cells in suspension) |
| Shear stress | Very low (no impeller in the bed) | Moderate (impeller + sparger) | Moderate (impeller + sparger) |
| Harvest method | Lysis buffer perfused through bed | Enzymatic detachment + bead separation | Cell lysis in vessel |
| Online monitoring | Reservoir only (pH, DO, temp) | Direct (pH, DO, temp in bulk) | Direct (pH, DO, temp in bulk) |
| Scalability limit | 500-600 m2 per vessel | Proven to 2,000 L STR | Proven to 2,000 L STR |
| Transfection compatibility | PEI, lipofection (in-situ) | PEI (in-situ on beads) | PEI, electroporation |
| Regulatory track record | Multiple Phase I-III programmes | Established (vaccines) | Growing (gene therapy) |
| Labour intensity | Low (closed system) | Medium (bead transfer steps) | Low (standard STR ops) |
How to Scale Up a Fixed-Bed Bioreactor Process
Fixed-bed bioreactors scale by maintaining bed height constant and increasing the cross-sectional area of the vessel. This is fundamentally different from stirred-tank scale-up (which relies on matching P/V, tip speed, or kLa) and is one of the key advantages of the platform.
Three parameters must be conserved during scale-up:
- Superficial velocity (cm/min) through the bed, which determines nutrient and oxygen delivery rate per unit carrier surface.
- Dissolved oxygen at the bed outlet, ensuring that cells at the bottom of the bed (furthest from fresh medium) receive adequate oxygen. The DO gradient across the bed must remain within the same range as at bench scale.
- Volumetric perfusion rate (vessel volumes per day, vvd), which maintains the same medium exchange rate per cell.
Worked Example: Scaling from iCELLis Nano to 500+
Process development scale: iCELLis Nano, 10 cm bed, 4.0 m2 surface area
Optimised parameters:
- Seeding density: 8,000 cells/cm2
- Perfusion rate: 2.5 vvd
- Medium volume: 2.0 L
- Feed: DMEM/F12 + 5% FBS
- Transfection: PEI at 1 µg/cm2, DNA at 0.5 µg/cm2
- Harvest: 72 h post-transfection, 3 freeze-thaw cycles + Benzonase
Target production scale: iCELLis 500+, 10 cm bed, 333 m2
Scale factor: 333 / 4.0 = 83x increase in surface area
Scaled parameters:
- Seeding: 8,000 cells/cm2 x 333 x 104 cm2 = 2.66 x 1010 cells required
- Medium volume: scales proportionally with surface area, ~166 L
- Perfusion rate: 2.5 vvd (unchanged)
- PEI: 1 µg/cm2 x 3.33 x 106 cm2 = 3.33 g PEI
- Plasmid DNA: 0.5 µg/cm2 x 3.33 x 106 cm2 = 1.67 g DNA
- Bed height: 10 cm (unchanged)
Key check: if Nano DO at bed outlet was 40% air saturation at 2.5 vvd, verify the 500+ achieves the same. If not, increase perfusion rate incrementally.
The simplicity of this scale-up approach is why fixed-bed bioreactors accelerate process development timelines. A process optimised at the Nano scale (weeks of development) transfers to the 500+ with typically only 1-2 engineering runs to confirm the DO profile and perfusion rate. Compare this with microcarrier scale-up, which requires establishing new bead-to-bead transfer protocols and re-optimising impeller speed, sparger design, and surface-area-to-volume ratios at each scale.
Viral Vector Yields in Fixed-Bed Bioreactors
Fixed-bed bioreactors have demonstrated competitive yields for AAV, lentiviral, and adenoviral vectors. Published data show that fixed-bed platforms match or exceed cell-stack yields on a per-surface-area basis, with the added advantage of closed-system operation and reduced manual handling.
