A systematic approach towards long-term, serum-free cultivation of fish cells with the RTgill-W1 cell line as example

Highlights

• •First systematically developed serum-free culture medium for RTgill-W1 cell line.
• •Novel 96-well plate assay enabled screening/optimization of medium components.
• •Cells transitioned to serum-free medium and sub-cultured for 35 passages to date.
• •Successful cryopreservation/thawing multiple times.
• •Response to positive control according to OECD test guideline 249 as wildtype cells.

Abstract

Continuous cultures of vertebrate cells, referred to as cell lines, comprise invaluable biological material for a seemingly unlimited range of applications: from basic cell biology to toxicology to biotechnology, to name a few. Cell lines of fish, such as the cold-water fish rainbow trout (Oncorhynchus mykiss), are no exception but are still little tapped into compared to the widely applied mammalian cell lines. Yet, fish cell lines hold promises as alternatives to animal use in, e.g., toxicology and cell-based food production. However, while the number of fish cell lines is growing, a severe bottleneck for their defined and ethical use is lack of a fully described culture medium that supports long-term viability and proliferation, prerequisites to exploit them. More specifically, routine culture of these cells still depends on fetal bovine serum, an undefined, animal-derived component. Here, we report on the first systematically developed, fully defined medium for long-term culture of fish cells, using the rainbow trout gill cell line, RTgill-W1, as example. We first identified components that are apparently provided by serum. A purposefully designed short-term proliferation assay then allowed the screening of the serum replacement components regarding cell viability and proliferation support. Cells were then taken into long-term culture using the designed mix – yielding 35 passages in the fully defined medium thus far. They can be cryopreserved/thawed in this medium; and for the functions tested, they resemble their parent strain. We thus provide impetus to move fish cell line culture to fully defined media for basic research and application.

Graphical abstract

1. Introduction

Our ability to culture vertebrate cells outside of organisms, i.e. in vitro, dates back >100 years (Roberts and Mather, 1998). However, it was only when such cells could be more permanently cultured, allowing them to be viable and proliferate in the artificial environment provided to them, that the tremendous value of using in vitro vertebrate cell cultures for exploring the fundamentals of cell biology and for inspiring innovations and new applications became apparent. Specific examples include the use of cell lines as alternatives to animal testing in chemical risk assessment (Knight et al., 2021) and in the production of cell-based food (Edelman et al., 2005). While the main focus has been on cell lines derived from mammals, cell lines derived from fish alike offer tremendous value and a wide, fish-specific application range. Indeed, procedures for vertebrate cell cultures are generally similar between cell types but culture conditions need to obey species-specific requirements. For example, culture temperature should lie in the optimal range of the organism from which the cell line was originally derived.

The first cell line derived from a fish was the rainbow trout (Oncorhynchus mykiss) gonadal cell line, RTG-2, established by Wolf and Quimby with the intention of providing a research tool to study fish viruses (Wolf and Quimby, 1962). To date, the Cellosaurus returns about 944 hits for fish cell lines, stemming from 210 species of fish with representation from various organs, such as gill (44 hits), intestine (5), liver (55), brain (80). While this repertoire of fish cell lines is still small compared to that of mammals and considering that an estimated 30′000 fish species inhabit the earth (Nelson et al., 2016), it holds tremendous potential to support measures to preserve fish as an ecological and economically valuable organism group. Fish cells reflect the properties of fish in terms of, for example, genetic make-up and nutritional requirements and can uniquely be derived from healthy tissue, i.e. without being cancerous or needing genetic manipulation (Bols et al., 2005, 2017; Schirmer, 2006). Fish cell lines originating from rainbow trout form the largest fish “invitrome” with about 74 cell lines (https://www.cellosaurus.org/; last access: 06.03.2024).

However, prolonged cultivation of most vertebrate cell lines, be they mammalian or fish, still requires the use of undefined animal sera. For example, the rainbow trout cell lines can usually conveniently be cultured at 18–22 °C in free gas exchange with air in the basal medium L-15 developed by Leibovitz (Leibovitz, 1963) but require 5–15 % serum supplementation for long-term cell viability and proliferation, with fetal bovine serum (FBS, referring to serum from an unborn cow, frequently also termed Fetal Calf Serum - FCS) being the common serum of choice (Bols et al., 2017).

The use of FBS in cell culture is fraught with scientific, ethical, and practical challenges (van der Valk et al., 2018; van der Valk, 2022). Scientifically, the largely unknown and variable composition of FBS, including batch-to-batch variability, introduces an unaccountable element into experiments, jeopardizing mechanistic understanding and reproducibility (Liu et al., 2023). For example, inconsistencies in FBS can influence gene expression and cellular phenotype (Wei et al., 2016; Zhao and Klimecki, 2015). Ethically, the collection of FBS raises significant concerns due to the potential for fetal distress and the suffering of unborn calves during blood collection (Versteegen et al., 2021). Furthermore, the reliance on FBS contributes to unpredictable shortages in supply, complicating research and production schedules (van der Valk et al., 2018).

