Making Cultured Meat Clean; Bye FBS
Fetal Bovine Serum is commonly harvested by inhumane practices, which is why more ethical alternatives are needed
By Polina Parker
Imagine for a second that we’re living in a world where any species on our planet feels safe. By safe I don’t mean ducks lying on a beach and drinking cola, or dolphins having a separate island to relax at weekends. I just mean not keeping live animals in terrible conditions when slaughtering for your Friday party 🎊
There is hope for that if we’re able to scale up meat alternatives, such as a lab-grown one. You may not care about animal welfare, but probably about the future of our planet, because this biotechnology has a chance to decrease greenhouse gas emissions by up to 92%, land by 90%, and water use.
Cultured meat is a part of cellular agriculture that’s produced by deriving stem cells from an animal, following their proliferation and differentiation into muscle. These processes are possible thanks to a culture medium that provides cells with necessary nutrients. Using scaffolding material differentiated muscle or adipocyte cells can take the structure of steak, for example.
What?? I’m not gonna eat a cheeseburger made by a scientist, you crazy!! Yes, public acceptance is a burden. But don’t panic, at least right now, because cultured meat needs to fill much more challenging gaps that will influence whether it can scale up.
The least obvious example would be the fetal bovine serum that’s used in culture medium. A quick fact: FBS is derived from slaughtered cows, yes, I know, it completely contradicts the whole point behind cultured meat. That’s what got me excited to understand why it’s the case, and what are the ways to replace FBS 🐮
Why do we even use FBS?
The culture medium is composed of two main elements tailored to a species’ cell type. The first component is called basal media and provides the essential nutrients: a buffered solution containing glucose, inorganic salts, water-soluble vitamins, and amino acids. The second group includes growth factors that permit long-term maintenance, proliferation, and differentiation of cells. So far, the champion in these functions is our enemy — fetal bovine serum.
A basal medium formulation is enough for cells to grow for a short period, but for cells to proliferate over an extended time, animal sera have been used. Fetal bovine serum is a high-protein mixture containing growth and attachment factors, hormones, antioxidants, and lipids. Numerous growth factors, for example, Insulin-like growth factor 1 (IGF-1) and basic fibroblast growth factor (FGF-2), are present in FBS and activate cell growth and proliferation.
It’s derived from the fetal of slaughtered cows, up to 8% of cows in the slaughter line may be pregnant, making FBS a byproduct of the meat processing industry.
Yes, the use of FBS is not only against all the principles behind cultured meat, but it’s unsustainable and unreliable
- A single liter of serum requires one to three fetuses, with roughly two million fetal calves used in serum collection annually, totalling approximately 800,000 liters of FBS produced per year.
- Despite the long use of FBS (1950s), it has thousands of components and compositions that are unknown, in addition, it depends on geography, diet, hormones and antibiotics, fetal gestational age, and sex. Suggesting external costs in quality control testing, irreproducibility in laboratories across the globe, and conflicting results. FBS is a potential source of contamination from mycoplasma, viruses, and diseases (BSE — bovine spongiform encephalopathy).
- Lastly, there’s a limited global supply of serum, even though the cultivated meat industry is moving to serum-free formulations, cell therapies and stem research have increased the demand that exceeds current availability. The prices have already been increased 300% times over a few years.
With that said, let’s find out some existing FBS-free formulations, and recent experiments to make lab-grown meat truly sustainable!
Starting point — Beefy 8
As a foundation, simple basal media, or Beefy-8 is used. It consists of glucose, amino acids, vitamins, salts, fatty acids, minerals, and proteins. Lab scale production demonstrated a cost of $16/L, which could be reduced with further scale-up, whereas 20% FBS media for satellite cells costs $200–500/L.
Reducing FBS concentration. The experiment showed that B8 media combined with BSC-GM (20% FBS) significantly improved growth compared to BSC-GM alone over four days, this benefit remains with as much as 62.5% reduction in FBS. Even more, a serum reduction of 87.5% didn’t significantly reduce the cell growth over four days.
On the other hand, B-8 media alone encouraged cell growth only over three days (it didn’t continue into day four), suggesting that B8 can reduce the percent of serum, but not eliminate it fully. So, not the best option.
It’s time to go forward, Albumin
The next step was to test numerous supplements in different concentrations to overcome the Beefy-8 deficiency.
In this experiment, six growth factors improved BSC (bovine stem cell) proliferation, when compared to B-8: interleukin, curcumin, recombinant albumin, platelet-derived growth factor (PDGF-BB), linoleic and oleic acid. Of these, recombinant albumin was the driving factor, the imposter 😈
So, B8 media was supplemented with recombinant Albumin (800 μg/mL), and termed — Beefy-9. It was able to maintain short-term growth compared to serum-containing media in a short-term period.
