Biotechnol Lett (2015) 37:725–731 DOI 10.1007/s10529-014-1701-4
ORIGINAL RESEARCH PAPER
An investigation into the stability of commercial versus MG63-derived hepatocyte growth factor under flow cultivation conditions Giulia Meneghello • Michael P. Storm • Julian B. Chaudhuri • Paul A. De Bank Marianne J. Ellis
•
Received: 14 May 2014 / Accepted: 7 October 2014 / Published online: 22 October 2014 Ó Springer Science+Business Media Dordrecht 2014
Abstract The scale-up of tissue engineering cell culture must ensure that conditions are maintained while also being cost effective. Here we analyse the stability of hepatocyte growth factor (HGF) to investigate whether concentrations change under dynamic conditions, and compare commercial recombinant human HGF as an additive in ‘standard medium’, to HGF secreted by the osteosarcoma cell line MG63 as a ‘preconditioned medium’. After 3 h under flow conditions, HGF in the standard medium degraded to 40 % of its original concentration but HGF in the preconditioned medium remained at 100 %. The concentration of secreted HGF was 10 times greater than the working
G. Meneghello P. A. De Bank Department of Pharmacy and Pharmacology, Centre for Regenerative Medicine, University of Bath, Bath BA2 7AY, UK G. Meneghello M. P. Storm J. B. Chaudhuri M. J. Ellis (&) Department of Chemical Engineering, Centre for Regenerative Medicine, University of Bath, Bath BA2 7AY, UK e-mail: [email protected] G. Meneghello Department of Neuroscience, Janssen Research and Development, Janssen Pharmaceutica, Turnhoutseweg 30, 2340 Beerse, Belgium J. B. Chaudhuri School of Engineering and Informatics, University of Bradford, Bradford BD7 1DP, UK
concentration of commercially-available HGF. Thus HGF within this medium has increased stability; MG63-derived HGF should therefore be investigated as a cost-effective alternative to current lyophilised powders for use in in vitro models. Furthermore, we recommend that those intending to use HGF (or other growth factors) should consider similar stability testing before embarking on experiments with media flow. Keywords Bioreactor Hepatocyte growth factor Liver Perfusion Regenerative medicine Tissue engineering
Introduction The liver is one of the most investigated tissue engineering systems due to the application of in vitro models for toxicology and drug metabolism and pharmacokinetics (DMPK) studies, as a bioartificial organ for temporary support of diseased liver, and for the biomanufacture of regenerative medicine hepatic constructs. The liver is very complex due to its exocrine and endocrine functions and also, due to the variation of function along the sinusoid, known as zonation, all of which are controlled by an array of hormonal, extracellular and O2 concentration cues (Gebhardt 1992). Considerations for in vitro models include the cell type selected, i.e. primary cultures, a cell line or stem cell-derived hepatocytes, and the use
123
726
of co-cultures with one or more nonparenchymal cell types. A suitable environment, (i.e. 2D v 3D culture, matrix material and architecture, fluid dynamics) and operating conditions must also be provided (Davidson et al. 2010, 2012). The media used in vitro is crucial to achieving and maintaining hepatic function. To date there is no single established culture medium for the maintenance of hepatic function. Media recipes can vary between laboratories and there are many commercial brands with different components currently available on the market. Dong et al. (2008) carried out a design of experiments approach to optimize the culture medium recipe for the human hepatoma cell line C3A based on seven different factors, with hepatocyte growth factor (HGF) and oncostatin M being identified as the most effective for improving metabolism, although they also noted that the recipe may need modifying based on the phenotype. HGF is one of the most important factors for liver development and regeneration (Michalopoulos and Zarnegar 1992) and it is routinely used in vitro for the differentiation of stem cells into hepatocytes and for the maintenance of their differentiated state. Hence, if hepatocytes or hepatocyte-like cells are to be derived from stem cells, HGF will be required in the culture medium. Commercial HGF is expensive and large quantities are needed for both stem cell differentiation and the subsequent maintenance of their differentiated state, therefore systems for the continuous and inexpensive production of growth factors would be of benefit to the field. A cheap-to-run cell factory is one option that may meet this need. The human osteosarcoma cell line MG63 spontaneously secretes HGF into cell culture medium without any external stimulation (Taichman et al. 2001). Here, HGF released by MG63 cells was quantified, and ‘preconditioned medium’ was tested in a dynamic environment by circulation through silicone tubing, and compared to a ‘standard medium’, with commercial HGF as an additive, to verify its potential for liver tissue engineering applications.
