Growing Growth Factors: Plant based production

This week marks the CellGS market debut of plant derived TGF-β3. Recombinant growth factors are among the most expensive consumables in a cell culture laboratory, and for a long time there were only three realistic ways to make them: bacteria, yeast or mammalian cells. A fourth route has been quietly maturing for the best part of two decades. Plant molecular farming, the production of recombinant proteins in engineered plants or plant cell cultures, is now supplying commercial growth factors and cytokines to stem cell researchers, cell therapy developers and the cultivated meat sector.

The interest is not purely scientific. Growth factors are the dominant cost line in serum-free media, and the economics of several emerging industries depend on bringing that cost down. Plants offer a production route with no bioreactor, no animal-derived components and no bacterial endotoxin, and with a scale-up path that looks more like agriculture than bioprocessing.

Here we look at how plant expression systems work, what the published evidence says about the proteins they produce, where the commercial activity is, and what remains unresolved. We close with a short section on transforming growth factor beta 3 (TGF-β3), a protein with a long history in both plant expression research and cell culture practice.

Why Plants At All

Plant molecular farming has a history of nearly forty years, beginning with monoclonal antibodies expressed in plants in 1989 and hepatitis B surface antigen in 1992.¹ The early vision of edible vaccines grown in bananas did not survive contact with dosage control and regulatory reality, and the field consolidated around purified products from a small number of hosts.

The attraction for protein production rests on a handful of properties:

• No animal or microbial contaminants. Plants do not carry mammalian viruses, prions or oncogenes, and they do not produce bacterial endotoxin. For products destined for cell therapy or regenerative medicine, this removes a whole class of release testing anxiety.
• Agricultural scale-up. Once a stable line exists, increasing output means planting more. The capital cost of a greenhouse is a fraction of a stainless steel fermentation suite.
• Eukaryotic folding machinery. Unlike E. coli, plant cells can form disulphide bonds and carry out post-translational modification, which matters for structurally complex growth factors.
• Containment options. Barley is self-pollinating, and chloroplast transgenes are maternally inherited and so are not transmitted in pollen. Both features address the gene flow concerns that have dogged transgenic crops.

Against this, Rybicki's 2026 review of the state of the art is candid about why the pharmaceutical industry has not embraced the technology: relatively low yields of active ingredient per gram of biomass, inconsistent product quality from whole plants, and persistent difficulties with scale-up and downstream processing.¹ The same review notes that modern plant molecular farming has narrowed considerably, with most activity now using transient expression in Nicotiana benthamiana rather than the wide range of transgenic crops explored in the 1990s.

That consolidation is worth keeping in mind. The companies selling plant-made growth factors today mostly use stable transgenic systems that the review describes as niche, which suggests the reagent market and the pharmaceutical market have different requirements and are diverging rather than converging.

The Main Plant Platforms

Four distinct approaches dominate commercial and near-commercial growth factor production, and they differ in more than just the species involved.

The chloroplast route has produced some striking accumulation figures. Codon optimisation for chloroplast usage raised TGF-β3 accumulation 75-fold over the native sequence in tobacco plastids, to the point where the 13 kDa monomer band on a Coomassie-stained gel was more intense than the RuBisCO small subunit.² A fusion of green fluorescent protein and epidermal growth factor reached 10.2% of total soluble protein, equivalent to 1.57 g per kg of fresh leaf.³ Basic fibroblast growth factor expressed the same way reached a more modest 0.1% of total soluble protein.⁴ The spread across those three proteins indicates that yield is protein-specific rather than a property of the system.

The barley seed platform has the longest commercial record. Human Flt3 ligand expressed in barley endosperm was purified to near homogeneity by two chromatography steps and showed biological activity comparable to the commercial reference, with yields similar to a bacterial expression system.⁵ The protein was glycosylated, carrying alpha-1,3-fucose and alpha-1,2-xylose, the plant-specific sugars discussed below. That work was published in 2010 and the products derived from it have been on the market since.

What Plant Production Does And Does Not Solve

Two technical issues come up repeatedly:

Glycosylation. Plant glycoproteins carry core beta-1,2-xylose and alpha-1,3-fucose, and lack the beta-1,4-galactose and sialic acid found on mammalian glycoproteins.¹ This has been raised for years as an obstacle for plant-made therapeutics. In practice the position is more relaxed than the theory suggested. Plant-made influenza virus-like particle vaccines given to subjects with pre-existing plant allergies elicited only transient antibody responses that were not associated with hypersensitivity symptoms, and the approved plant-made enzyme taliglucerase alfa is produced in carrot cells without modified glycosylation.¹ Where glycan structure does matter, glycoengineered plant lines now exist that can produce galactosylated, fucose-free and even sialylated glycans.¹ For research-grade growth factors used in culture medium, glycan differences are a characterisation question rather than a safety one, though they can affect half-life and potency and should be checked against the reference material.

