Blog

Table of Contents

• Introduction
• 1.Start by thinking at the end
• 2. Scale up train
• 3. Design of a large-scale production reactor
  • 3.1 Sufficient oxygen transfer capacity
  • 3.2 Cell death through shear
• Conclusion 
• References

 Introduction

Scaling up your production process is far from being an easy task. It requires a talented engineering team, with good oversight of the project and the end goal in mind! The scale up process requires time, usually from 3-10 years, and great attention to details. Rushing to meet the deadlines and overseeing something can lead to errors and end up being more costly. The more thorough  you are at each scale up step the less room for error you allow. However, this does not mean that unexpected things can’t happen, so try to prepare for the unexpected as best you can!

So you can be a little more prepared, we have prepared a roadmap that will help you get to your end goal with as few mistakes as possible!

1. Start by thinking at the end 

The most important thing is that you start with the end in mind. To be able to design a production process that will cover the worldwide yearly demand for your product, first, the amount of the  product  demand needs to be established. This will determine what kind of production scale you will need to plan.

Before designing the final production reactor, a few further practical considerations need to be made:

First, put the emphasis on the fermentation step. It requires the most capital and it influences the extent of downstream processing. At this stage, the product loss in downstream recovery needs to be taken into account, which is usually 0.5. This means that at the end of the whole process you will obtain less product, due to loss in every purification step. You will have to make a calculated trade-off between quantity and quality of your product.

Second, you have to decide the working time of the reactor. Will it be working every week of the year, 24 hours, or not. If you are not using single use bioreactors you also have to consider cleaning times between batches.

Third, you have to choose the operation mode of your production bioreactor. After you have decided upon that, you have to take into account the operation mode maximum production concentration. In other words, how much product can be maximally produced in one operation mode run.

With this information you can calculate the total product produced per run. By calculating the total product produced per run with your chosen operation mode you will be able to calculate the total reactor volume needed to achieve your market demand.  

After these initial decisions and estimations have been made, you can start with your actual scale up train.

2. Scale-up train 

Instead of going straight for the production reactor, first use large-scale computational models to identify critical scale-up parameters and do small scale lab and pilot runs to evaluate your critical process parameters. Run the tests early on and often!

Let’s focus on a scale up train for a biopharmaceutical process using mammalian cells. 

For an animal cell culture in a large production bioreactor, a scale-up train is needed to start the production reactor with the desired cell density. To make sure the equipment is utilized optimally, and residence time is no longer than needed, the cells in the inoculum must be in the exponential growth phase. Even though the scale-up is a time-consuming, it is essential because animal cells do not grow at low cell density. 

For the scale-up train, cells are taken from a cell bank and grown in reactors of stepwise increasing volume until enough cells are produced to inoculate the final production reactor. In scale-up, disposable reactors are becoming more popular due to the ease of use and lower operating and validation costs. The whole scale-up process is quite vulnerable to contamination. To prevent contamination, a scale-up reactor often has a bigger volume than the inoculum volume. When the cells have stopped growing, the volume is increased by adding more medium to the cells. This minimalizes the number of passages between reactors and the risk of contamination. 

3. Design of a large-scale production reactor

Large-scale conditions can be simulated and recorded in stirred lab bioreactors with the implementation of control hardware and software.

While designing a large-scale production bioreactor there are many important factors that need to be considered and scaled up very carefully. 

First one being sufficient oxygen supply and transfer throughout the bioreactor. The second factor is maintaining sufficient mixing that will maintain homogenous conditions, but at the same time not cause too high shear forces, leading to cell death.

3.1 Sufficient oxygen transfer capacity

All cells in the large-scale bioreactor need to be supplied with a sufficient amount of oxygen. This means that the cells need to be supplied with at least the same amount of oxygen that they will take up. This can be expressed as follows:

Where OTR is the oxygen transfer rate and OUR is the oxygen uptake rate. 

The respective formulas for OUR and OTR are: 

With the assumption made above, the following equation can be formulated: 

Where, Kol is the oxygen transfer coefficient, A is the specific surface area and C*ol-Col  describes the gradient of concentrations and is also named “the driving force”.   The driving force is the difference between C*ol, which is the oxygen concentration at the top layer of the medium that is in equilibrium with gas phase and Col is oxygen concentration in the bulk of the medium.

The Col is usually a setpoint of DO=50, where most mammalian cells grow optimally. To get the maximum value for the driving force, there needs to be 100% pure oxygen in the headspace (gas phase). After calculating and determining all the variables, A can be calculated using the rewritten equation:

Now, that A is known, it can be used to calculate the gas flow F necessary to supply the cells in this bioreactor with O2. This formula is used, if the sparging is used for aeration and it needs to be taken into account. 

Where, db is the diameter of the bubbles formed which is given, vbw is the rising velocity of the bubbles, and D is the diameter of the reactor.

Finally, gas flow rate (F) needed to supply the bioreactor can be calculated!

3.2 Cell death through shear

If sparging is used for aeration it contributes to cell damage through shear. Therefore, it is necessary to prove that cell death through shear is negligible and for that the growth rate of the cells needs to be higher than the death rate. To calculate the death rate, the hypothetical volume needs to be determined. This can be done experimentally in a small bubble column with small-scale sparging. Data is obtained from samples of every half an hour in a period of 3 hours. The experiments are conducted under two different scenarios with varying gas flow F (dm3/h) and column height H (m).

The hypothetical killing volume can be calculated, using this equation:

Here, Vk is the hypothetical killing volume, Vb is the bubble volume, db is the diameter of the bubble, D and H are the diameter and height of the bubble column respectively.

In addition, there are several more scale-dependent fermentation parameters, but for now we will sum them up in this table:

Conclusion

When undergoing such a big process as scale up is don’t underestimate the power of planning. Do all the risk assessments and make mitigation plans accordingly. Use the knowledge from previous similar projects and consult skilled engineers. But first of all, start by setting realistic expectations!

Refferences

• Seed train optimization for cell culture -PubMed (https://pubmed.ncbi.nlm.nih.gov/24297426/#:~:text=Seed%20train%20optimization%20for%20cell%20culture%20For%20the,are%20described%20for%20the%20seed%20train%20mapping%2C%20%E2%80%A6)
• 4 Process Involved in Scale up Criteria | Microbial Bioprocessing (https://www.biologydiscussion.com/cell-biology/4-process-involved-in-scale-up-criteria-microbial-bioprocessing/7967)
• Scale-up of industrial microbial processes (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5995164/)