Materials and methods Cell line MG63 human osteosarcoma cells, obtained from the European collection of cell culture (ECACC), were
123
Biotechnol Lett (2015) 37:725–731
seeded from 10,000 cells ml-1 (5,000 cells cm-2) to 100,000 cells ml-1 (50,000 cells cm-2) in 24-well plates (0.5 ml media/well) and cultured in complete medium prepared with Dulbecco’s modified Eagle’s medium (DMEM) containing 10 % (v/v) fetal calf serum (FCS), 1 % (v/v) sodium pyruvate, 1 % (v/v) non-essential amino acids and 1 % (v/v) antibiotic/ antimycotic solution, and incubated for 3 days at 37 °C and 5 % CO2. After 3 days, the medium was aspirated from the cells and stored at -80 °C for the subsequent analysis of HGF concentration by ELISA. Media The preconditioned medium was prepared by diluting the MG63 aspirate with complete medium to obtain an initial HGF concentration of 2 ng ml-1. The standard medium was prepared using commercial recombinant human HGF (R&D Systems; #294-HG) at an initial concentration of 2 ng ml-1 in DMEM ? 10 % FCS. For the static studies, the two media were placed in separate glass bottles and incubated for 3 h at 37 °C and 5 % CO2. Samples were collected from each bottle every 30 min and stored at -80 °C until analysis. Flow studies For flow studies, the media were circulated inside platinum-cured silicone tubing (internal diam. 2.4 mm, Cole Parmer) by means of a peristaltic pump in a closed recycling system at 37 °C and 5 % CO2 for 3 h at a flow rate of 2.5 ml min-1, Re & 30, sx & 0.02 Pa (this is the shear stress at the wall, i.e. the maximum shear stress experienced), based on properties of water at 37°C. Samples were taken every 30 min from the feed bottle and stored at -80 °C until analysis. As the MG63-derived HGF was produced by a human cell line, it was almost certainly in its native conformation (heterodimer of a 69 kDa a chain and 34 kDa b chain), while the commercial HGF used in this study consisted of a mixture of a disulphide-linked heterodimer and unprocessed single chain (inactive precursor). The quoted molecular weights of the commercial product are 79.7 kDa (single chain precursor), 53.7 kDa (a chain) and 26 kDa (b chain). This reduced mass in comparison to the MG63-derived HGF is likely to reflect altered glycosylation due to the
Biotechnol Lett (2015) 37:725–731
727
insect cell line used to produce the commercial product (Okada et al. 2010; Schiller et al. 2012; Verma et al. 1998). HGF quantification For HGF quantification, samples were thawed at 37 °C and the concentration of HGF present in the samples was quantified using a human HGF Duoset ELISA development system kit (R&D Systems) according to the manufacturer’s instructions. Scattering assay Scattering assay was carried out essentially according to Lai et al. (2012). C3A cells (a HepG2 subclone) were plated on tissue culture plastic at a density of 3,000 cells cm-2 in EMEM supplemented with 10 % (v/v) FCS (Gibco), 2 mM glutamine, penicillin/streptomycin and allowed to grow into colonies for 48 h. Media was changed to control medium (EMEM plus supplements as above with the addition of nonessential amino acids and sodium pyruvate) or control media conditioned for 48 h by MG63 cells and diluted to contain 5 ng ml-1 MG63-secreted HGF. Colonies were scored at the time points indicated for scattering. A colony was considered scattered when 50 % of the cells had lost contact with their neighbouring cells and exhibited a fibroblast-like morphology. Greater than 100 colonies were scored for each condition at each time point. Statistical analysis was carried out using GraphPad Prism software.
Results and discussion The amount of HGF secreted by MG63 cells correlated with an increasing number of cells, although this began to plateau at the highest cell number investigated (Fig. 1). With seeding at 80,000 cells ml-1 (40,000 cells cm-2) and above, the MG63 s became over-confluent by day 3 and the decreased secretion of HGF is likely to result from limited availability of HGF from cells within large clumps. Taichman et al. (2001) found that after four days, an initial cell number of 20,000 cells cm-2 secreted 16 ng HGF ml-1 compared to 18 ng HGF ml-1 secreted from cells seeded at 20,000 cells cm-2 (10,000 cells
Fig. 1 HGF released by MG63 cells seeded at different concentrations 3 days post-seeding. Data represent the mean ± SD (n = 4) and the trendline was fitted with nonlinear regression (within the limits 10,000 \ cell number (cells ml-1) \ 100,000)
ml-1) after 3 days this study; this equates to 0.9 pg HGF cell-1 (9 9 10-4 ng HGF cell-1) in our studies (It should be noted that the HGF cell-1 value cannot be calculated from the Taichman paper). Although the time frames and culture conditions differ between the two experiments the values are of the same order of magnitude suggesting MG63 secretion of HGF is repeatable in different laboratories. The concentrations of secreted HGF are an order of magnitude greater than that required in hepatocyte media. Therefore, large amounts of HGF could be continuously available due to the continual secretion from these cells and the capacity to dilute the preconditioned medium with additives for the hepatocytes. This shows an efficient level of production of HGF by the MG63 cell line and demonstrates that the MG63 system is a viable alternative to commercial suppliers for obtaining HGF. A cost analysis shows that the cheapest HGF is currently $130 (£75) for 10 lg, the product used here being the most expensive at £417 for 10 lg. Based on the amount of media and plasticware needed to produce the same weight of HGF, the MG63-secreted HGF costs around £60 per 10 lg, using T75 flasks which, while simple to use, are not media-volume efficient. This should be seen as an in-house cost to researchers and there is scope to develop a dynamic co-culture system to promote differentiation/maintain function of hepatocytes. Importantly, we have also demonstrated that the MG63-secreted HGF is biologically active. As HGF causes hepatocytes to scatter (Naldini et al. 1991) and,
123
728
Fig. 2 MG63-conditioned medium induces scattering of hepatocytes. a C3A hepatocytes were cultured in conditioned medium containing 5 ng ml-1 HGF (filled square) and cell scatter compared to that in control medium (open square). Data represent the mean ± SD of triplicate samples ([100 colonies per condition). Representative images of C3A colonies following 48 h exposure to control b and conditioned c medium. Scale bars = 100 lm
in order to demonstrate the activity of the secreted HGF, we incubated C3A cells, a HepG2 subclone, in MG63-conditioned medium. As expected, this HGFcontaining medium (with HGF at 5 ng ml-1) induced a marked level of cell scattering in comparison to control medium (Fig. 2). Thus, MG63 cells are capable of generating biologically active HGF in a cost-effective manner. However, the authors recognise there are potential safety issues in using a human tumour cell line to produce a therapy and therefore suggest the use of MG63-derived HGF for in vitro models only. When the standard and preconditioned media were maintained under static conditions the HGF concentration in the standard medium decreased from 100% to 93 ± 1% over 3 h, while no decrease was seen in the HGF concentration in the preconditioned medium (Fig. 3). In comparison, when the standard and preconditioned media were maintained under flow conditions, the HGF concentration in the standard
123
Biotechnol Lett (2015) 37:725–731
medium decreased from 100 % to 35.6 % ± 0.5 % over the 3 h, while no decrease was seen in the HGF concentration in the preconditioned medium (Fig. 4). To our knowledge no studies on the effect of shear on HGF stability have been reported. However, it is recognised that shear stress can cause protein denaturation and agglomeration (Kost and Condreay 1999; Baldauf et al. 2009; Biddlecombe et al. 2007) and may cause irreversible unfolding (Ashton et al. 2009), but this does not explain why such a significant difference was observed in the concentrations of commercial and MG63-secreted HGF following perfusion. One explanation may be the slight differences in the media used to circulate the two forms of HGF. While both media were DMEM with 10 % (v/v) FCS, the standard medium containing commercial HGF lacked antibiotic/antimycotic, non-essential amino acids and any soluble factors, in addition to HGF, secreted by MG63 cells. Although it could be postulated that one or more of these additional components elicited a protective effect on the MG63-derived HGF, this is unlikely, especially considering the magnitude of the difference between protein concentrations. HGF binds to serum albumin (Basilico et al. 2008) but both media contained 10 % (v/v) FCS so this is unlikely to be responsible for any protective effect. MG63 cells secrete fibronectin (Franceschi et al. 1988; Kamihagi et al. 1994) and glycosaminoglycans (Kumarasuriyar et al. 2009), and HGF binds to both (Rahman et al. 2005; Zhu and Clark 2014). Future investigations to explore the protective effects of culture medium could include: (a)
(b) (c)
Depletion of HGF from MG63-conditioned medium and addition of commercial HGF to see if this changes its degradation. Addition of non-essential amino acids and antibiotic/antimycotic to standard medium. Western blots of both types of medium probed with an anti-HGF antibody, potentially identifying bands of different MWs where HGF is bound to another molecule.
Another possible explanation for the differences observed is the HGF itself. The standard medium contained commercial HGF. This is a recombinant human protein expressed in the fall armyworm Spodoptera frugiperda. While this species has the metabolic machinery to perform many of the posttranslational modifications found in mammalian
Biotechnol Lett (2015) 37:725–731
729
Fig. 3 Concentration of HGF under static conditions. a Standard medium, b preconditioned medium. Concentrations are expressed as percentages of the initial concentration of HGF which was 2 ng ml-1. Data represent the mean ± SD of three samples (n = 3)
Fig. 4 Concentration of HGF under flow conditions. a Standard medium, b preconditioned medium secreted by MG63 cells. The media were circulated continuously inside silicone tubing at
2.5 ml min-1. Concentrations are expressed as percentages of the initial concentration of HGF which was 2 ng ml-1. Data represent the mean ± SD of three samples (n = 3)
proteins, it lacks the ability to synthesize the correct Nglycans (Okada et al. 2010; Schiller et al. 2012; Verma et al. 1998); of the five glycosylation sites on HGF, one is an O-glycosylation site and four are N-glycosylation sites (Fukuta et al. 2005). This could explain the lack of stability of this protein under perfusion, as glycosylation has been shown to be an important regulator of protein folding, stability and function. Although Fukuta and coworkers assessed the effect of removing glycans from HGF on stability and activity and found no effect (Fukuta et al. 2005), their study was performed under static conditions. It may be that the differences in glycosylation between the two HGF species are sufficient that the non-human glycans are unable to
prevent destabilization and unfolding under shear, resulting in direct loss of activity or increased absorption to the silicone tubing. Further investigation into how glycosylation of HGF affects stability of the protein could include experiments to enzymatically cleave glycans from MG63-derived HGF as described by Fukuta et al. (2005). Non-parenchymal cells (NPCs) play an important role in liver repair due in part to the secretion of HGF by endothelial cells (Maher 1993; Malik et al. 2002). The use of NPCs for conditioned media is also well established; for example, the use of NPCs (human liver endothelial (TMNK-1), stellate (TWNT-1) and cholangiocyte (MMNK-1) cell lines), plus a deleted variant of HGF (dHGF) has been used as part of the
123
730
Biotechnol Lett (2015) 37:725–731
protocol for in vitro differentiation of mouse embryonic stem cells into hepatocyte-like cells (SotoGutierrez et al. 2007) although the effect of the conditioned media versus the dHGF was not described. A study was carried out on the co-culture of primary hepatocytes and NPCs under static and flow conditions, with and without the addition of HGF and heparin-binding epidermal growth factor-like growth factor (HB-EGF) (Kan et al. 2004). The authors concluded that perfusion, co-culture and growth factors were beneficial to in vitro culture, however the effects of the NPCs providing a conditioned media were not accounted for, nor the stability of the growth factors under flow conditions. Based on the findings presented here it is recommended that the stability of HGF (and other media factors) are assessed in the system prior to cell culture to ensure the true effects of these factors on the cells are being analysed.
Conclusion The ability to produce high levels of bioactive HGF from MG63 cells provides an inexpensive and efficient way to produce this growth factor. We have demonstrated that, in both static and dynamic systems, MG63-derived HGF in preconditioned medium is more stable than commercially available HGF in standard medium and is, therefore, more suitable for bioreactor applications. As proteins are, in general, susceptible to degradation by a number of different mechanisms, these data demonstrate that it is imperative that their fate is thoroughly investigated in order to optimise cell growth and fate within bioreactor culture systems. Acknowledgments This work was supported by a Marie Curie Early Stage Training Fellowship award (Meneghello). The authors would like to thank Dr Alexander Ciupa for his technical assistance. Conflict of interest authors.
No conflicts of interest exist for any of the
References Ashton L, Dusting J, Imomoh E et al (2009) Shear-induced unfolding of lysozyme monitored in situ. Biophys J 96:4231–4236
123
Baldauf C, Schneppenheim R, Stacklies W et al (2009) Shearinduced unfolding activates von willebrand factor a2 domain for proteolysis. J Thromb Haemost 7:2096–2105 Basilico C, Arnesano A, Galluzzo M et al (2008) A high affinity hepatocyte growth factor-binding site in the immunoglobulin-like region of met. J Biol Chem 283:21267–21277 Biddlecombe JG, Craig AV, Zhang H et al (2007) Determining antibody stability: creation of solid-liquid interfacial effects within a high shear environment. Biotechnol Prog 23:1218–1222 Davidson AJ, Ellis MJ, Chaudhuri JB (2010) A theoretical method to improve and optimize the design of bioartificial livers. Biotech Bioeng 106:980–988 Davidson AJ, Ellis MJ, Chaudhuri JB (2012) A theoretical approach to zonation in a bioartificial liver. Biotech Bioeng 109:234–243 Dong J, Mandenius C-F, Luebberstedt M et al (2008) Evaluation and optimization of hepatocyte culture media factors by design of experiments (doe) methodology. Cytotechnology 57:251–261 Franceschi RT, Romano PR, Park KY (1988) Regulation of type-i collagen-synthesis by 1,25-dihydroxyvitamin-d3 in human osteo-sarcoma cells. J Biol Chem 263:18938–18945 Fukuta K, Matsumoto K, Nakamura T (2005) Multiple biological responses are induced by glycosylation-deficient hepatocyte growth factor. Biochem J 388:555–562 Gebhardt R (1992) Metabolic zonation of the liver–regualtion and implications for liver-function. Pharma Therapeut 53:275–354 Kamihagi K, Katayama M, Ouchi R et al (1994) Osteonectin/ sparc regulates cellular secretion rates of fibronectin and laminin extracellular-matrix proteins. Biochem Biophy Res Commun 200:423–428 Kan P, Miyoshi H, Ohshima N (2004) Perfusion of medium with supplemented growth factors changes metabolic activities and cell morphology of hepatocyte-nonparenchymal cell coculture. Tissue Eng 10:1297–1307 Kost TA, Condreay JP (1999) Recombinant baculoviruses as expression vectors for insect and mammalian cells. Curr Opin Biotechnol 10:428–433 Kumarasuriyar A, Murali S, Nurcombe V et al (2009) Glycosaminoglcan composition changes with mg-63 osteosarcoma osteogenesis in vitro and induces human mesenchymal stem cell aggregation. J Cell Physiol 218:501–511 Lai JK, Wu HC, Shen YC, Hsieh HY, Yang SY, Chang CC (2012) Kru¨ppel-like factor 4 is involved in cell scattering induced by hepatocyte growth factor. J Cell Sci 125:4853–4864 Maher JJ (1993) Cell-specific expression of hepatocyte growthfactor in liver-up-regulation in sinusoidal endothelial-cells after carbon-tetrachloride. J Clin Invest 91:2244–2252 Malik R, Selden C, Hodgson H (2002) The role of non-parenchymal cells in liver growth. Semin Cell Dev Biol 13:425–431 Michalopoulos GK, Zarnegar R (1992) Hepatocyte growthfactor. Hepatology 15:149–155 Naldini L, Weidner KM, Vigna E et al (1991) Scatter factor and hepatocyte growth factor are indistinguishable ligands for the MET receptor. EMBO J 10:2867–2878
Biotechnol Lett (2015) 37:725–731 Okada T, Ihara H, Ito R et al (2010) N-Glycosylation engineering of lepidopteran insect cells by the introduction of the beta 1,4-N-acetylglucosaminyltransferase iii gene. Glycobiology 20:1147–1159 Rahman S, Patel Y, Murray J, Patel KV, Sumathipala R, Sobel M, Wijelath ES (2005) Novel hepatocyte growth factor (HGF) binding domains on fibronectin and vitronectin coordinate a distinct and amplified Met-integrin induced signalling pathway in endothelial cells. BMC cell biol 6(1):8 Schiller B, Hykollari A, Yan S et al (2012) Complicated Nlinked glycans in simple organisms. Biol Chem 393:661–673 Soto-Gutierrez A, Navarro-Alvarez N, Zhao D et al (2007) Differentiation of mouse embryonic stem cells to
731 hepatocyte-like cells by co-culture with human liver nonparenchymal cell lines. Nat Protoc 2:347–356 Taichman RS, Reilly MJ, Verma RS et al (2001) Hepatocyte growth factor is secreted by osteoblasts and cooperatively permits the survival of haematopoietic progenitors. Brit J Haematol 112:438–448 Verma R, Boleti E, George AJT (1998) Antibody engineering: comparison of bacterial, yeast, insect and mammalian expression systems. J Immunol Methods 216:165–181 Zhu J, Clark RAF (2014) Fibronectin at select sites binds multiple growth factors and enhances their activity: expansion of the collaborative ecm-gf paradigm. J Invest Dermatol 134:895–901
123
ORIGINAL RESEARCH PAPER
An investigation into the stability of commercial versus MG63-derived hepatocyte growth factor under flow cultivation conditions Giulia Meneghello • Michael P. Storm • Julian B. Chaudhuri • Paul A. De Bank Marianne J. Ellis
•
Received: 14 May 2014 / Accepted: 7 October 2014 / Published online: 22 October 2014 Ó Springer Science+Business Media Dordrecht 2014
Abstract The scale-up of tissue engineering cell culture must ensure that conditions are maintained while also being cost effective. Here we analyse the stability of hepatocyte growth factor (HGF) to investigate whether concentrations change under dynamic conditions, and compare commercial recombinant human HGF as an additive in ‘standard medium’, to HGF secreted by the osteosarcoma cell line MG63 as a ‘preconditioned medium’. After 3 h under flow conditions, HGF in the standard medium degraded to 40 % of its original concentration but HGF in the preconditioned medium remained at 100 %. The concentration of secreted HGF was 10 times greater than the working
G. Meneghello P. A. De Bank Department of Pharmacy and Pharmacology, Centre for Regenerative Medicine, University of Bath, Bath BA2 7AY, UK G. Meneghello M. P. Storm J. B. Chaudhuri M. J. Ellis (&) Department of Chemical Engineering, Centre for Regenerative Medicine, University of Bath, Bath BA2 7AY, UK e-mail: [email protected] G. Meneghello Department of Neuroscience, Janssen Research and Development, Janssen Pharmaceutica, Turnhoutseweg 30, 2340 Beerse, Belgium J. B. Chaudhuri School of Engineering and Informatics, University of Bradford, Bradford BD7 1DP, UK
concentration of commercially-available HGF. Thus HGF within this medium has increased stability; MG63-derived HGF should therefore be investigated as a cost-effective alternative to current lyophilised powders for use in in vitro models. Furthermore, we recommend that those intending to use HGF (or other growth factors) should consider similar stability testing before embarking on experiments with media flow. Keywords Bioreactor Hepatocyte growth factor Liver Perfusion Regenerative medicine Tissue engineering
Introduction The liver is one of the most investigated tissue engineering systems due to the application of in vitro models for toxicology and drug metabolism and pharmacokinetics (DMPK) studies, as a bioartificial organ for temporary support of diseased liver, and for the biomanufacture of regenerative medicine hepatic constructs. The liver is very complex due to its exocrine and endocrine functions and also, due to the variation of function along the sinusoid, known as zonation, all of which are controlled by an array of hormonal, extracellular and O2 concentration cues (Gebhardt 1992). Considerations for in vitro models include the cell type selected, i.e. primary cultures, a cell line or stem cell-derived hepatocytes, and the use
123
726
of co-cultures with one or more nonparenchymal cell types. A suitable environment, (i.e. 2D v 3D culture, matrix material and architecture, fluid dynamics) and operating conditions must also be provided (Davidson et al. 2010, 2012). The media used in vitro is crucial to achieving and maintaining hepatic function. To date there is no single established culture medium for the maintenance of hepatic function. Media recipes can vary between laboratories and there are many commercial brands with different components currently available on the market. Dong et al. (2008) carried out a design of experiments approach to optimize the culture medium recipe for the human hepatoma cell line C3A based on seven different factors, with hepatocyte growth factor (HGF) and oncostatin M being identified as the most effective for improving metabolism, although they also noted that the recipe may need modifying based on the phenotype. HGF is one of the most important factors for liver development and regeneration (Michalopoulos and Zarnegar 1992) and it is routinely used in vitro for the differentiation of stem cells into hepatocytes and for the maintenance of their differentiated state. Hence, if hepatocytes or hepatocyte-like cells are to be derived from stem cells, HGF will be required in the culture medium. Commercial HGF is expensive and large quantities are needed for both stem cell differentiation and the subsequent maintenance of their differentiated state, therefore systems for the continuous and inexpensive production of growth factors would be of benefit to the field. A cheap-to-run cell factory is one option that may meet this need. The human osteosarcoma cell line MG63 spontaneously secretes HGF into cell culture medium without any external stimulation (Taichman et al. 2001). Here, HGF released by MG63 cells was quantified, and ‘preconditioned medium’ was tested in a dynamic environment by circulation through silicone tubing, and compared to a ‘standard medium’, with commercial HGF as an additive, to verify its potential for liver tissue engineering applications.
Materials and methods Cell line MG63 human osteosarcoma cells, obtained from the European collection of cell culture (ECACC), were
123
Biotechnol Lett (2015) 37:725–731
seeded from 10,000 cells ml-1 (5,000 cells cm-2) to 100,000 cells ml-1 (50,000 cells cm-2) in 24-well plates (0.5 ml media/well) and cultured in complete medium prepared with Dulbecco’s modified Eagle’s medium (DMEM) containing 10 % (v/v) fetal calf serum (FCS), 1 % (v/v) sodium pyruvate, 1 % (v/v) non-essential amino acids and 1 % (v/v) antibiotic/ antimycotic solution, and incubated for 3 days at 37 °C and 5 % CO2. After 3 days, the medium was aspirated from the cells and stored at -80 °C for the subsequent analysis of HGF concentration by ELISA. Media The preconditioned medium was prepared by diluting the MG63 aspirate with complete medium to obtain an initial HGF concentration of 2 ng ml-1. The standard medium was prepared using commercial recombinant human HGF (R&D Systems; #294-HG) at an initial concentration of 2 ng ml-1 in DMEM ? 10 % FCS. For the static studies, the two media were placed in separate glass bottles and incubated for 3 h at 37 °C and 5 % CO2. Samples were collected from each bottle every 30 min and stored at -80 °C until analysis. Flow studies For flow studies, the media were circulated inside platinum-cured silicone tubing (internal diam. 2.4 mm, Cole Parmer) by means of a peristaltic pump in a closed recycling system at 37 °C and 5 % CO2 for 3 h at a flow rate of 2.5 ml min-1, Re & 30, sx & 0.02 Pa (this is the shear stress at the wall, i.e. the maximum shear stress experienced), based on properties of water at 37°C. Samples were taken every 30 min from the feed bottle and stored at -80 °C until analysis. As the MG63-derived HGF was produced by a human cell line, it was almost certainly in its native conformation (heterodimer of a 69 kDa a chain and 34 kDa b chain), while the commercial HGF used in this study consisted of a mixture of a disulphide-linked heterodimer and unprocessed single chain (inactive precursor). The quoted molecular weights of the commercial product are 79.7 kDa (single chain precursor), 53.7 kDa (a chain) and 26 kDa (b chain). This reduced mass in comparison to the MG63-derived HGF is likely to reflect altered glycosylation due to the
Biotechnol Lett (2015) 37:725–731
727
insect cell line used to produce the commercial product (Okada et al. 2010; Schiller et al. 2012; Verma et al. 1998). HGF quantification For HGF quantification, samples were thawed at 37 °C and the concentration of HGF present in the samples was quantified using a human HGF Duoset ELISA development system kit (R&D Systems) according to the manufacturer’s instructions. Scattering assay Scattering assay was carried out essentially according to Lai et al. (2012). C3A cells (a HepG2 subclone) were plated on tissue culture plastic at a density of 3,000 cells cm-2 in EMEM supplemented with 10 % (v/v) FCS (Gibco), 2 mM glutamine, penicillin/streptomycin and allowed to grow into colonies for 48 h. Media was changed to control medium (EMEM plus supplements as above with the addition of nonessential amino acids and sodium pyruvate) or control media conditioned for 48 h by MG63 cells and diluted to contain 5 ng ml-1 MG63-secreted HGF. Colonies were scored at the time points indicated for scattering. A colony was considered scattered when 50 % of the cells had lost contact with their neighbouring cells and exhibited a fibroblast-like morphology. Greater than 100 colonies were scored for each condition at each time point. Statistical analysis was carried out using GraphPad Prism software.
Results and discussion The amount of HGF secreted by MG63 cells correlated with an increasing number of cells, although this began to plateau at the highest cell number investigated (Fig. 1). With seeding at 80,000 cells ml-1 (40,000 cells cm-2) and above, the MG63 s became over-confluent by day 3 and the decreased secretion of HGF is likely to result from limited availability of HGF from cells within large clumps. Taichman et al. (2001) found that after four days, an initial cell number of 20,000 cells cm-2 secreted 16 ng HGF ml-1 compared to 18 ng HGF ml-1 secreted from cells seeded at 20,000 cells cm-2 (10,000 cells
Fig. 1 HGF released by MG63 cells seeded at different concentrations 3 days post-seeding. Data represent the mean ± SD (n = 4) and the trendline was fitted with nonlinear regression (within the limits 10,000 \ cell number (cells ml-1) \ 100,000)
ml-1) after 3 days this study; this equates to 0.9 pg HGF cell-1 (9 9 10-4 ng HGF cell-1) in our studies (It should be noted that the HGF cell-1 value cannot be calculated from the Taichman paper). Although the time frames and culture conditions differ between the two experiments the values are of the same order of magnitude suggesting MG63 secretion of HGF is repeatable in different laboratories. The concentrations of secreted HGF are an order of magnitude greater than that required in hepatocyte media. Therefore, large amounts of HGF could be continuously available due to the continual secretion from these cells and the capacity to dilute the preconditioned medium with additives for the hepatocytes. This shows an efficient level of production of HGF by the MG63 cell line and demonstrates that the MG63 system is a viable alternative to commercial suppliers for obtaining HGF. A cost analysis shows that the cheapest HGF is currently $130 (£75) for 10 lg, the product used here being the most expensive at £417 for 10 lg. Based on the amount of media and plasticware needed to produce the same weight of HGF, the MG63-secreted HGF costs around £60 per 10 lg, using T75 flasks which, while simple to use, are not media-volume efficient. This should be seen as an in-house cost to researchers and there is scope to develop a dynamic co-culture system to promote differentiation/maintain function of hepatocytes. Importantly, we have also demonstrated that the MG63-secreted HGF is biologically active. As HGF causes hepatocytes to scatter (Naldini et al. 1991) and,
123
728
Fig. 2 MG63-conditioned medium induces scattering of hepatocytes. a C3A hepatocytes were cultured in conditioned medium containing 5 ng ml-1 HGF (filled square) and cell scatter compared to that in control medium (open square). Data represent the mean ± SD of triplicate samples ([100 colonies per condition). Representative images of C3A colonies following 48 h exposure to control b and conditioned c medium. Scale bars = 100 lm
in order to demonstrate the activity of the secreted HGF, we incubated C3A cells, a HepG2 subclone, in MG63-conditioned medium. As expected, this HGFcontaining medium (with HGF at 5 ng ml-1) induced a marked level of cell scattering in comparison to control medium (Fig. 2). Thus, MG63 cells are capable of generating biologically active HGF in a cost-effective manner. However, the authors recognise there are potential safety issues in using a human tumour cell line to produce a therapy and therefore suggest the use of MG63-derived HGF for in vitro models only. When the standard and preconditioned media were maintained under static conditions the HGF concentration in the standard medium decreased from 100% to 93 ± 1% over 3 h, while no decrease was seen in the HGF concentration in the preconditioned medium (Fig. 3). In comparison, when the standard and preconditioned media were maintained under flow conditions, the HGF concentration in the standard
123
Biotechnol Lett (2015) 37:725–731
medium decreased from 100 % to 35.6 % ± 0.5 % over the 3 h, while no decrease was seen in the HGF concentration in the preconditioned medium (Fig. 4). To our knowledge no studies on the effect of shear on HGF stability have been reported. However, it is recognised that shear stress can cause protein denaturation and agglomeration (Kost and Condreay 1999; Baldauf et al. 2009; Biddlecombe et al. 2007) and may cause irreversible unfolding (Ashton et al. 2009), but this does not explain why such a significant difference was observed in the concentrations of commercial and MG63-secreted HGF following perfusion. One explanation may be the slight differences in the media used to circulate the two forms of HGF. While both media were DMEM with 10 % (v/v) FCS, the standard medium containing commercial HGF lacked antibiotic/antimycotic, non-essential amino acids and any soluble factors, in addition to HGF, secreted by MG63 cells. Although it could be postulated that one or more of these additional components elicited a protective effect on the MG63-derived HGF, this is unlikely, especially considering the magnitude of the difference between protein concentrations. HGF binds to serum albumin (Basilico et al. 2008) but both media contained 10 % (v/v) FCS so this is unlikely to be responsible for any protective effect. MG63 cells secrete fibronectin (Franceschi et al. 1988; Kamihagi et al. 1994) and glycosaminoglycans (Kumarasuriyar et al. 2009), and HGF binds to both (Rahman et al. 2005; Zhu and Clark 2014). Future investigations to explore the protective effects of culture medium could include: (a)
(b) (c)
Depletion of HGF from MG63-conditioned medium and addition of commercial HGF to see if this changes its degradation. Addition of non-essential amino acids and antibiotic/antimycotic to standard medium. Western blots of both types of medium probed with an anti-HGF antibody, potentially identifying bands of different MWs where HGF is bound to another molecule.
Another possible explanation for the differences observed is the HGF itself. The standard medium contained commercial HGF. This is a recombinant human protein expressed in the fall armyworm Spodoptera frugiperda. While this species has the metabolic machinery to perform many of the posttranslational modifications found in mammalian
Biotechnol Lett (2015) 37:725–731
729
Fig. 3 Concentration of HGF under static conditions. a Standard medium, b preconditioned medium. Concentrations are expressed as percentages of the initial concentration of HGF which was 2 ng ml-1. Data represent the mean ± SD of three samples (n = 3)
Fig. 4 Concentration of HGF under flow conditions. a Standard medium, b preconditioned medium secreted by MG63 cells. The media were circulated continuously inside silicone tubing at
2.5 ml min-1. Concentrations are expressed as percentages of the initial concentration of HGF which was 2 ng ml-1. Data represent the mean ± SD of three samples (n = 3)
proteins, it lacks the ability to synthesize the correct Nglycans (Okada et al. 2010; Schiller et al. 2012; Verma et al. 1998); of the five glycosylation sites on HGF, one is an O-glycosylation site and four are N-glycosylation sites (Fukuta et al. 2005). This could explain the lack of stability of this protein under perfusion, as glycosylation has been shown to be an important regulator of protein folding, stability and function. Although Fukuta and coworkers assessed the effect of removing glycans from HGF on stability and activity and found no effect (Fukuta et al. 2005), their study was performed under static conditions. It may be that the differences in glycosylation between the two HGF species are sufficient that the non-human glycans are unable to
prevent destabilization and unfolding under shear, resulting in direct loss of activity or increased absorption to the silicone tubing. Further investigation into how glycosylation of HGF affects stability of the protein could include experiments to enzymatically cleave glycans from MG63-derived HGF as described by Fukuta et al. (2005). Non-parenchymal cells (NPCs) play an important role in liver repair due in part to the secretion of HGF by endothelial cells (Maher 1993; Malik et al. 2002). The use of NPCs for conditioned media is also well established; for example, the use of NPCs (human liver endothelial (TMNK-1), stellate (TWNT-1) and cholangiocyte (MMNK-1) cell lines), plus a deleted variant of HGF (dHGF) has been used as part of the
123
730
Biotechnol Lett (2015) 37:725–731
protocol for in vitro differentiation of mouse embryonic stem cells into hepatocyte-like cells (SotoGutierrez et al. 2007) although the effect of the conditioned media versus the dHGF was not described. A study was carried out on the co-culture of primary hepatocytes and NPCs under static and flow conditions, with and without the addition of HGF and heparin-binding epidermal growth factor-like growth factor (HB-EGF) (Kan et al. 2004). The authors concluded that perfusion, co-culture and growth factors were beneficial to in vitro culture, however the effects of the NPCs providing a conditioned media were not accounted for, nor the stability of the growth factors under flow conditions. Based on the findings presented here it is recommended that the stability of HGF (and other media factors) are assessed in the system prior to cell culture to ensure the true effects of these factors on the cells are being analysed.
Conclusion The ability to produce high levels of bioactive HGF from MG63 cells provides an inexpensive and efficient way to produce this growth factor. We have demonstrated that, in both static and dynamic systems, MG63-derived HGF in preconditioned medium is more stable than commercially available HGF in standard medium and is, therefore, more suitable for bioreactor applications. As proteins are, in general, susceptible to degradation by a number of different mechanisms, these data demonstrate that it is imperative that their fate is thoroughly investigated in order to optimise cell growth and fate within bioreactor culture systems. Acknowledgments This work was supported by a Marie Curie Early Stage Training Fellowship award (Meneghello). The authors would like to thank Dr Alexander Ciupa for his technical assistance. Conflict of interest authors.
No conflicts of interest exist for any of the
References Ashton L, Dusting J, Imomoh E et al (2009) Shear-induced unfolding of lysozyme monitored in situ. Biophys J 96:4231–4236
123
Baldauf C, Schneppenheim R, Stacklies W et al (2009) Shearinduced unfolding activates von willebrand factor a2 domain for proteolysis. J Thromb Haemost 7:2096–2105 Basilico C, Arnesano A, Galluzzo M et al (2008) A high affinity hepatocyte growth factor-binding site in the immunoglobulin-like region of met. J Biol Chem 283:21267–21277 Biddlecombe JG, Craig AV, Zhang H et al (2007) Determining antibody stability: creation of solid-liquid interfacial effects within a high shear environment. Biotechnol Prog 23:1218–1222 Davidson AJ, Ellis MJ, Chaudhuri JB (2010) A theoretical method to improve and optimize the design of bioartificial livers. Biotech Bioeng 106:980–988 Davidson AJ, Ellis MJ, Chaudhuri JB (2012) A theoretical approach to zonation in a bioartificial liver. Biotech Bioeng 109:234–243 Dong J, Mandenius C-F, Luebberstedt M et al (2008) Evaluation and optimization of hepatocyte culture media factors by design of experiments (doe) methodology. Cytotechnology 57:251–261 Franceschi RT, Romano PR, Park KY (1988) Regulation of type-i collagen-synthesis by 1,25-dihydroxyvitamin-d3 in human osteo-sarcoma cells. J Biol Chem 263:18938–18945 Fukuta K, Matsumoto K, Nakamura T (2005) Multiple biological responses are induced by glycosylation-deficient hepatocyte growth factor. Biochem J 388:555–562 Gebhardt R (1992) Metabolic zonation of the liver–regualtion and implications for liver-function. Pharma Therapeut 53:275–354 Kamihagi K, Katayama M, Ouchi R et al (1994) Osteonectin/ sparc regulates cellular secretion rates of fibronectin and laminin extracellular-matrix proteins. Biochem Biophy Res Commun 200:423–428 Kan P, Miyoshi H, Ohshima N (2004) Perfusion of medium with supplemented growth factors changes metabolic activities and cell morphology of hepatocyte-nonparenchymal cell coculture. Tissue Eng 10:1297–1307 Kost TA, Condreay JP (1999) Recombinant baculoviruses as expression vectors for insect and mammalian cells. Curr Opin Biotechnol 10:428–433 Kumarasuriyar A, Murali S, Nurcombe V et al (2009) Glycosaminoglcan composition changes with mg-63 osteosarcoma osteogenesis in vitro and induces human mesenchymal stem cell aggregation. J Cell Physiol 218:501–511 Lai JK, Wu HC, Shen YC, Hsieh HY, Yang SY, Chang CC (2012) Kru¨ppel-like factor 4 is involved in cell scattering induced by hepatocyte growth factor. J Cell Sci 125:4853–4864 Maher JJ (1993) Cell-specific expression of hepatocyte growthfactor in liver-up-regulation in sinusoidal endothelial-cells after carbon-tetrachloride. J Clin Invest 91:2244–2252 Malik R, Selden C, Hodgson H (2002) The role of non-parenchymal cells in liver growth. Semin Cell Dev Biol 13:425–431 Michalopoulos GK, Zarnegar R (1992) Hepatocyte growthfactor. Hepatology 15:149–155 Naldini L, Weidner KM, Vigna E et al (1991) Scatter factor and hepatocyte growth factor are indistinguishable ligands for the MET receptor. EMBO J 10:2867–2878
Biotechnol Lett (2015) 37:725–731 Okada T, Ihara H, Ito R et al (2010) N-Glycosylation engineering of lepidopteran insect cells by the introduction of the beta 1,4-N-acetylglucosaminyltransferase iii gene. Glycobiology 20:1147–1159 Rahman S, Patel Y, Murray J, Patel KV, Sumathipala R, Sobel M, Wijelath ES (2005) Novel hepatocyte growth factor (HGF) binding domains on fibronectin and vitronectin coordinate a distinct and amplified Met-integrin induced signalling pathway in endothelial cells. BMC cell biol 6(1):8 Schiller B, Hykollari A, Yan S et al (2012) Complicated Nlinked glycans in simple organisms. Biol Chem 393:661–673 Soto-Gutierrez A, Navarro-Alvarez N, Zhao D et al (2007) Differentiation of mouse embryonic stem cells to
731 hepatocyte-like cells by co-culture with human liver nonparenchymal cell lines. Nat Protoc 2:347–356 Taichman RS, Reilly MJ, Verma RS et al (2001) Hepatocyte growth factor is secreted by osteoblasts and cooperatively permits the survival of haematopoietic progenitors. Brit J Haematol 112:438–448 Verma R, Boleti E, George AJT (1998) Antibody engineering: comparison of bacterial, yeast, insect and mammalian expression systems. J Immunol Methods 216:165–181 Zhu J, Clark RAF (2014) Fibronectin at select sites binds multiple growth factors and enhances their activity: expansion of the collaborative ecm-gf paradigm. J Invest Dermatol 134:895–901
123