| Vector type | Platform | Cell line | Yield | Reference |
|---|---|---|---|---|
| AAV | iCELLis Nano | HEK293 | 2.2 x 1010 vg/cm2 | Lesch et al., 2015 |
| AAV | iCELLis Nano | A549 packaging | 4.5 x 108 vg/cm2 | Lesch et al., 2015 |
| Adenovirus (Ad5) | iCELLis 500 (100 m2) | HEK293 | 1.2 x 1016 VP/batch | Lesch et al., 2015 |
| Adenovirus | Scale-X hydro | HEK293 | 1.11 x 1011 VP/mL | Leinonen et al., 2020 |
| Adenovirus | iCELLis Nano | HEK293 | 8.53 x 1010 VP/mL | Leinonen et al., 2020 |
| Lentivirus | Scale-X hydro | HEK293T | 9.8 x 105 TU/cm2 | Leinonen et al., 2020 |
| Lentivirus | iCELLis Nano | HEK293T | 4.7 x 105 TU/cm2 | Leinonen et al., 2020 |
| Lentivirus | Fixed-bed vs Cell Factory | HEK293T | 3.2-7.3x higher titre | Valkama et al., 2018 |
The iCELLis fixed-bed bioreactor has a proven GMP regulatory track record: Zolgensma (onasemnogene abeparvovec, Novartis Gene Therapies), an FDA-approved AAV9 gene therapy for spinal muscular atrophy, is manufactured using adherent HEK293 cells in the iCELLis system. This makes the iCELLis one of a small number of bioreactor platforms with an approved commercial gene therapy product.
Two points are worth noting about the yield data. First, per-surface-area yields in fixed-bed bioreactors are generally comparable to or slightly below optimised cell stacks for the same cell line and process (iCELLis yielded ~47% of the Cell Factory control for AAV in the Lesch et al. 2015 study), meaning the primary advantage of fixed-bed is scale, closed-system operation, and labour reduction rather than inherently higher productivity per cm2. Second, the transition to stable producer cell lines for AAV (eliminating the need for transient transfection) can further boost yields in fixed-bed systems because every cell in the bed contributes to vector production without the transfection efficiency ceiling.
Process Transfer from Cell Stacks to Fixed-Bed
Transferring an existing adherent cell process from multi-layer cell stacks (e.g., Corning CellSTACK, Nunc Cell Factory) to a fixed-bed bioreactor requires matching three critical parameters: seeding density per unit surface area, perfusion rate per unit surface area, and reagent doses per unit surface area. The surface area becomes the common denominator for all scaling calculations.
The transfer workflow follows five steps:
- Calculate the equivalent surface area. For example, 40 x 10-layer CellSTACKs (each 6,320 cm2) = 253,000 cm2 = 25.3 m2. This maps to an iCELLis 500+ with 66 m2 (2 cm bed) or a Scale-X Nitro for development.
- Seed at the same cells/cm2. If cell stacks are seeded at 15,000 cells/cm2, use the same density in the fixed bed.
- Match the medium exchange rate. If cell stacks receive a full medium change every 48 hours, set perfusion to deliver the same volume per cm2 per day.
- Scale transfection reagents per cm2. PEI, DNA, and lipofectamine doses that were optimised per cell stack surface area transfer directly to fixed-bed surface area.
- Adjust harvest for the bed. Instead of aspirating supernatant and scraping cells, lysis buffer (Triton X-100 or freeze-thaw) is perfused through the bed. The bed retains cell debris while product flows through, simplifying clarification.
If your cell-stack process has not been optimised using a design of experiments approach, the transfer to fixed-bed is an opportunity to systematically screen key parameters (seeding density, PEI:DNA ratio, MOI for stable lines, harvest time, and perfusion rate) rather than transferring sub-optimal conditions. A 25-1 fractional factorial screen in the iCELLis Nano requires only 16 runs and typically identifies 2-3x yield improvements.
Cost Analysis: Cell Stacks vs Fixed-Bed vs Suspension
The economic case for fixed-bed bioreactors rests on two factors: dramatically reduced labour costs (from eliminating manual cell-stack manipulation) and smaller cleanroom footprint. Published cost models for adherent cell viral vector manufacturing show 30-60% reduction in cost of goods when transitioning from cell stacks to fixed-bed systems at equivalent scale.
| Cost category | Cell stacks (40 x 10-layer) | iCELLis 500+ (200 m2) | Suspension STR (200 L) |
|---|---|---|---|
| Consumables (vessel, media, reagents) | $35,000 - 50,000 | $25,000 - 40,000 | $15,000 - 25,000 |
| Direct labour (hours per batch) | 80 - 120 h | 20 - 30 h | 15 - 25 h |
| Cleanroom footprint | Grade B, 40-60 m2 | Grade C/D, 15-25 m2 | Grade C/D, 10-20 m2 |
| Contamination risk | High (many open manipulations) | Low (closed system) | Low (closed system) |
| Batch failure rate (est.) | 5-10% | 1-3% | 1-3% |
| Cell line requirement | Adherent (native) | Adherent (native) | Suspension-adapted |
The suspension STR option has the lowest consumable cost because it avoids the single-use fixed-bed cartridge (the most expensive single component in the iCELLis system). However, suspension adaptation of HEK293 cells requires 3-6 months of cell line development and may alter cell phenotype, transfection efficiency, or vector quality attributes. For programmes already committed to adherent HEK293 or Vero platforms, fixed-bed represents the best combination of cost reduction and process continuity.