Folding of cysteine-knot proteins. This is the harder problem. The TGF-β family forms a disulphide-linked homodimer with a cysteine knot, and the chloroplast is not an oxidising compartment. PUblished tobacco chloroplast work on TGF-β3 accumulated the protein in insoluble aggregates and required chemical refolding to recover activity.² Work on TGF-β1 in N. benthamiana ran into a related problem from the other direction: the latency-associated peptide was not cleaved by a furin-like protease in planta, and mature TGF-β1 expressed without the peptide was inactive until furin was co-expressed.⁶ Anyone evaluating a plant-made TGF-β should therefore be asking about the refolding or processing step and the bioassay data that follow it, not just the expression yield.

Downstream processing is the quieter constraint. Techno-economic work on horseradish peroxidase found that transient expression in N. benthamiana was only economic above roughly 5 kg of product per year, although fusing the enzyme to an antibody to add value changed the calculation entirely.⁷ Scale and product value interact, and the trade-off needs settling early rather than after a facility has been built.

The Market Case: Growth Factors Are The Cost

The commercial momentum behind plant-made growth factors comes mostly from one number. In serum-free media for cultivated meat, growth factors and recombinant proteins account for the bulk of the cost, and serum-free media in turn makes up at least half of variable operating costs in manufacture.⁸ The Good Food Institute's analysis of the widely used Essential 8 formulation put 99% of medium cost in the growth factors, with FGF-2 and TGF-β alone responsible for 96%.

That is the opportunity the plant companies have targeted. Bright Biotech, a Manchester company founded in 2019 and using chloroplast expression in tobacco, raised $3.2m in an oversubscribed seed round in late 2022 led by FoodLabs with Big Idea Ventures and CPT Capital participating. Its stated calculation is that replacing commercial growth factors in animal-free medium with plant-made equivalents takes medium cost from $376 to $21.70 per litre. Core Biogenesis, founded in France in 2020, raised a $10.5m Series A in 2022 to build a Strasbourg facility for growth factors and cytokines expressed in Camelina sativa seeds, and has since reported four transgenic camelina lines expressing Activin A, FGF-2, human serum albumin and TGF-β, all with confirmed bioactivity. Tiamat Sciences, working on the same problem from North Carolina, raised seed funding in 2021 and 2023.

Several of these companies have found that regenerative medicine is a more immediate market than cultivated meat. Core Biogenesis stated early that its first target was the cost and complexity of cell therapy manufacture rather than food, and its products are now distributed to European research customers with claimed purities above 95% and prices 30 to 50% below market standards. The same platform has moved sideways into cosmetics, where oleosome-fused EGF and FGF-2 are being positioned as retinol alternatives.

ORF Genetics, the Icelandic company whose barley platform produced the Flt3 ligand work cited above, is the longest-established player and shows what a mature version of this looks like. It runs three product lines from one platform: ISOkine for stem cell research, MESOkine for cultivated meat and DERMOkine for skincare, with barley grown hydroponically in greenhouses powered by geothermal energy. Its growth factors are supplied as animal-component-free and endotoxin-free, with stated endotoxin levels below 0.005 ng per µg. In February 2026 it announced a distribution partnership with Tebubio covering the ISOkine range across Europe, which is a reasonable marker of the category moving from novelty to catalogue item.

Estimates for plant molecular farming vary widely depending on what is counted, with one 2025 forecast putting the market at $219m in 2024 growing to $1.34bn by 2032. These figures are directional rather than precise, and they bundle vaccines and antibodies together with reagent proteins. 

TGF-β3 Applications

TGF-β3 is one of three closely related isoforms in the transforming growth factor beta family. The isoforms share 70 to 80% amino acid identity in their mature regions and bind the same receptors, yet their biological effects are not interchangeable.⁶ Mature human TGF-β3 is a 112 amino acid chain released from a 412 amino acid precursor by furin-like cleavage of the latency-associated peptide, and the active molecule is a disulphide-linked homodimer. The mature sequence is 100% identical to mouse, dog and horse TGF-β3, which is convenient for anyone working across species.

Three applications account for most laboratory use:

Chondrogenesis and cartilage tissue engineering. This is the dominant use. The classic protocol cultures human mesenchymal stem cells in micromass pellets with 100 nM dexamethasone and 10 ng/ml TGF-β3, producing a matrix containing type II collagen, aggrecan and anionic proteoglycans within 14 days.⁹ Later work found that TGF-β3 has higher chondrogenic potential than TGF-β1 and drives more rapid differentiation,¹⁰ which is why it became the default isoform for cartilage work despite TGF-β1 being cheaper and more widely available. The effect is enhanced by appropriate matrix: hyaluronic acid coated surfaces with TGF-β3 supplementation gave a pronounced increase in type II collagen, aggrecan and Sox9 expression over either alone.