A first step towards serum-free (FBS-free) media was pioneering efforts in developing basal media — a set of nutrient supplements supporting maintenance of cells in vitro — consisting of defined and essential components that allow for standardization in cell culture research (Yao and Asayama, 2017). This innovation was primarily driven by the urgent demand from the biopharmaceutical industry, particularly for cell lines such as the Chinese hamster ovary (CHO), where serum components hindered recombinant protein production efficiency (Ritacco et al., 2018). However, in the majority of the cases, the proprietary nature of specific media compositions poses challenges in knowledge transfer to other cell lines and limits broader applicability and dissemination of optimized serum-free culture methodologies (van der Valk et al., 2010). The 3Rs-Centre Utrecht Life Sciences at Utrecht University maintains an FBS-free database (https://fcs-free.sites.uu.nl/) with over 400 entries for animal component-free, fully defined media using mammalian cells, though many of the formulations are proprietary (Van Der Valk and Gstraunthaler, 2017). There is not a single entry for a purposefully designed, fully transparent serum-free medium for fish. Yue et al. (2016) attempted to culture the rainbow trout gill, RTgill-W1, cell line in two proprietary serum-free media (Turbo Doma and InVitrus VP-6), previously suggested for fish cell lines by Ackermann and Fent (1998). Only one medium allowed passaging of the cells though never more than passage 10 – this was the InVitrus VP-6 with unknown nature of the medium composition being a major caveat (Yue et al., 2016). A notable exception is in the research field of derivation and expansion of human induced pluripotent stem cells and human embryonic stem cells, which are vital for applications like regenerative medicine and drug discovery and where efforts have been made over the past decade to disclose transparently serum-free media formulations (Akopian et al., 2010). Altogether, this led to the development of the Essential 8 medium, a refinement of the previously developed TeSR1 medium. Both formulations are being openly disclosed (Chen et al., 2011).

We here set out to purposefully develop the first serum-free medium for fish cell lines. We focus on the RTgill-W1 rainbow trout gill cell line (Bols et al., 1994), an important alternative to traditional tests for environmental risk assessments using fish. Cell population growth of this cell line has been demonstrated to inform about the ability of chemicals to reduce fish growth (Stadnicka-Michalak et al., 2015). In addition, short-term chemical exposures of RTgill-W1-derived epithelia effectively predict the acute toxicity of chemicals traditionally measured in juvenile or adult fish according to OECD test guideline (TG) 203, the still most widely applied, highest severity vertebrate test in chemical environmental risk assessment. Indeed, the RTgill-W1 cell line is at the core of the first cell-based OECD test guideline (OECD TG 249: Fish Cell Line Acute Toxicity - The RTgill-W1 cell line assay) and ISO standard (ISO 21,115: Water quality - Determination of acute toxicity of water samples and chemicals to a fish gill cell line (RTgill-W1) for environmental protection. It eliminates the need for fish, but, whilst chemical exposure of the cells is performed in a specifically designed medium, L-15/ex, which is completely animal component-free, long-term culture of the RTgill-W1 cells still requires at least 5 % FBS. Hence, if a serum-free medium can be obtained, not only the RTgill-W1 cell line assay as such, but also its routine culture could be performed entirely free of animal sacrifice.

2. Methods

2.1. Media and reagents

All supplements were of analytical reagent quality and suitable for cell culture. A detailed list of supplements tested for the serum-free formulation can be found in Table S2. Aqueous soluble supplements were prepared with either filter-sterilized water or L-15, according to manufacturers’ instructions. For the serum-free mixture testing and optimization, some of the individual supplements were prepared as concentrated stock solutions. Trace element stock solution (TE, 1000x) was freshly prepared before each addition to the serum-free medium by dissolving the pre-measured corresponding weight of each chemical salt with sterile water. The solution was sterilized by syringe filtration through a 0.22 µm filter (ST16541-Q, Sartorius), aliquoted for single use and stored at −20 °C.

2.2. Routine culture of the RTgill-W1 cell line

The commercially available rainbow trout gill cell line RTgill-W1 (ATCC, reference: ATCC CRL-2523, Cellosaurus database, reference: RRID:CVCL_6441) was used in this study. Culture of RTgill-W1 cells was performed under a laminar flow bench using sterile materials and solutions. Cells were maintained in 75 cm² cell culture flasks in 10 mL L-15 medium supplemented with 5 % FBS (v/v) at 19 ± 1 °C under regular atmospheric conditions (referred to as standard conditions). For exposure to serum-free media, confluent cells were washed twice with Versene® (Invitrogen) followed by detachment with trypsin (0.25 % in phosphate-buffer saline without calcium and magnesium; Biowest) and resuspended in the L-15/FBS medium, applying a cell splitting ratio of 1:2, every 2 weeks. Cells obtained in suspension were counted via Casy TTC Cell Counter & Analyser System according to the manufacturer's instructions (Schärfe System). Gill cells with different passage numbers (from P-57 to P-87) were used for the experiments. Mycoplasma tests (MycoAlert PLUS Mycoplasma Detection Kit, Lonza) were carried out six times during this study, showing no signs of contamination.

2.3. 96-well plate short-term assay for medium optimization

Plate layout: Layout was designed to include both control and experimental conditions as follows: The outer wells of the plate were left unseeded to serve as no cell controls and were filled with culture medium ensuring that any background signal could be accounted for. The inner wells were utilized for specific assays, with rows B, C, and D designated for measuring cell viability and rows E, F, and G for cell counting. Each column from 2 to 11 included six technical replicates for each condition: a serum-free media mixture, a positive proliferation control (5 % FBS), and a no proliferation control (L-15 medium).

Optimization of cell number and duration of short-term proliferation assay: To determine optimal cell seeding density and assay duration, cells were seeded in L-15/FBS with an initial density of 2 500, 5 000, 7 500, 10 000, and 12 500 cells in 200 µL per well in 96-well culture plates (665,180, Greiner). After 24 h of attachment, the cells were washed once with L-15 before either new L-15 or L-15/FBS was added, while the outer wells were filled only with 200 μL of L-15. Time points were initially set at 1, 2, 5 and 6 days after medium exchange for all cell seeding densities and further supplemented with day 4 and day 7 for 5 000, 7 500 and 10 000 cells/well as starting density. For each time point, three wells per seeding density were used to assess cell viability, while the remaining three wells of the same plate were utilized for cell counting to determine proliferation rates as described in the section „Analysis of cell viability and cell number”. The optimal starting density was determined to be 7 500 cells/well per 200 µL media and was used for all further experimentation.

Testing serum-free formulations: Using the determined starting density of 37 500 cells/mL (cell density ∼ 24 000 cell/cm²) cells were seeded in μclear® black 96-well culture plates specially purposed for bottom imaging (6,555,097, Greiner) using the plate layout described above. After 24 h of attachment, the cells were washed once with L-15, followed by exposure to 200 µL of serum-free media mixtures (time point Day 0). Cells were subsequently returned to the incubator at standard conditions for 6 days. For each time point, the outcomes were quantified by the established short-term cell proliferation assay using fluorescence-based cell viability assessment, cell count and morphology. Notably, no media exchange was performed throughout the assay duration.

2.4. Adaptation to serum-free media

A 5 mL of cell suspension in L-15/FBS was seeded into an uncoated 25-cm² cell culture flask (cell density ∼ 20 000 cell/cm², approx. 102 000 cell/mL). After 24 h of attachment, cell media was exchanged to the serum-free medium by either direct transfer (100 % serum-free medium, cell referred to as “Sink-Or-Swim” - SOS) or gradual reduction of the FBS content by diluting with serum-free medium (2.5 %, 1 %, 0.5 % and 0 % FBS until 100 % serum-free) during a 20-day period (cells referred as “weaned” - W). The first attempt at cell passaging was done when 0 % FBS and cell confluency were reached. The splitting ratio of cells was 1:2 with each flask split into two new ones. From this point on, cells were maintained in 25 cm² cell culture flasks in 5 mL of freshly prepared serum-free medium, and the medium was changed routinely every 7 days. Whenever at least 70–80 % confluency was reached, cells were passaged by using Accutase® (Sigma, SCR005) for the first ten passages and then by using TrypZean® (Sigma, T3449). In the case of the latter, the trypsinization reaction was stopped with the addition of a defined Trypsin inhibitor from Glycine max (Sigma, T6414), whereas Accutase® did not require a stopping reagent.

2.5. Analysis of cell viability and cell number

The previously described cell viability assay, using a combination of the two fluorescent indicator dyes alamarBlue™ (Invitrogen) and 5-Carboxyfluorescein Diacetate, Acetoxymethyl Ester (CFDA-AM, Invitrogen), was applied after modification for use in 96 well-plates (Dayeh et al., 2005). In short, exposure was terminated by removing exposure media via a multichannel pipette followed by washing the cells once with L-15 (1x) and adding 100 µL per well of indicator dyes working solution. Fluorescence was measured on a Tecan Infinite M200 fluorescence multi well plate reader, first for alamarBlue™ (excitation 530 nm, emission 593 nm), then CFDA-AM (excitation 493 nm, emission 541 nm) after 30 min incubation in the rows B, C and D. Reading was performed from the bottom of the plate. The gain was manually set at 100 with 15 flashes, 20 µs integration time, 0 lag time, and 0 settle time for both dyes.

Cell number was evaluated using the direct whole well cell counting method where cell DNA was labelled using Hoechst 33342 dye (16.23 mM, H3570, Sigma). To initiate this step, exposure was terminated by removing the media, followed by washing the cells once with L-15 (1x) before the addition of a 100 µl staining solution in rows E, F and G. Plates were protected from light and incubated for 30 min. Cells were then fixed by paraformaldehyde solution (4 % in PBS, Sigma). After an additional 10 min incubation, wells were rinsed three times with L-15. Samples were acquired by bottom imaging using the Biotek Cytation 5 (Agilent). Beforehand, inspection of each control was performed via manual microscopy mode using phase contrast and Hoechst 33342 channel (excitation 365 nm, emission 477 nm). The standardized settings for the Hoechst 33342 channel are LED power 9, integration time 65 ms, and gain of 2. These settings were then used to automatically take overlapping montage images of the whole well using the 10x wide field of view objective. Autofocus was achieved via laser autofocus with manual reference images. In total, 36 automatically determined overlapping images were taken in the Hoechst 33342 channel for each whole well image. Images were preprocessed using the Gen5 software to construct the final whole well images. The edges of the wells were cropped. Each image was then treated with the post-imaging procedure “preprocessing” using the following settings: dark background, auto flattening, and an image smoothing strength of 0 cycles of a 3 × 3 filter. The resulting individual images were treated with the “stitching” post-imaging procedure using the following settings: linear blend stitching and crop to remove black rectangles or borders. Acquired images were downsized to 75 % and exported as color pictures in *.png or *.jpg format. The final complete image was used for primary mask cellular analysis by Gen5 software to automatically place object masks around nuclei in each image. For cellular analysis, object counting analysis was performed on the Hoechst 33342 channel to highlight each nuclei.

2.6. Cell proliferation and doubling time

To construct the cell proliferation curves, cells were initially seeded into 96-well plates at an initial density of 10 000 cells per well following the 96-well plate layout. Cells were seeded in six technical replicates either in serum-free media (RTgill-W1-sf W adapted) or 5 % FBS (RTgill-W1). At sampling times, rows B, C, and D were assessed for cell viability using a cell viability assay (alamarBlue™) as previously described. Rows E, F, and G were used for cell counting using the cell number method described above. The population doubling time (DT) was calculated utilizing Eq. (1)

$$DT=tx\frac{ln2}{\ln Xe/Xb}$$ (1)

where t is the incubation time, Xb is the cell number at the beginning of the incubation time, and Xe is the cell number at the end of the incubation time. Cell proliferation curves were subsequently constructed using the cell count data obtained at the specified time points.

2.7. Cryopreservation of cells in serum-free medium

Upon reaching confluence, cells were harvested by detachment (see section Culture of the RTgill-W1 cell line) and with each portion derived from a 1 × 25 cm² flask corresponding to a cryovial (∼ 1 × 10⁶ cell/mL). An aliquot was taken for the measurement of cell viability and cell number using Casy TTC Cell Counter & Analyser System (see section Cell culture). The remaining cell suspension was centrifuged (1000 x g for 3 min), the resulting cell pellet re-suspended in 1 mL of ice-cold freezing medium, composed of 90 % serum-free medium and 10 % dimethyl sulfoxide (DMSO), and subsequently transferred to a Nunc cryo-vial (Fisher Scientific, Wohlen, Switzerland). These samples were maintained on ice for a duration of 10 min, followed by storage at −80 °C. Subsequently, after storage at −80 °C for a minimum of 7 days, the samples were transferred to a liquid nitrogen tank, where they were securely stored for a minimum of 7 additional days before testing thawing success.

2.8. Thawing procedure

The cryovials were thawed in a water bath and subsequently, as soon as the pellet had dissolved, 4 mL of fresh serum-free culture medium was added. Cells then underwent centrifugation at 1000 x g for a duration of 3 min. The resulting cell pellet was resuspended in fresh 5 mL of serum-free medium and aliquots taken for cell counting and viability assessment by Casy TTC Cell Counter & Analyser System (cells after thawing). Cells were then seeded into a 25 cm² flask. Cell recovery was quantified using Eq. (2)

$$Cellrecovery(\%)=\frac{Numberofviablecellsafterthawing}{Numberofviablecellsbeforefreezing}x100$$ (2)

2.9. RTgill-W1 cell line acute toxicity assay

The cell viability assay was performed as instructed by the established protocol according to OECD TG 249 (2021).

2.10. Statistical analysis

Corresponding functions in GraphPad Prism 9.0 software were used. Student's unpaired t tests (two-tailed) were used to compare differences between two groups while data from three or more independent groups were analyzed using a one-way analysis of variance (ANOVA). For all statistic tests, *P < 0.05, **P < 0.01, or ***P < 0.001 were considered as thresholds for statistical significance. Unless indicated otherwise in the figure legends, all data presented are mean values ± SD. GraphPad Prism 9.0 and BioRender were used for data graphing and graphical design.

3. Results

3.1. Selection of FBS replacement components

We selected the Leibovitz’ L-15 (SI Appendix, Table S1) basal medium as the base of our fish cell-focused framework. This medium is well known for the culturing of fish cells (Leibovitz, 1963), and many cell lines, including the RTgill-W1 cell line, were initiated in L-15 as a basal medium (Bols et al., 1994; Kawano et al., 2011). Our primary focus, therefore, was to identify the limitations of L-15 concerning components which are presumably supplied by FBS. Our secondary focus was to account for the needs arising from the epithelial nature of RTgill-W1 cells.

To provide for a systematic selection, we first added components needed for any vertebrate cell, i.e., insulin, transferrin, and selenium (ITS), trace elements, vitamins and albumin. We then considered additional components, such as certain hormones, growth factors and lipids in a second step. We identified 15 components, covering vitamins, metals and minerals, lipids, hormones, and proteins, from the literature (SI Appendix, Table S2) and added epidermal growth factor (EGF) and hydrocortisone to account for the epithelial-like nature of RTgill-W1 cells. These components can be categorized as substrates (starting points for, e.g., enzymatic activity in metabolic pathways), components for biosynthesis and metabolism (essential for, e.g., synthesis of macromolecules or for modulating specific cellular functions), and/or structural building blocks (for, e.g., lipids in biological membranes) (SI Appendix, Table S2).

Based on this systematic selection, we designed the composition and component concentrations of the serum-replacement medium (Table 1). We pooled components known to simultaneously act on cells and accommodated commercially available mixtures, such as ITS, together with ethanolamine (labelled ITS-X) and a chemically defined lipid mixture. We used ITS/ITS-X and hydrocortisone according to the manufacturer's recommendations and conducted preliminary range-finding experiment over the span provided by the manufacturer for the lipid mixture. The concentrations of biotin and vitamin B12 were comparable to those first documented for hippocampus neuronal cells to leverage the well-established benefits of these vitamins in supporting cellular metabolism and function within a defined culture system (Brewer and Cotman, 1989; Brewer et al., 1993). The chosen EGF concentration was informed by research on human mesenchymal stromal and ovine tracheal cells, where a similar concentration effectively thickened the cell layer (Jung et al., 2010; O'Boyle et al., 2018), endorsing its potential to enhance cellular responses within the fish cell model. In the case of trace elements, we performed an ICP-MS analysis of FBS to determine which concentrations to supplement the L-15 with (SI Table S3/S4). Finally, we tested albumin concentrations in a range of 0.781–1.5 g/L based on the Certificate of Analysis of the Eurobio Scientific FBS used throughout the study.

Table 1.

Composition of the serum-free medium used for adaptation of the RTgill-W1 cell line.

Supplements	Abbreviation	Components	Concentration (mg/L)
			in stock solution	dilution factor in the final medium
Insulin-Transferrin-Selenium Ethanolamine	ITS-X	Insulin	1 000	100x
		Transferrin	550
		Selenium	0.67
		Ethanolamine	200
Bovine Serum Albumin	BSAreg		100 000	100x
Chemically Defined Lipid Concentrate	CD lipid	Arachidonic Acid	2	500x
		Cholesterol	220
		DL-alpha-Tocopherol Acetate	70
		Ethyl Alcohol 100 %	na
		Linoleic Acid	10
		Linolenic Acid	10
		Myristic Acid	10
		Oleic Acid	10
		Palmitic Acid	10
		Palmitoleic Acid	10
		Pluronic F-68	90 000
		Stearic Acid	10
		Tween 80®	2 200
Trace element mixture	TE	CuSO4×5H2O	318	1000x
		FeSO4 x 7H2O	10 880
		ZnSO4 x 7H2O	11 990
		Cl2H8MnO4	177.7
		L-Ascorbic acid	633
Biotin	Biotin		200	100x
Vitamin B12	VitaminB12		800	200x
Epidermal growth factor (recombinant)	rEGF		10	at first 1000x, later 600x
Hydrocortisone	H		1.81	100x

3.2. Short-term cell proliferation assay for the screening of FBS replacement components

To screen the selected components, we set-up a 96-well plate assay designed to differentiate viable but non-proliferating from actively proliferating cells. The method estimates cell proliferation and viability by measuring changes in fluorescence intensities of specific dyes indicating cell metabolic activity, cell membrane integrity and cell count, combined with phase-contrast imaging for cell morphology analysis (Fig. 1A, see Materials and Methods for details). An increase in fluorescence intensity of the live-cell indicator dyes for metabolic activity and cell membrane integrity, combined with fluorescent-based quantification of cell numbers, therefore provides a direct measure of cell proliferation. We started by allowing cells to attach for 24 h in L-15 with 5 % FBS before changing the medium to the desired exposure medium. The first measurement of cellular parameters was then taken 24 h later, i.e., at 48 h in culture and 24 h exposure in the respective medium (Day 1).

Fig. 1.

Systematic approach towards tailored serum-free medium design. (A) Workflow of de novo designed 96-well plate cell proliferation assay. (B) Assessment of cell viability based on (i) metabolic activity (alamarBlue™) and (ii) cell membrane integrity (CFDA-AM) from day 1 to day 7 in the respective medium using 7 500 cells as initial seeding density. For days 2, 5, 6: N = 3; for days 4 and 7: N = 2. (iii) Mean cell count obtained by direct whole well nuclei counting of each culturing condition on days 1 and 6 in the respective media, N = 18. **** P < 0.0001, one-Way ANOVA Šídák's multiple comparison test. (C) Screening of chosen supplements on metabolic activity of RTgill-W1 cells at Day 6 in l-15 vs. l-15 with the added components. N = 3. **** P < 0.0001, one-Way ANOVA Tukey's multiple comparison test. (D) Representative phase-contrast images of RTgill-W1 cells at Day 6 in either L-15 with 5 % FBS (left) or L-15 with FBS replacement components, i.e. serum-free (right). Scale bar 300 µm.

Using this experimental set-up and a starting density of 5 000 to 10 000 cells per well, we identified Days 5–7 as optimal to achieve the desired resolution of only viable vs. proliferating cells (SI Appendix, Figure S1). We found that L-15 medium alone keeps cells viable with stable values for cell metabolic activity and cell membrane integrity until at least Day 6 with only a slight decrease at Day 7 (Figure 1Bi-ii). On the contrary, L-15 with 5 % FBS leads to a significant increase in metabolic activity and membrane integrity for at least seven days. We therefore decided to use a 5-day period from Day 1 to Day 6 for quantifying the influence of the selected components on the cells. This period also resulted in a several-fold increase in cell number (Figure 1Biii), which correlates well with cell metabolic activity and membrane integrity, our two measures of cell viability (SI Figure S2).

We next compared the metabolic activity of cells exposed to each selected component (Table 1) to exposure in L-15 alone (Fig. 1C). We kept each component leading to an improvement in our medium formulation and tested its influence in the mixture. Following this strategy, we found that all components enhanced metabolic activity from about 3 to 6.5-fold compared to L-15.

The simplest medium, though among the least effective, contained only ITS-X and BSAreg (Fig. 1C), selected from a pre-screen of BSA variants in combination with either ITS or ITS-X (SI Figure S3A). We next tested the influence of commercially available Chemically Defined (CD) lipids in our formulations. Pre-screening of different lipid mixture concentrations (250x, 500x, 1000x dilutions) revealed an optimum of 500x dilution in combination with ITS-X/BSA (SI Appendix, Figure S3B), which we used throughout in our medium component list. Further, the trace element (TE) mixture showed statistically significant improvement in cell metabolic activity upon addition to the ITS-X/BSA and CD.

Both the vitamin Biotin and B12 individually enhanced cell metabolic activity over the ITS-X/BSA/TE mixture to a small and comparable extent with a further slight enhancement if added in combination (Fig. 1C). We additionally assessed the impact of recombinant EGF (rEGF) and the hormone hydrocortisone (H) on cell metabolic activity. Both showed a comparably positive impact with slightly better performance of hydrocortisone alone. We nevertheless included both components in the final serum-free formulation to account for their different functionalities (SI Appendix,Table S2 and Fig. 2C) and because both are necessary for the adaptation of cells to a serum-free medium for long-term proliferation (SI Appendix, Figure S4).

Fig. 2.

Adaptation strategies for transitioning of RTgill-W1 to serum-free medium. (A) Two approaches have been tested: direct (SOS cells) and gradual (W cells) serum replacement. (B) Timeline of cell adaptation to the serum-free medium. (C) Representative phase contrast images of the cells throughout different passages during adaptation to serum-free medium. Confluent cells are presented in passages 1, 3, 10, and 15. Pictures were obtained using a 10x objective in phase contrast module by Cytation 5. Scale bar 300 µm.

Overall, we added 27 components to L-15 to arrive at our serum-free medium (Table 1). Examination of phase-contrast images of RTgill-W1 cells cultured in this medium until Day 6 revealed a morphology remarkably similar to that of cells cultured in the presence of 5 % FBS (Fig. 1D).

3.3. Adaptation of RTgill-W1 cells to serum-free medium

To confirm the proliferation performance obtained during the screening process in the 96-well plate culture format, we scaled up the culture conditions to using 25 cm² cell culture flasks. Our aim was to test if the RTgill-W1 cells can be adapted to the newly developed serum-free medium for long-term proliferation and routine culture (Fig. 2A). We used two distinct strategies to assess the adaptability of RTgill-W1 cells to the newly formulated serum-free medium: "Sink-Or-Swim" (SOS), which directly replaced L-15 with 5 % FBS with completely serum-free medium, and "weaned" cells (W), which involved a gradual transition from serum-containing to serum-free medium. Both strategies successfully facilitated the adaptation of the RTgill-W1 cell line to the serum-free medium, but the time required for full adaptation differed (Fig. 2B).

We seeded the SOS cells at about 20 % higher density in 25 cm² culture flasks (4.25 × 10⁵ cells/flask in 5 mL medium) compared to the W cells (3.5 × 10⁵ cells/flask in 5 mL of medium) to facilitate a workable starting cell density despite the abrupt change in the medium composition. In case of the SOS cells after 13 days, cells were sufficiently confluent (>80 %) to attempt first passaging using Accutase®. After passaging, cells started to grow very slowly but consistently. A monolayer formed during about five weeks, after which we observed the formation of gaps, which subsequently remained unpopulated (Fig. 2C). Notably, this phenomenon ceased with subsequent passaging.

The W cells were gradually adapted over a period of three weeks without sub-culture. Once the medium was serum-free for 24 h, we performed a first split using Accutase®. Proliferation was significantly slower in the serum-free medium initially, and cells reached confluency before passaging only after about six weeks. However, adapted RTgill-W1 cells exhibited steady cell proliferation and reached confluency every 10 to 14 days after the third passage.

Both SOS and W cells were subsequently continuously cultured. At the time of writing the manuscript, SOS-adapted cells reached 17 passages before proliferation slowed and senescence set in at passage 19. In contrast, W cells were cultured up to passage 35 thus far without any signs of slower proliferation and senescence. We further successfully repeated the weaning process in two additional independent trials. Hence, we used W cells for all subsequent investigations.

3.4. Functionality of RTgill-W1-sf (serum-free) cells

Cell proliferation: We next evaluated growth characteristics of cells adapted to serum-free medium (RTgill-W1-sf) by cell metabolic activity and cell number and compared them to conventionally cultured RTgill-W1 cells. We also investigated the influence on cell proliferation of the cell dissociation method during sub-culturing (Fig. 3A). Serum-free cells demonstrated their capacity to proliferate and form monolayers, exhibiting typical growth patterns characterized by a lag, exponential, and stationary phase. Growth characteristics of the serum-free cells differed little from the conventionally cultured cells with the only discernable effect being an earlier on-set of the stationary phase for the serum-free cells ≥10 days. However, cell population doubling times revealed an influence of the cell dissociation method. Based on the cell numbers over time (Fig. 3A), population doubling times amounted to 3.71 ± 0.56, 6.81 ± 0.96 and 4.84 ± 0.62 days for regular trypsin (pig porcine) in conventionally cultured cells, and Accutase® (derived from crustacean) and TrypZean® (recombinant expressed in corn, used together with trypsin inhibitor from soy) for serum-free cells, respectively, with the doubling time for Accutase® being significantly higher (Fig. 3B). Hence, we subsequently passaged RTgill-W1-sf cultures using TrypZean®.

Fig. 3.

Functionality assessment of RTgill-W1-sf. (A) Proliferation curves of RTgill-W1-sf cells obtained by the W adaptation method passaged using Accutase® or recombinant trypsin (TrypZean®) and RTgill-W1 (FBS), N=3. (B) Doubling times calculated from (A) ** P < 0.01, (C) RTgill-W1-sf cell recovery calculated post thawing for the cells obtained by weaning (W) or direct (SOS) approach of adaptation to the serum-free medium. (D) The concentration-response curve of RTgill-W1-sf and RTgill-W1 exposed to 3,4-Dichloroaniline for 24 h N=3.

Cryopreservation: To establish cryopreservation, we subjected RTgill-W1-sf cells to multiple subcultures followed by cryopreservation in liquid nitrogen using a 10 % dimethyl sulfoxide (DMSO) solution in the serum-free medium. After storage for extended times and subsequent thawing, we measured the number of viable cells (Fig. 3C). In every batch that underwent the freeze-thaw process, cell recovery consistently approached approximately 80 %, underscoring the robustness of the cryopreservation protocol and the feasibility of long-term storage of RTgill-W1-sf cells. Both cell cultures adapted via the SOS and the W methods were able to grow and form a monolayer, but SOS cells took longer to reach P3 (Passage 3) after thawing (data not shown).

Performance of RTgill-W1-sf according to OECD TG 249: The OECD Test Guideline 249 (Fish Cell Line Acute Toxicity - The RTgill-W1 cell line assay) is increasingly used for assessing the acute toxicity of chemicals to fish. We therefore tested if our RTgill-W1-sf cells pass the criteria of the positive control, 3,4-Dichloroaniline, according to OECD TG 249 (Table 2 and (OECD 2021)). Both RTgill-W1-sf and RTgill-W1 responded similarly with no significant differences in EC50 values, demonstrating the suitability of our medium for use in OECD TG 249 (Fig. 3D).

Table 2.

Comparison of EC50 values of RTgill-W1-sf and RTgill-W1 (FBS) to values reported in OECD TG 249^*† (N = 3).

		EC50 ± SD (mg/L)
		RTgill-W1-sf	RTgill-W1 (FBS)	OECD TG249
Cell metabolic activity	alamarBlue™	37.5 ± 6.6	42.2 ± 2.4	43.6 ± 6.1
Cell membrane integrity	CFDA-AM	73.7 ± 9.1	67.4 ± 5.8	62.5 ± 18.9
Lysosomal membrane integrity	NeutralRed	50.6 ± 2.4	52.7 ± 1.2	58.6 ± 18.6

^⁎The validity criteria for the OECD TG 249 indicate that the EC50 values should be in the range of 28.4 mg/L to 58.9 mg/L, 15.3 mg/L to 109.8 mg/L, 12.1 mg/L to 105.0 mg/L for cell metabolic activity, cell membrane integrity and lysosomal membrane integrity, respectively.