However, seeding cells into Beefy-9 media directly after passaging showed ineffectiveness, as cells didn’t re-adhere to tissue culture plastic. There were two possible explanations for this outcome: the coating used was insufficient to enable cell adhesion (in the absence of adhesion factors in serum), and the second was the high concentrations of rAlbumin that outcompeted laminin to absorb the tissue-culture plastic. To overcome these limitations, BSC was used with various concentrations of adhesive proteins, and rAlbumin was added one day after passaging to allow cells to adhere. Since the priority was simplicity, focus was placed on low-cost recombinant adhesive proteins that didn’t require pre-coating.
Comparing different peptides, results showed that truncated vitronectin (Vtn-N 1,5μg/cm²) showed superior cell adhesion and growth than PDL (Poly-D-lysine coatings) alone, laminin alone, PDL + laminin, lower concentrations of Vtn-N with or without PDL. A winner, in other words.
In this and the following experiments we always need to consider how the removal/ addition/ any change affects differentiation (formation into muscle 🥩). To measure it, these factors are typically used:
MyoD and Myogenin — transcription factors that bind to DNA and turn on necessary genes for muscle cell differentiation. MyoD is one of the first proteins to be expressed whereas Myogenin comes into play later to help with the final stages of muscle cell maturation.
MHC — motor protein that is responsible for muscle contraction, different types determine its functional properties.
Pax7 keeps and reserves satellite cells (stem cells in muscle tissue) so they can be used when muscle repair is needed.
Alright, coming back to Beefy-9!
Differentiation and expansion
When the delayed rAlbumin and 1,5 μg/cm2 of Vtn-N were established in Beefy-9, the myogenicity was verified. Cells were expanded in the media for 5 days showing the formation of multi-nucleated myotubes that stained positive for the MHC. Then, the cost was reduced by lowering the FGF-2 growth factor to 5 ng/mL and 1.25 ng/mL without significantly affecting the growth. However, the complete removal of FGF-2 significantly affected the cell growth and morphology.
Long-term culture
Phenotypic analysis revealed the maintenance of BSC stemness (gives rise to differentiated cells) over long-term culture: 96% of cells in BSC-GM and B9 stained Paired-box 7 positive (early satellite cell marker), with no significant difference between types. Similarly, qPCR and analysis of gene expression for the Pax3 gene showed no significant difference, suggesting that B9 maintains satellite cell identity to a similar degree as serum-containing medium. In both media, the formation of Myogenin and MHC-positive multi-nucleated myotubes was shown. Adding to that, improved differentiation appeared in B9, compared to serum-containing media, in terms of myotubes size, density, and fusion index. That could be possible due to the amount of doubling cells undergone in BSC-GM, representing a more aged population.
Just tell me, how much?
Even when using purchased growth factors and without bulk ordering, the cost of B9 is less than serum-containing ($217 to $290/L). The cost depends on the concentration of FGF-2 growth factor, with high and low dropped to $74 and $46/L — resulting in a 75% cost reduction on a per-liter basis. On the other hand, these cost savings are mitigated by reduced growth. That’s why an optimal concentration of 3.2 mg/mL of rAlbumin was found, which may offer the best ratio of cell growth to media cost — $148.
Even though the addition of recombinant Albumin to B8 makes it suitable for BSC expansion, it resulted in a 50–400% increase in cost, compared to B8 alone.
Low-cost extracted proteins
So, the approach to using the extraction from low-cost and abundant sources, such as Oilseed protein isolates (OPI), was suggested.
Specifically, the isolates from four plant sources were tested: Inca peanut, soybean, rapeseed, and cottonseed. They were prepared using alkali extraction, isoelectric precipitation, centrifugation, filtration, and ultrafiltration, resulting in isolates. Since Inca protein was originally produced for human consumption, it was expensive ($27.6/kg), the three other OPIs were the byproduct of food oil production at a relatively low cost ($0.4/kg).
Following extraction and concentration, OPI yields were 6.5g/kg (IPPI), 188g/kg (SPI), 39g/kg (RPI), and 39g/kg (CPI). Using yields and available data on protein costs, the OPI starting material costs were $4.255, $1.75, $6.10, and $3.21 per kg, respectively. Not including processing costs, the costs of SPI, RPI, and CPI are lower than the $45.000/kg for rAlbumin 🤯
Protein isolates replacing rAlbumin
To test which of the OPIs works best in different media formulations (1. B8 with no supplementation, 2. B8 with OPIs, or 3. B9), bovine stem cells were cultured.
In this round, rapeseed isolate (RPI) at 0.4 mg/mL performed the best, completely recovering rAlbumin. That’s how the new media was established and termed Beefy-R: consisting of Beefy-8 with 0.4 mg/mL RPI.
