Thermostable FGF2-G3 for weekend-free stem cell culture

Qkine licensed the patented FGF2-G3 technology from Enantis/Masaryk University to manufacture and provide thermostable (heat stable) FGF2-G3 for stem cell culture and emerging applications such as cellular agriculture.

At Qkine, we have combined the excellent science behind the FGF2-G3 technology with our protein manufacturing expertise to provide the best quality protein. His tags may cause issues for scientists looking to translate discoveries to clinical or scale-up applications. We therefore optimized the production of FGF2-G3 to remove His tags present in the academic constructs. Tags are challenging to explain to regulatory bodies and introduce scientific uncertainties.

Two forms of FGF-2 are used frequently in stem cell culture: the 146 aa (or 145 aa) form and a 154 aa form with a 9 residue extension at the N-terminus. The 9 aa pro-segment of FGF-2 is not required for biological activity and may have roles in the localization of FGF-2. The original FGF2-G3 (Qk053) corresponds to the 154 aa protein. Furthermore, we wanted to allow scientists to compare directly with their existing FGF-2, so we introduced the nine amino acid substitutions into the 145 aa form of FGF-2 to make Qk052 FGF2-G3 (145 aa).

Paul Burridge and colleagues at Northwestern University, Chicago, published a protocol for B8 media. This media uses thermostable FGF2-G3, along with optimized media component concentration and composition, therefore reducing media cost and facilitating weekend-free pluripotent stem cell culture regimes. B8 media effectively maintains the homogeneity and differentiation potential of human iPSC (Kuo et al. and updated in Lyra-Leite et al.).
The main changes in B8 media compared to the industry-standard E8 media are:

• FGF-2 (used at 100 ng/ml) replaced with FGF2-G3 (used at 5 ng/ml)
• pH reduced to 7-7.1 and osmolarity to 310 osm/L
• the concentration of TGF beta 1 (Qk010) reduced to 2 ng/ml

Koledova and colleagues at Masaryk University explored the receptor specificity of FGF2-G3 (FGF2 STAB®) in detail using a BaF3 cell proliferation assay. FGF2-G3 exhibited the same profile of receptor activation as wild-type FGF-2. FGF2-G3 induced proliferation of cells expressing FGFR1c, FGFR2c, FGFR3c and FGFR1b, but not FGFR2b and FGFR3b. Interestingly, they also reported a lower EC50 value (10-100x lower) for FGF2-G3 compared to the wild-type and suggested the stability increased receptor affinity in this assay.
In the same study, the authors evaluated the impact of exogenous heparin on FGF-2 wild-type and FGF2-G3 bioactivity (EC50) and signalling dynamics. Heparin has a stabilising effect on wild-type FGF-2 and, perhaps consequently, impacts FGF-2 signalling dynamics. FGF2-G3 signalling was less heparin-dependent in ERK1/2 signalling assay in both the FGFR over-expressing cell lines and in primary fibroblast.

Petr Dvorak and Jiri Damborsky, researchers at Masaryk University, published their computer-assisted protein engineering approach. This approach was used to generate forms of FGF-2 with greater thermal and chemical stability, improving the half-life at 37 °C from 10 hours to more than 20 days. In a multi-step process, Dvorak and colleagues identified an optimal set of nine amino acid substitutions that stabilized FGF-2 whilst not affecting the FGF receptor 1 (FGFR1) or FGF receptor 2 (FGFR2) interface. This hyperstable FGF-2 is known as FGF2-G3, or FGF2-STAB® (Dvorak et al., 2017). The biological half-life of FGF2-G3 and previous stabilized FGF-2 iterations (FGF2-G0 and FGF2-G2) were evaluated by phosphorylation of ERK1/2 through FGF-2 receptor signalling in FGF-2 starved human embryonic stem cells. In these assays, FGF-2 was incubated in conditioned hESC media prior to the assay to mimic time in culture. The biological activity of wild-type FGF-2 was <50% after a 10 hour incubation with conditioned media. In contrast, no significant reduction in FGF2-G3 biological activity was observed after up to 20 days of incubation with media at 37 °C. In the same study, FGF2-G3 and wild-type FGF-2 had equivalent ability to maintain hESC pluripotency and expression of pluripotency markers Oct-4 and nanog.

Patent (WO2017089016A1) was filed by Masaryk University and Enantis. The patent describes the FGF2-G3 technology and its potential for regenerative medicine and stem cell culture applications.

E8 media, developed by Chen and colleagues (Chen et al., 2011 and Beers et al., 2012), was adopted as a standard base media for human induced pluripotent stem cell culture. This media provided a xeno-free, chemically defined stem cell medium with FGF-2 and a TGF beta protein supporting maintenance of pluripotency. Many variations of E8 have been optimized for specific iPSC cell lines and applications.

Ludwig and colleagues established a feeder-free media for human embryonic stem cell culture (mTESR). FGF-2 was identified as the most important factor for maintaining hESC pluripotency in feeder-free media. The original publications used a zebrafish FGF-2, although subsequent mTESR formulations used a human FGF-2 protein. An alternative culture media, StemPro, which also used FGF-2 as a core component, was developed in parallel (Wang, 2007). hESC media are reviewed in Lee et al., 2011.

The unusual initiation start sites and multiple isoforms of FGF-2 were characterized. FGF-2 has five alternative initiation sites, four of which are atypical CUG codons and one AUG. The product produced from the AUG site is the 18 kDa protein corresponding to the 154 aa form adopted for use in stem cell culture media. FGF-2 does not contain a signal sequence and is secreted through the membrane using a mechanism dependent on the tertiary structure and exposed thiols. Interestingly, instead of a signal sequence, a 9 aa N-terminal pro-segment is subject to proteolytic cleavage to form a 146 aa mature segment. This 146 aa protein is the form originally purified from tissues.

FGF-2, also known as basic FGF or bFGF, was isolated from several animal tissues as one of two main mitogenic proteins for endothelial cells (acidic and basic FGF). Esch et al. determined the amino acid sequence of FGF-2 from bovine pituitary. This corresponds to the 146 aa form of FGF-2.