Reflections on cold sterilization
According to a Good Food institute, bioreactor costs are one of the biggest hurdles for alternative protein manufacturing. I mean we are talking bigger limiting factor than scaffolding, which I would say is a rather big challenge. Further they also show that bioreactors are the second most urgent thing in need of cost reductions, only surpassed by prices of growth factors and peptides. Though I didn't come across similar data for fermentation, I'd be surprised if the topic wasn't equally, if not more, important here.
In this block post I'd like to come with a few reflections on what drives bioreactor costs, and how changing some of the ways things are done today can dramatically reduce costs, especially in large scale. Steam sterilization – a main driver of cost
Depending on size there are various drivers in the cost, but commonly, one of the biggest contributors is the equipment and requirements to handle high pressure steam during sterilization. On the smaller sizes (500L and smaller) it is mainly related to equipment and sensors, and while the components themself are not very expensive, wiring, welding, and handling of the extra components adds up. On the larger sizes the requirement to the tank drives the price up exponentially. This is not just due to being able to handle pressure, but to a much higher degree to be able to handle the thermal expansion during sterilization, and requirements to being able to handle vacuum which can occur during the cooling of the tank.
But what to do about it? After all we need aseptic conditions.
What does it mean to be aseptic?
While it is not uncommon to talk about aseptic processes as being equipment or media which experience a certain heat for a given time. However, I'd like to draw the attention to the fact that we do this to achieve an adequate likelihood that our batch will fail, also known as sterility assurance level or SAL. In a pharmaceudical context a 1 in a million chance is the usual standard, which is often adopted by the food industry as well. This makes sense in an end product scenario, where we want to avoid harmful organisms and often a long shelf life. However, in bioreactors it usually serves a different purpose. Since both filamentous fungi and mammalian cells grow slow compared to other strains it is crucial for a successful production not to be outcompeted by other organisms. This change in purpose has a significant consequence - now we are not talking about an end product quality metric, but a production parameter. From this it makes sense to reflect on a cost vs benefit in setting a high SAL.
What happens if there is a contamination? Well, the batch is lost, and results in a loss in productivity. However, there are many other causes which also results in loss in productivity, and I would argue that the operator, component and sensor failure or inadequate sterilization of the media are a much higher cause of loss in productivity than one in a million. Thus, I would argue that it does not make sense to pay for a higher security than one i a thousand. Even for a fast fermentation time (say one batch per day) this would mean one lost batch per 3 years, and for most we are talking once a decade.
Reaching sterility
After having defined our sterility level of 1:1000 risk of contamination, it makes sense to look at the likely hood of contaminants coming into the system in the first place. After the initial sterilization, the process should be sterile, so how does it become contaminated? Well, that's simple, that is during our cleaning process. Often because contractors try to save money on designing the CIP system.
That said, a properly designed CIP system should not result in any measurable contaminants (which is after all the whole point of cleaning). For a cheese factory I know that it is not uncommon to ask that five 10 x 10 cm swap tests show a result of 0 CFU after CIP, In that case, let's estimate a worst case contamination scenario, where a sixed swap test would have found contaminants. This gives a worst-case scenarios of 1/6 CFU pr 100 cm2 or 16,7 CFU/m2. Doing a bit of simple math this results in the following sterilization needs
Why is this important? Well, because chemical sterilization can reach a log reduction of between 4 to 6 (depending on organism) in a matter of 5-15 minutes, comparing it with the hour or two used to sterilize a bioreactor with steam I would be surprised if adequate levels could not be reached with an oxidizing agent and a good hygienic design. This would not only save a lot of costs, but likely also shorten the time between batches, and reduce the amount of energy required. I haven't done the full analysis of how this would impact a sustainability report, with the high amounts of energy used for sterilization I would expect it to have a positive impact on industrial sizes.
Conclusion
So what is the conclution here? I believe there is a big potential in cold sterilization, but to acheave this, we need to stop talk about log reductions as a mean to sterility, and rather talk about bioburden after CIP and what SAL we want to acheave. In case there is a detection of CFU's in the swap test, I would argue that estimating contamination levels should be quite straight forward.