Compared to B9, Beefy-R resulted in:
- Improved growth over four passages
- Over 2 weeks, twice as many cells were produced
- Increased population doublings (PD) from 10.6 to 11.7
All media types showed an increase in doubling time with increased passages. Beefy-R resulted in slower deterioration (declining) of doubling rate than B-9, but faster than BSC-GM. Shows that RPIs are not only replacing albumin but offer improved functionality, though not to the degree of BSC-GM, still.
BSCs myogenicity and identity
P.S. We’re gonna use the factors I mentioned above to analyze the effectiveness of the media.
BSC-GM showed a significant increase in MyoD (10 folds), and a decrease in Myogenin compared to B9 and Beefy-R, suggesting that BSC-GM gained an earlier muscle phenotype. Beefy- R showed an increase in MHC (muscle movement) compared to BSC-GM, suggesting enhanced differentiation of cells, following proliferative phenotype.
Proliferative cells showed Pax7 expression and heterogeneous MyoD expression for all media types. qPCR and immunostaining revealed that cells in Beefy-R media (RPI) maintained myogenic phenotype at least as Beefy-9 but in a later stage rather than cells in serum-containing media.
Self-sufficient cells
The dominant imposter (cost contributor 👿) in Beefy-R media is recombinant FGF2, taking >60% of the cost. Logically, we want to replace this growth factor, as we did with rAlbumin, right?
But wouldn’t it be more fun and sustainable for cells to express the FGF2 growth factor themselves? We’re about to answer this question!
FGFs (FGF-2 is just a part of this family) act by binding to cell surface FGF receptors and activating mitogenic pathways. That results in growth and cell fusion.
Before the breakthrough
Initially, it was observed that mouse myoblasts (MM14) naturally express some FGFs, but they’re still dependent on exogenous FGF2 to proliferate. Also, the overexpression of additional FGF2 can partially reduce the dependency, but can’t fully recover growth.
Finally, the overexpression of Ras mutant can amplify the effects of endogenously expressed FGF2 in MM14, but it’s not sufficient to induce comparable proliferation in serum-free media.
Based on these findings Andrew Stout and his team engineered induced BSCs to express FGF2, Ras, or both FGF2/ Ras.
The absence of exogenous FGF2
Studies showed that cells in serum-free media couldn’t proliferate without exogenous FGF2, whereas engineered cells (FGF2, Ras, and FGF2/Ras) grew continuously under the same conditions. The combined expression of both FGF2 and Ras suggested improved growth compared to single gene expression and was able to fully recover the proliferation of control cells in 40 ng/mL rFGF. Cells engineered for autocrine signaling had similar growth whether or not rFGF was present in the media.
During proliferation, cells were assessed via immunostaining for markers, such as Pax 7 (satellite cell phenotype), MyoD, Myogenin, and MHC. Staining revealed high levels of expression of Myod and Pax7, over 97.5% of nuclei in all cell types stained Pax 7 positive. Results showed no significant difference except in FGF2 cells (which was reduced Pax3), suggesting that Ras and FGF2/Ras offer the highest phenotypic similarity to controls.
Differentiation sucks
All cell types were able to differentiate into myotubes. Gene expression showed that Ras and FGF2/Ras showed higher expression of Pax3 than control cells, but lower of MyoD and Myogenin. Resulting in a less differentiated phenotype.
Additional time could overcome this discrepancy, as cells may still experience lingering signaling two days after doxycycline (antibiotic) removal.
Constructing Lego 🧊
Another way of looking at serum-free alternatives is to construct, or chemically define a media, by testing how different growth factors affect the overall growth. Yes, it sounds much easier than the cellular engineering above.
This study develops a chemically defined serum-free medium that supports the attachment and expansion of bovine myoblasts. This process feels like a Lego for me! I listed the main concepts used to define the efficient medium. It all started with Ham’s F10 Nutrient mix as a basal media.
1) Presence/ absence affects proliferation
The goal of the first step was to check whether insulin, transferrin, and sodium selenite were necessary. If yes, compare the existing solution containing insulin (I), transferrin (T), sodium selenite (S), and ethanolamine (X), known as ITS-X, to the initially adopted media. Comparing the media containing I/T/S with the one containing 1% ITS-X commercial mix suggests that 1% of ITS-X increased proliferation.
When the absence of a component doesn’t significantly reduce proliferation or cell attachment, it can be removed. For example, in this case, due to the cost of vitronectin, it was tested whether it can be excluded. The highest yields without vitronectin were achieved rather in the presence of fetuin, or fetuin and fibronectin. So, it was removed from the media.
2) Act synergistically
The SFM was supplemented with Asc-2-p and sulfate HS to optimize the media. Even though, there’s no significant effect from the presence of both, they were included in the media since heparan sulphate and Asc-2-p may act synergistically with other growth factors.