Pluripotent stem cell maintenance. TGF-β is one of the four growth factors in Essential 8 and its derivatives, alongside insulin, transferrin and FGF-2, and as noted above it is one of the two proteins that dominate the cost of that formulation. Anyone running iPSC culture at scale has a direct financial interest in where their TGF-β comes from.

Scarless wound healing. The interest here came from the observation that TGF-β3 is relatively elevated in tissues that heal without scarring, such as early gestational foetal skin and oral mucosa. Renovo developed recombinant human TGF-β3 as avotermin, trade name Juvista, and three double-blind placebo-controlled Phase I/II studies reported improvements in scar appearance after intradermal administration. The Phase III REVISE trial in more than 350 scar revision patients missed both its primary and secondary endpoints in February 2011, Renovo's share price fell 75% in a day, Shire terminated its licensing agreement, and the company wound down.

TGF-β3 is often associated with anti-scarring. The biology that motivated the programme is still valid, but the discrepancy between the controlled Phase I/II populations and the broad Phase III population been fully explained. It remains an open question rather than a settled negative, but the clinical claim has not been made good.

The plant connection runs through all of this. TGF-β3 was one of the first growth factors seriously pursued in chloroplasts precisely because a sustainable source free of animal pathogens would suit a product intended for wound application, and the 2011 tobacco work concluded that chloroplasts were an attractive production platform for it.² The protein's structure is also what makes it a demanding test of any expression system, for the reasons set out above.

The question of equivalence has not been settled in a general way. Individual proteins have been shown to match their conventionally produced counterparts in specific assays, but there is no body of comparative work covering the growth factors most used in culture across the cell types they are used on. Buyers are currently reliant on supplier bioassay data, which varies in how closely it resembles their own application.

Nor is it clear how the cost advantage behaves at different scales. The arguments for plant production are strongest at kilogram scale, where capital cost per gram dominates. Most research customers buy in 25 to 1,000 µg packs, where purification and quality control costs are a far larger fraction of the total, and the advantage may be thinner than the headline comparisons imply.

There is also an uncomfortable mismatch between where the field's technical momentum sits and where the products are. Rybicki describes stable transgenic and chloroplast expression as niche relative to transient expression in N. benthamiana,¹ yet every commercial growth factor platform listed above uses a stable system. Either the reagent market has different requirements that make stable lines the right answer, or the growth factor companies are building on a foundation the wider field is moving away from. It is too early to say which.

References

1. Rybicki EP. Plant molecular farming for pharmaceuticals: the state of the art. npj Sci Plants. 2026;2:22.
2. Gisby MF, Mellors P, Madesis P, et al. A synthetic gene increases TGFβ3 accumulation by 75-fold in tobacco chloroplasts enabling rapid purification and folding into a biologically active molecule. Plant Biotechnol J. 2011;9(5):618-628.
3. Wang Y, Fan J, Wei Z, Xing S. Efficient expression of fusion human epidermal growth factor in tobacco chloroplasts. BMC Biotechnol. 2023;23(1):1-9.
4. Stable expression of basic fibroblast growth factor in chloroplasts of tobacco. Int J Mol Sci. 2016;17(1):19.
5. Erlendsson LS, Muench MO, Hellman U, et al. Barley as a green factory for the production of functional Flt3 ligand. Biotechnol J. 2010;5(2):163-171.
6. Wilbers RHP, Westerhof LB, van Raaij DR, et al. Co-expression of the protease furin in Nicotiana benthamiana leads to efficient processing of latent transforming growth factor-β1 into a biologically active protein. Plant Biotechnol J. 2016;14(9):1695-1704.
7. Walwyn DR, Huddy SM, Rybicki EP. Techno-economic analysis of horseradish peroxidase production using a transient expression system in Nicotiana benthamiana. Appl Biochem Biotechnol. 2015;175:841-854.
8. Exploring cost reduction strategies for serum free media development. npj Sci Food. 2024.
9. Mackay AM, Beck SC, Murphy JM, Barry FP, Chichester CO, Pittenger MF. Chondrogenic differentiation of cultured human mesenchymal stem cells from marrow. Tissue Eng. 1998;4(4):415-428.
10. Barry F, Boynton RE, Liu B, Murphy JM. Chondrogenic differentiation of mesenchymal stem cells from bone marrow: differentiation-dependent gene expression of matrix components. Exp Cell Res. 2001;268(2):189-200.

Cell Guidance Systems supplies a range of recombinant growth factors for human and other species, alongside PODS® depot growth factors for sustained release in culture. For three-dimensional and chondrogenic culture work, our PeptiGel® synthetic peptide hydrogels and defined surfaces and ECMs may be of interest, and further background is available in our PODS® technology resources.

Main Image: Tobacco field. Credit: Shutterstock

Learn more about cytogenetics and cell characterisation from Cell Guidance Systems: