What is Dr. Zubair Khalid's research focus?

Dr. Zubair Khalid specializes in molecular virology, mRNA vaccine development, and computational biology, with a focus on avian pathogens like IBDV and Avian Reovirus.

Where is Dr. Zubair Khalid currently working?

Dr. Zubair Khalid is a Postdoctoral Research Associate at the University of Maryland (UMD), specifically within the Department of Animal and Avian Sciences.

Upstream Bioprocessing: A Practical Guide for Industry Scientists

By Dr. Zubair Khalid, DVM, MS, PhD · 2026-08-11

Introduction to Upstream Bioprocessing

Definition and Scope

Upstream bioprocessing encompasses all unit operations involved in preparing, growing, and maintaining cells or microorganisms for the purpose of producing a biological product. In biopharmaceutical manufacturing, this begins with the thaw of a working cell bank vial and ends with the harvest of the culture broth containing the product of interest—typically a recombinant protein, monoclonal antibody, virus, or plasmid. The upstream process establishes the foundation upon which all subsequent purification steps depend; product titer, quality attributes, and impurity profiles are largely determined by decisions made in this phase.

The scope of upstream processing includes media preparation, inoculum expansion, the main cell culture step (whether batch, fed-batch, or perfusion), and the initial harvest operations that separate cells from the product-containing fluid. Each of these stages must be executed under controlled conditions that maintain cell viability, genetic stability, and product consistency. For a working scientist, understanding the interplay between cell physiology, bioreactor engineering, and process economics is essential for designing robust, scalable processes that meet regulatory requirements.

Upstream vs. Downstream Processing

The biopharmaceutical manufacturing workflow is conventionally divided into two major phases. Upstream processing in biotechnology covers all steps from cell thaw through harvest, where the objective is to generate the maximum quantity of product with the desired quality. Downstream processing begins at the harvested culture fluid and includes clarification, capture, purification, viral inactivation, and formulation. The boundary between the two is the harvest operation.

The distinction matters practically because the two phases require different skill sets, equipment, and analytical approaches. Upstream work is dominated by cell biology and fermentation engineering; downstream work is dominated by protein chemistry and separation science. However, the phases are deeply interconnected. A high-titer upstream process that generates excessive host cell proteins or DNA will burden downstream purification. Conversely, an overly aggressive harvest procedure can damage product integrity. Understanding the downstream and upstream interface is critical for process integration and overall yield optimization.

Key Unit Operations in Upstream Processing

Media Preparation and Sterilization

The upstream process begins with the preparation of culture media. Media formulations must supply all essential nutrients—carbohydrates, amino acids, vitamins, trace metals, and buffers—in concentrations that support cell growth and product formation. For mammalian cells, media are typically complex mixtures containing 50 to 100 components. Common basal media include DMEM/F12, RPMI-1640, and proprietary formulations such as CD CHO or EX-CELL. These are often supplemented with glutamine (2–6 mM), glucose (4–10 g/L), and, for serum-containing processes, fetal bovine serum (5–10%).

Sterilization is a critical step. Heat-labile components such as growth factors and certain vitamins must be filter-sterilized (0.1–0.22 µm pore size), while heat-stable components can be sterilized in situ in the bioreactor at 121°C for 20–30 minutes. In practice, most industrial processes use a combination: the bulk basal medium is heat-sterilized in the vessel, and heat-sensitive supplements are added aseptically after cooling. For single-use systems, pre-sterilized media bags are purchased from suppliers, eliminating the need for in-house sterilization.

Inoculum Expansion

The inoculum expansion stage scales the cell population from a cryopreserved vial to a volume sufficient to inoculate the production bioreactor. A typical expansion scheme for a 2,000 L production bioreactor might proceed as follows:

Thaw one vial of the working cell bank (typically 1 mL containing 1–10 × 10⁷ cells).
Transfer cells to a 125 mL shake flask or T-flask containing 20–30 mL of growth medium.
After 3–5 days, when the culture reaches 2–4 × 10⁶ cells/mL, expand into a 1 L shake flask (100–200 mL working volume).
Scale through a series of spinner flasks or small stirred-tank bioreactors (5 L, 50 L, 500 L) with a split ratio of 1:4 to 1:10 at each step.
Transfer to the production bioreactor at a seeding density of 0.2–0.5 × 10⁶ cells/mL.

The number of passages must be minimized to reduce the risk of genetic drift and phenotypic instability. Most processes are designed to complete expansion in 10–15 population doublings from thaw to production inoculation. Each expansion step typically takes 3–4 days for mammalian cells, meaning the entire inoculum train requires 2–3 weeks. For microbial systems, the timeline is compressed to 1–3 days.

Cell Culture Modes

Three principal modes of operation are used in industrial cell culture:

Batch culture is the simplest mode. All nutrients are added at the start, and the culture proceeds without further additions until harvest. Cell growth proceeds until a nutrient is depleted or a toxic byproduct (typically lactate or ammonia) accumulates to inhibitory levels. Batch cultures typically achieve cell densities of 2–5 × 10⁶ cells/mL for mammalian cells and product titers of 0.5–2 g/L for monoclonal antibodies. The advantages are simplicity and reduced contamination risk; the disadvantages are low productivity and inefficient use of bioreactor capacity.

Fed-batch culture is the dominant mode in industry for monoclonal antibody production. A basal medium supports initial growth, and a concentrated feed solution is added periodically or continuously to replenish nutrients and extend the culture duration. Feeding strategies are designed to maintain glucose above 2–3 g/L and glutamine above 1–2 mM while controlling lactate accumulation. Fed-batch cultures typically run 12–18 days, achieve cell densities of 10–30 × 10⁶ cells/mL, and produce titers of 3–10 g/L. The extended culture duration requires careful control of pH, dissolved oxygen, and osmolarity, which rises as feed is added.

Perfusion culture involves continuous addition of fresh medium and simultaneous removal of spent medium while retaining cells in the bioreactor. Cell retention devices include alternating tangential flow (ATF) filters, tangential flow filtration (TFF) modules, and acoustic settlers. Perfusion allows cell densities of 50–100 × 10⁶ cells/mL and enables continuous operation for 30–60 days. The product is harvested continuously in the permeate, which simplifies downstream integration. Perfusion is used for unstable products, for processes requiring high volumetric productivity, and increasingly in integrated continuous manufacturing. The trade-offs are higher complexity, greater medium consumption, and more demanding process control.

Harvest and Cell Removal

The harvest step separates cells and debris from the product-containing culture fluid. For fed-batch cultures, the entire bioreactor contents are processed at the end of the run. For perfusion, harvest is continuous throughout the run.

The primary harvest step is typically depth filtration or centrifugation. Depth filters, composed of diatomaceous earth and cellulose, remove cells and debris through mechanical sieving and adsorptive binding. A typical train uses a 3–5 µm primary filter followed by a 0.5–1 µm secondary filter. Centrifugation using disc-stack centrifuges is preferred for large volumes (>1,000 L) because it concentrates the cell mass and reduces filter loading. The centrate then passes through a 0.2 µm sterile filter before downstream processing.

Cell removal efficiency is critical because residual cells and debris can release proteases and host cell proteins that degrade product quality. The harvest operation must also be completed quickly—typically within 2–4 hours—to minimize product exposure to degradative enzymes at process temperature.

Cell Lines and Expression Systems

Mammalian Cell Lines

Chinese hamster ovary (CHO) cells are the workhorse of the biopharmaceutical industry, used to produce the majority of approved monoclonal antibodies and recombinant proteins. CHO cells are favored because they perform complex post-translational modifications, including glycosylation, that are required for therapeutic protein function. They grow well in suspension culture, adapt to serum-free media, and can be genetically engineered to produce high titers. Common derivatives include CHO-K1, CHO-DG44, and CHO-S. The dihydrofolate reductase (DHFR) and glutamine synthetase (GS) selection systems are used to amplify and maintain transgene expression.

Human embryonic kidney (HEK) 293 cells are used primarily for the production of viral vectors, virus-like particles, and proteins requiring human-specific glycosylation. HEK293 cells are easily transfected, making them suitable for transient expression systems that enable rapid production of research-grade material. Stable pools and clones are also used for continuous production. The main limitation is that HEK293 cells are less robust than CHO cells in high-density suspension culture.

Other mammalian lines include mouse myeloma (NS0, Sp2/0), baby hamster kidney (BHK), and human retinal cells (PER.C6). These are used for specific products where their characteristics offer advantages, such as particular glycosylation profiles or high specific productivity.

Microbial Systems

Escherichia coli is the most widely used microbial host for proteins that do not require glycosylation, including many enzymes, antibody fragments (Fab, scFv), and peptide hormones. E. coli grows rapidly (doubling time 20–30 minutes), reaches high cell densities (50–100 g/L dry cell weight), and is inexpensive to culture. Proteins are produced either intracellularly as inclusion bodies (requiring refolding) or secreted to the periplasm. The absence of glycosylation machinery limits its use for complex therapeutic proteins.

Saccharomyces cerevisiae and Pichia pastoris are yeast hosts used for proteins requiring some post-translational modification but not mammalian-type glycosylation. P. pastoris is particularly valued for its strong inducible promoters (AOX1) and high secretion efficiency. Yeast systems are used for vaccines, insulin, and various enzymes.

Other microbial systems include Aspergillus and Trichoderma for secreted enzymes, and Lactococcus for certain vaccine antigens.

Selection Criteria

The choice of expression system is driven by product requirements and process economics:

Criterion	Mammalian (CHO)	Microbial (E. coli)	Yeast (P. pastoris)
Glycosylation	Complex, human-like	None	High-mannose
Product titer	3–10 g/L	10–50 g/L	5–20 g/L
Culture duration	12–18 days	2–5 days	3–7 days
Medium cost	High	Low	Low
Product complexity	High	Low–moderate	Moderate
Regulatory precedent	Extensive	Extensive	Moderate

For a therapeutic monoclonal antibody, CHO cells are the default choice because effector functions depend on human-compatible glycosylation. For a non-glycosylated enzyme used in diagnostics, E. coli offers faster development timelines and lower cost. The decision also considers the cell line development timeline, which ranges from 6–12 months for stable mammalian clones to 2–4 months for microbial systems.

Culture Media and Feed Strategies

Media Components

Culture media must provide six categories of nutrients:

Carbohydrates serve as the primary carbon and energy source. Glucose is standard at 4–10 g/L in basal media. Some processes use galactose or fructose to reduce lactate production, as these sugars are metabolized more slowly. Glutamine (2–6 mM) is a secondary carbon source and the primary nitrogen donor for nucleotide synthesis.

Amino acids are required for protein synthesis and metabolism. Essential amino acids (arginine, cysteine, isoleucine, leucine, lysine, methionine, phenylalanine, threonine, tryptophan, tyrosine, valine) must be supplied in the medium. Non-essential amino acids (alanine, asparagine, aspartate, glutamate, glycine, proline, serine) are synthesized by the cell but are often added to reduce metabolic burden.

Vitamins and cofactors include B-group vitamins (biotin, folic acid, niacinamide, pantothenate, pyridoxal, riboflavin, thiamine) and lipoic acid. These are required in micromolar to nanomolar concentrations.

Trace metals include iron (as ferric citrate or transferrin), zinc, copper, selenium, and manganese. Iron is critical for cytochrome function and is often a limiting nutrient in serum-free media.

Lipids and precursors include cholesterol, fatty acids, and ethanolamine. CHO cells require exogenous lipids in serum-free conditions.

Buffers and salts maintain osmolarity (260–320 mOsm/kg) and pH. Sodium bicarbonate (2–4 g/L) is the primary buffer, working in equilibrium with CO₂ in the gas phase. HEPES is sometimes added at 10–25 mM for additional buffering capacity.

Fed-Batch Feeding

The fed-batch feeding strategy determines the trajectory of cell growth, metabolism, and product formation. Two broad approaches are used:

Fixed-schedule feeding adds a defined volume of feed at fixed intervals (e.g., 5% of initial volume every 2 days). This is simple to implement but does not account for culture-to-culture variability.

Demand-based feeding uses online measurements of glucose, lactate, or cell density to trigger feed additions. For example, a concentrated feed (200–400 g/L glucose, 20–40 g/L glutamine, plus amino acids) is added whenever glucose drops below 3 g/L. This approach maintains nutrients in the desired range and reduces metabolic waste production.

The feed composition is typically 5–10× concentrated relative to basal medium, with glucose and glutamine at 50–200 mM. Amino acids are included at levels that prevent depletion of the most limiting component, often cysteine or tyrosine. Some feeds include yeastolate or peptone hydrolysates to provide undefined growth factors.

Lactate control is a key objective. CHO cells in exponential growth produce lactate at 1–2 g/L/day. When glucose is maintained at moderate levels (2–6 g/L) and glutamine is limited, cells shift to a more oxidative metabolism, consuming lactate rather than producing it. Temperature shifts (from 37°C to 31–34°C) at the transition to stationary phase are commonly used to reduce metabolic rate and extend culture duration.

Perfusion and Continuous Culture

Perfusion media are formulated to support continuous cell growth at high density. The medium exchange rate (1–5 reactor volumes per day) must supply sufficient nutrients and remove inhibitory byproducts. Perfusion media are typically more dilute than fed-batch feeds but are used in much larger volumes.

The key operational parameter is the cell-specific perfusion rate (CSPR), defined as the volume of medium supplied per cell per day. A typical CSPR for CHO cells is 0.05–0.2 nL/cell/day. If the CSPR is too low, nutrients become limiting and cell viability declines; if too high, medium is wasted and product is diluted.

Perfusion processes can be operated in two modes: steady-state perfusion, where cell density is held constant by a cell bleed, and intensified perfusion, where cells are allowed to accumulate to maximum density. The latter, often called concentrated fed-batch or N-1 perfusion, is used to generate high-density inoculum for fed-batch production, reducing the seed train duration by 1–2 weeks.

Bioreactor Design and Operation

Bioreactor Types

Stirred-tank bioreactors are the industry standard. They consist of a cylindrical vessel with a height-to-diameter ratio of 1.5–3:1, a marine or pitched-blade impeller, and sparger for gas delivery. Working volumes range from 1 L (development) to 20,000 L (production). The impeller provides mixing and gas dispersion; typical tip speeds are 1–2 m/s, and power inputs are 20–100 W/m³.

Single-use bioreactors use disposable plastic bags or rigid containers instead of stainless steel. They are available in volumes from 50 mL to 2,000 L (with some systems up to 5,000 L). The advantages are reduced capital cost, elimination of cleaning and sterilization validation, and faster turnaround between batches. The disadvantages are higher consumable cost, limited sensor options, and reduced heat transfer capacity.

Other configurations include airlift reactors (used for shear-sensitive cells), hollow-fiber reactors (used for high-density perfusion), and wave-mixed reactors (used for seed expansion). For most industrial applications, stirred-tank and single-use systems dominate.

Process Parameters

Temperature is controlled at 36–37°C for mammalian cells during growth. A shift to 31–34°C is commonly applied during the production phase to reduce metabolism and improve product quality. Microbial cultures use 30–37°C depending on the organism.

pH is controlled at 7.0–7.2 for mammalian cells and 6.8–7.4 for microbial cultures. CO₂ sparging and base addition (sodium bicarbonate or sodium hydroxide) are used to control pH. Lactate production by cells acidifies the culture; base addition compensates.

Dissolved oxygen (DO) is maintained at 30–60% of air saturation for mammalian cells and 20–50% for microbial cultures. DO is controlled by sparging air or oxygen through the sparger and by increasing agitation. Oxygen transfer rate must match cellular consumption, which increases with cell density.

Agitation provides mixing and oxygen transfer. The shear stress generated by agitation must be balanced against cell damage. For mammalian cells, impeller tip speeds above 2.5 m/s can cause significant cell death. The addition of shear protectants such as Pluronic F-68 (0.1–1 g/L) mitigates damage.

Osmolarity rises during fed-batch as feed is added. Values of 300–400 mOsm/kg are tolerated; above 450 mOsm/kg, growth and productivity decline.

Control and Monitoring

Modern bioreactors use distributed control systems (DCS) or programmable logic controllers (PLC) to maintain process parameters within setpoints. PID (proportional-integral-derivative) control loops are standard for temperature, pH, and DO. Cascade control is used for DO, where agitation is increased first, followed by oxygen enrichment.

Gas blending systems mix air, oxygen, nitrogen, and CO₂ to achieve the desired DO and pH. Mass flow controllers regulate gas flow rates. For large bioreactors, multiple spargers (microsparger for oxygen, macrosparger for CO₂ stripping) are used.

Scale-Up and Scale-Down Considerations

Scale-Up Challenges

Scaling up from a 2 L development bioreactor to a 2,000 L production bioreactor is not a simple geometric translation. The key challenges are:

Mixing time increases with vessel size. A 2 L reactor mixes in 2–5 seconds; a 10,000 L reactor takes 30–60 seconds. This creates concentration gradients in pH, nutrients, and dissolved gases that can stress cells.

Oxygen transfer becomes more difficult at scale because the surface-area-to-volume ratio decreases. The volumetric oxygen transfer coefficient (kLa) must be maintained by increasing agitation and sparging, but this increases shear and foam formation.

CO₂ stripping becomes less efficient at scale because the reduced surface area limits CO₂ removal. Elevated CO₂ (above 100 mmHg) inhibits cell growth and alters metabolism.

Heat transfer is more challenging at scale because the surface-area-to-volume ratio decreases. Jacket cooling may be insufficient, requiring internal coils or external heat exchangers.

The standard scale-up criterion is to maintain constant power per unit volume (P/V) or constant impeller tip speed. In practice, a combination of criteria is used, and the process is validated at an intermediate scale (50–200 L) before full production scale.

Scale-Down Models

Scale-down models are essential for process development and troubleshooting. A well-designed scale-down model reproduces the key environmental conditions of the production bioreactor at small scale (1–10 L). This includes matching:

Mixing time and circulation time
kLa and oxygen transfer regime
CO₂ stripping rate
pH control deadband and base addition rate
Temperature control response

For perfusion processes, scale-down models must also match the cell retention device performance and the cell-specific perfusion rate.

Scale-down models are used to test process changes, evaluate new media formulations, and investigate deviations observed in production. A validated scale-down model is a regulatory expectation for process validation and change management.

Engineering Parameters

Several dimensionless and dimensional parameters guide scale-up:

Reynolds number (Re) characterizes flow regime. For stirred tanks, Re > 10,000 indicates turbulent flow, which is required for effective mixing.

Power number (Np) relates impeller power to fluid properties and impeller geometry. For a Rushton turbine, Np ≈ 5.5; for a pitched-blade impeller, Np ≈ 1.3.

kLa is the volumetric mass transfer coefficient for oxygen. Typical values are 5–20 h⁻¹ for mammalian cell culture and 100–1,000 h⁻¹ for microbial fermentation.

Mixing time (θm) is the time to achieve 95% homogeneity after a tracer addition. It scales with the 2/3 power of volume.

Shear rate at the impeller tip is approximated by tip speed divided by impeller diameter. Values above 10,000 s⁻¹ are damaging to mammalian cells.

Process Monitoring and Control

Sensors and Probes

Standard in situ sensors measure temperature, pH, DO, and pressure. These are sterilizable probes that transmit signals to the control system. pH probes use glass electrodes with a reference electrode; DO probes use polarographic or optical (luminescence quenching) principles.

Optical sensors are increasingly used for DO and pH because they are less prone to drift and do not consume oxygen. They are also compatible with single-use bioreactors.

Capacitance probes measure viable cell density in real time by detecting the dielectric properties of intact cell membranes. They provide a signal proportional to the viable biomass concentration, enabling real-time monitoring of cell growth.

Raman spectroscopy is a PAT tool that provides real-time measurements of glucose, lactate, glutamine, glutamate, and product titer. Raman probes are inserted into the bioreactor and collect spectra every 1–5 minutes. Multivariate models convert spectra to concentration values.

Process Analytical Technology (PAT)

PAT is a regulatory framework from the FDA that encourages real-time measurement of critical quality attributes and critical process parameters. The goal is to design, analyze, and control manufacturing processes through timely measurement of quality attributes.

For upstream processing, PAT applications include:

Real-time glucose and lactate monitoring for feed control
Viable cell density monitoring for harvest timing
Product titer monitoring for process optimization
Metabolite profiling for early detection of process deviations

The implementation of PAT requires multivariate data analysis (MVDA) to extract meaningful information from the large datasets generated by online sensors. Principal component analysis (PCA) and partial least squares (PLS) regression are commonly used.

Data Analysis and Automation

Modern bioprocessing generates enormous datasets. A single fed-batch run produces thousands of data points from online sensors, plus additional data from at-line and offline analyses. Data historians store this information for process monitoring, batch reporting, and regulatory compliance.

Automation extends beyond PID control to include:

Automated feeding based on glucose measurements
Automated sampling systems that withdraw and prepare samples for analysis
Automated cell counting and viability assessment
Integration with manufacturing execution systems (MES) for batch tracking

The trend toward continuous processing and digital twins—computational models that simulate the process in real time—is driving the adoption of more sophisticated data analytics. Machine learning algorithms are being developed to predict process outcomes and recommend corrective actions.

Common Pitfalls and Best Practices

Contamination Risks

Contamination is the most feared event in upstream processing. A single contaminated batch can result in product loss, facility shutdown, and regulatory scrutiny.

Bacterial contamination is the most common and is typically detected by turbidity, pH drop, or DO spike. The source is usually a breach in sterility—a leaking seal, a faulty filter, or an error in media preparation.

Mycoplasma contamination is insidious because it does not cause visible turbidity. Mycoplasma can alter cell metabolism and product quality without obvious signs. Routine testing using PCR or culture-based methods is essential.

Viral contamination is the most serious risk. The 2010 incident at Genzyme, where vesicular virus contaminated a bioreactor, resulted in a 6-month shutdown and significant product shortage. Mitigation strategies include raw material testing, viral filtration of media, and rigorous facility design.

Best practices to prevent contamination:

Use single-use systems where feasible to eliminate cleaning and sterilization risks.
Validate all sterilization cycles and filter integrity tests.
Implement environmental monitoring of cleanrooms and biosafety cabinets.
Train operators in aseptic technique and enforce strict gowning procedures.
Test all raw materials, especially animal-derived components, for adventitious agents.

Variability and Reproducibility

Batch-to-batch variability is a persistent challenge. Sources of variability include:

Raw material lot-to-lot variation (especially complex media components)
Cell bank heterogeneity and genetic drift
Inoculum density and viability differences
Environmental fluctuations in temperature, pH, and DO

Best practices to reduce variability:

Qualify raw material suppliers and test each lot before use.
Use defined, chemically characterized media where possible.
Establish a well-characterized cell bank with rigorous release testing.
Standardize inoculum expansion procedures, including passage number and seeding density.
Use statistical process control (SPC) to monitor key parameters and detect trends before they become excursions.

Scale-Up Pitfalls

Common scale-up failures include:

Oxygen limitation at scale due to inadequate kLa. This manifests as DO dropping below setpoint despite maximum agitation and oxygen flow.

CO₂ accumulation at scale due to poor stripping. This causes growth inhibition and altered metabolism.

pH gradients due to slow mixing. Base added at the top of the vessel may not be distributed before cells at the bottom experience a pH spike.

Foaming increases at scale due to higher gas flow rates. Antifoam agents (e.g., polypropylene glycol at 0.01–0.1%) are used, but they can interfere with oxygen transfer and downstream processing.

Shear damage from impeller and sparging at scale. This is mitigated by using lower agitation speeds, larger impellers, and shear protectants.

Best practices for scale-up:

Validate the scale-down model against production data before using it for process changes.
Conduct engineering runs at intermediate scale (10–20% of production volume) before full-scale implementation.
Monitor CO₂ partial pressure in addition to pH and DO.
Use computational fluid dynamics (CFD) to predict mixing and shear profiles at scale.
Document all scale-up decisions and the rationale behind them.

Frequently Asked Questions

What is upstream bioprocessing?

Upstream bioprocessing is the phase of biopharmaceutical manufacturing that covers all operations from the thaw of a cell bank vial through the growth and maintenance of cells to the harvest of the product-containing culture fluid. It includes media preparation, inoculum expansion, cell culture (batch, fed-batch, or perfusion), and initial harvest steps.

What is the difference between upstream and downstream bioprocessing?

Upstream bioprocessing involves growing cells and producing the biological product, while downstream bioprocessing involves purifying that product from the harvested culture fluid. Upstream ends at harvest; downstream begins with clarification and includes capture, purification, viral inactivation, and formulation.

What are the main steps in upstream bioprocessing?

The main steps are: (1) media preparation and sterilization, (2) inoculum expansion from a cell bank vial to production scale, (3) the main cell culture step in a bioreactor (batch, fed-batch, or perfusion), and (4) harvest, where cells are separated from the product-containing fluid.

What are the common cell lines used in upstream bioprocessing?

The most common cell lines are Chinese hamster ovary (CHO) cells for monoclonal antibodies and recombinant proteins, HEK293 cells for viral vectors and proteins requiring human glycosylation, and E. coli for non-glycosylated proteins. Yeast such as Pichia pastoris is used for certain secreted proteins and enzymes.

What is fed-batch culture?

Fed-batch culture is a mode of operation where nutrients are added periodically or continuously during the culture, but cells and product are not removed until harvest. This extends the culture duration and increases cell density and product titer compared to batch culture. It is the dominant mode for monoclonal antibody production.

What is perfusion culture?

Perfusion culture is a mode of operation where fresh medium is continuously added and spent medium is continuously removed while cells are retained in the bioreactor using a cell retention device. This allows very high cell densities and continuous product harvest. Perfusion is used for unstable products and in continuous manufacturing processes.

What are the key parameters to monitor in a bioreactor?

The key parameters are temperature, pH, dissolved oxygen (DO), agitation speed, gas flow rates, and pressure. Additional monitoring includes viable cell density (via capacitance probes), glucose and lactate concentrations (via Raman spectroscopy or at-line analyzers), and product titer.

What are common challenges in scaling up upstream processes?

Common challenges include maintaining adequate oxygen transfer (kLa), preventing CO₂ accumulation, avoiding pH and nutrient gradients due to slow mixing, managing foam, and preventing shear damage to cells. These challenges are addressed through careful engineering design, scale-down models, and validation at intermediate scales.

Key Takeaways

Upstream bioprocessing encompasses all steps from cell thaw to harvest and establishes the foundation for product yield and quality.
Fed-batch culture is the dominant mode for monoclonal antibody production, while perfusion is used for unstable products and continuous manufacturing.
CHO cells are the preferred host for complex therapeutic proteins due to their human-compatible glycosylation and robustness in suspension culture.
Media formulation and feeding strategy are the primary levers for controlling cell growth, metabolism, and productivity.
Scale-up requires careful attention to mixing time, oxygen transfer, CO₂ stripping, and shear, with validated scale-down models serving as the bridge between development and production.
Real-time monitoring using capacitance probes, Raman spectroscopy, and PAT tools enables proactive process control and reduces batch-to-batch variability.
Contamination prevention, raw material qualification, and rigorous process documentation are essential for regulatory compliance and consistent product quality.