3) Human perception
When comparing two basal mediums, Ham’s F10 was replaced with DMEM/F12 showing a better effect on the bovine myoblasts proliferation.
But then, it was considered that Dexamethasone is a synthetic drug raising concerns for human consumption in cultured meat. It was replaced by hydrocortisone, a glucocorticoid hormone occurring in FBS. In both cases, 44% of growth compared to the control was reached on day six.
4) Outperform FGF2
The next step aimed to identify the combination of growth factors that matches or even outperforms the growth in FBS. The following factors and cytokines were used: IGF-1, HGF, LIF, EGF, PDGF-BB, and VEGF. Resulting in an experimental matrix of 64 conditions.
- IGF-1 100 ng/mL. High IGF-1 concentrations (30–100 ng/mL) induce satellite cell proliferation and produce muscle hypertrophy.
- HGF 5 ng/mL. Heparin-binding protein activates satellite cells and regulates cell proliferation while acting synergistically with FGF-2 or PDGF.
- LIF 5 ng/mL. Cytokine LIF is added to inhibit spontaneous differentiation and stimulate cell proliferation.
- PDGF-BB 10 ng/mL. Mitogen that promotes human muscle stem cell proliferation when supplemented with FGF-2 and TGF-B1.
- EGF 10 ng/mL. Stimulate DNA synthesis with IGF-1 and PDGF.
- VEGF 10 ng/mL. Increase muscle-derived stem cell proliferation, apoptotic effect.
5) Main pioneers
As a result, three different growth factor combinations were nearly as effective as a serum-containing GM. SFM supplemented with IGF1, LIF, PDGF-BB, VEGF, and HGF supported exponential growth as effectively as FBS-containing media.
6) Optimized concentrations
Once the effective components were identified, the researchers performed an analysis to find optimized component concentrations. BSA (bovine serum albumin), ITS-X, ALA (α-linoleic acid), hydrocortisone, CGH (cow growth hormone), heparan sulfate, L-ascorbic acid, fibronectin, fetuin, human IL-6 were added at higher, lower and omitted completely.
- The absence of BSA resulted in a significant cell confluency (surface area covered).
- ITS-X didn’t have a significant effect in the short term. The absence of ALA significantly reduced cell proliferation.
- CGH had a deleterious effect on bovine satellite cells, so it was excluded (in fact, an increase in concentration didn’t result in the reduction of cell growth).
- Heparan sulphate didn’t have any positive impact on proliferation so it was removed.
- L-ascorbic acid didn’t show a significant effect in the short term but was projected to add long-term benefits, so it wasn’t removed.
- Even though the absence of fibronectin resulted in a significant increase in cell growth, it was kept in media due to commercial availability and the requirement for attachment factors.
- hIL-6 was necessary for the absence of LIF.
So what …?
Yeah, some stats and explanations above seem complex. This study is just an example of what chemically defined media is and what principles are used. Promise, that’s the last one.
For SFM long-term culture, DMEM/F12 basal media was supplemented with L-ascorbic acid, fibronectin, hydrocortisone, GlutaMAX, HSA, ITS-X, hIL-6, ALA, FGF-2, VEGF, IGF-1, HGF, PDGF-BB. It was exchanged every 3–4 days for long-term passaging. So, FGF-2 was doubled to 10 ng/mL to compensate for its degradation at 37 °C.
Yo, that’s a story of what has happened in the last 2 years. Based on this research, now I see two clear ways to address the FBS challenge:
Breaking down and replacing imposters
There’s a chance to find more effective and sustainable FBS alternatives by getting rid of the imposters (the most expensive ones/ difficult to get) and replacing them. As we saw with B9, Beefy-R, and chem-defined.
Things to keep in mind:
- By breaking down the existing formulations, we need to consider factors such as differentiation capacity (Ras expression), growth time, and adherence (in the case of rAlbumin which was derived one day after with a vitronectin peptide).
- Always seeking the right concentration, especially when removing growth factors (B8 wasn’t able to eliminate FGF requirements, so FGF2 was lowered to 5 ng/mL as the most optimal concentration).
- As we covered, FGF removal is possible but requires to sacrifice cell growth.
- Even if some factors don’t significantly increase/ decrease growth, they may add value by syncing with each other (Asc-2p and sulphate).
- Human perception affected the decision to choose the basal media between F-10 and DMEM/F12.
- The cost and current availability (the case with fibronectin), even though fetuin showed increased growth.
First Principles
It means rethinking right from the beginning. Why do we need to care and optimize growth factors ourselves if a cell can express them itself? Really.
Excited to go further with this question. Thank you!! I know, it was challenging reading my first review!!
If we haven’t met, I’m Polina, a mad scientist. Seek adventures and innovations, so now working in the cellular agriculture field to end animal suffering (a challenging one!). Hit me up on LinkedIn or Twitter, and stay updated with my monthly challenges.