For detailed economic modelling of upstream options including media cost, facility utilisation, and batch scheduling, use the Fermentation Economics Calculator.
Frequently Asked Questions
What is a fixed-bed bioreactor?
A fixed-bed bioreactor is a single-use vessel containing a packed bed of structured carriers (typically woven PET or BioNOC II fibre discs) through which culture medium is continuously perfused. Adherent cells attach and grow on the carrier surfaces, achieving surface-area-to-volume ratios of 100-500 m2 per vessel. Unlike microcarrier systems, the carriers remain stationary while medium flows through the bed.
What is the difference between iCELLis and Scale-X bioreactors?
The iCELLis (Cytiva/Pall) uses woven PET macrocarriers with 2 cm or 10 cm bed heights, offering 0.53-500 m2 across the Nano and 500+ models. Scale-X (Univercells) uses structured carriers in a modular vessel design, with the Nitro at 2.4 m2 and Carbo up to 600 m2. iCELLis has a larger installed base and more published data; Scale-X offers a higher maximum surface area and modular architecture.
How do you scale up a fixed-bed bioreactor process?
Fixed-bed bioreactors scale by keeping bed height constant and increasing the cross-sectional area. Conserve three parameters: superficial velocity (cm/min), dissolved oxygen at the bed outlet, and volumetric perfusion rate (vvd). A process optimised on the iCELLis Nano at 4 m2 transfers directly to the 500+ at 200-500 m2 with typically 1-2 confirmation runs.
Is a fixed-bed bioreactor better than microcarriers for viral vector production?
For adherent cell viral vector production at scales up to 500 m2, fixed-bed bioreactors offer advantages: 5-10x higher volumetric cell density than cell stacks, no bead-to-bead transfer, lower shear, built-in perfusion, and simpler harvest. Microcarrier STR provides better real-time monitoring and proven scalability beyond 200 L, but requires more complex harvest protocols and bead-to-bead transfer for seed train expansion.
What AAV titers can be achieved in fixed-bed bioreactors?
Published AAV titers in fixed-bed bioreactors range from 1010 to 1011 vg/mL in crude harvest lysate, corresponding to 1013-1015 total vg per run depending on scale and serotype. Per-surface-area yields of 1-5 x 1010 vg/cm2 have been reported for AAV5 in the iCELLis Nano. Yields improve further with stable producer cell lines.
Scale-Up Calculator
Compare scale-up criteria (P/V, tip speed, kLa, Reynolds, mixing time) for STR-based processes or verify your fixed-bed superficial velocity calculations.
DOE Generator
Design a screening experiment to optimise seeding density, PEI:DNA ratio, MOI, perfusion rate, and harvest time in your fixed-bed process.
Seed Train Planner
Plan the inoculum expansion from cell bank to fixed-bed seeding. Calculate the number of T-flasks and cell stacks needed to hit your target cells/cm2.
Related Tools
- Media Estimator — Calculate media and supplement volumes for fixed-bed perfusion culture
- Fermentation Economics Calculator — Model COGS for upstream manufacturing including vessel, media, and labour costs
- Cell Therapy Planner — Plan adherent cell expansion for gene and cell therapy manufacturing
References
- Lesch HP, et al. (2015). Process development of adenoviral vector production in fixed bed bioreactor: from bench to commercial scale. Human Gene Therapy, 26(8), 560-571. doi:10.1089/hum.2015.081
- Valkama AJ, et al. (2018). Optimization of lentiviral vector production for scale-up in fixed-bed bioreactor. Gene Therapy, 25(1), 39-46. doi:10.1038/gt.2017.91
- Leinonen HM, et al. (2020). Benchmarking of Scale-X bioreactor system in lentiviral and adenoviral vector production. Human Gene Therapy, 31(5-6), 376-384. doi:10.1089/hum.2019.247
- Goral MI, et al. (2024). Innovative fixed bed bioreactor platform for scalable adherent cell bioprocessing. Biotechnology Journal, 19(4), e2300635. doi:10.1002/biot.202300635