^†Chemical exposure media in both cases RTgill-W1 (FBS) and RTgill-W1-sf is in L15/ex.

3.5. Towards the replacement of animal (including human)-origin supplements

While we succeeded in developing a serum-free medium for routine culture of RTgill-W1 cells, we acknowledge that three of the 27 components added to L-15 are still of animal origin (SI Appendix, Table S5). Serum albumin (BSA) originates from bovine blood; transferrin in the ITS-X supplement originates from human serum; and cholesterol in the lipid mixture is sourced non-invasively from sheep wool. Our further efforts therefore focused on replacing these three components to move our serum-free medium (for convenience below referred to as SFM) to being fully animal/human component-free and preferably of recombinant nature, while avoiding undefined supplements such as hydrolysates from plant sources. We devised a three-step strategy resulting in four different formulations (Fig. 4A): Firstly, we directly replaced the ITS-X supplement with an animal origin-free version of this supplement (AOF-ITS) at P3 in the newly adapted RTgill-W1-sf cells. This formulation of the serum-free medium was named *SFM. Secondly, we considered replacing BSA with recombinant human serum albumin (rHSA), expressed either in Oryza sativa (SFM I) or Pichia pastoris (SFM II). Thirdly, we replaced the complex mixture of chemically defined lipids with a mixture consisting of only two essential fatty acids (oleic and linoleic acid), recombinant cholesterol, and vitamin E (SFM III, SI Appendix, Table S6). The 96-well plate assay for short-term cell proliferation revealed a strong negative impact of O. sativa-derived rHSA on both cell metabolic activity and cell number (SFM I). The rHSA derived from P. pastoris (SFM II) showed less pronounced negative impact on cell count, but still exhibited reduced metabolic activity (Fig. 4B). We confirmed the particularly negative impact of O. sativa-derived rHSA by microscopy: cells showed signs of cellular stress along with reduced proliferation (Fig. 4D, top row). We found the same results when comparing with cells cultured in serum-free medium with rHSA-derived from P. pastoris (Fig. 4D, bottom row) that were passaged seven times in SFM II (w P. pastoris) before slowing in proliferation and eventually dying off. In subsequent experimentation, we investigated the impact of substituting cholesterol (originating from sheep wool) from the lipid mixture (CD lipid) with a recombinant one. Due to the unavailability of a commercial CD lipid mixture with recombinant cholesterol in small batches, we formulated a customized lipid blend comprising recombinant cholesterol, a limited selection of fatty acids, vitamin E and the emulsifier Tween 80 and shear stress protectant Pluronic F-68 in the same ratio as in the CD lipids (SI Appendix,Table S6). Substitution of cholesterol from the lipid mixture (CD lipid) with recombinant cholesterol yielded varying effects on cell metabolic activity and cell number, contingent upon the source of albumin (BSA or rHSA) in the serum-free formulations. Notably, the presence of bovine serum albumin (BSA) alongside recombinant cholesterol did not significantly influence cell metabolic activity (SFM* BSA-rCholesterol). However, the incorporation of recombinant human serum albumin (rHSA from P. pastoris) in the formulations led to a decrease in both cellular metabolic activity and cell number (SFM III), indicating that rHSA exerts a sustained inhibitory effect that overrides mitigation by the lipid mixture containing recombinant cholesterol.

Fig. 4.

Towards an animal (including human)-origin free medium for RTgill-W1. (A) Stepwise replacement strategy of animal-derived components in the serum-free media (SFM). (B) Testing the influence of the rHSA from two different expression systems in the serum-free media on cell metabolic activity and cell proliferation at Day 6. Data are presented as mean ± standard deviation.*P < 0.05. N=3. (C) Adaptation strategy for transitioning of RTgill-W1-sf cells to SFM I and SFM II. (D) Representative phase contrast images of RTgill-W1-sf cells over 21 days after the direct transfer to SFM I and SFM II consisting of either rHSA expressed in O. sativa (top panel) or rHSA expressed in P. pastoris (bottom panel). Scale bar 300 µm. (E) Testing the influence of recombinant cholesterol in the lipid mixture in combination with either rHSA or BSA on cell proliferation and cell metabolic activity at Day 6. ANOVA multiple comparison test, ns-non-significant, N=3.

4. Discussion

We developed and characterized the first systematically derived and fully transparent serum-free culture medium for fish cell lines, using the RTgill-W1 rainbow trout gill cell line for proof-of-concept. We decided on the L-15 culture medium to serve as a base to which we added serum-replacement components. The L-15 culture medium has been proven repeatedly to be suitable for cells of cold water (Bols et al., 2017) and even warm water fish (Tierbach et al., 2020).

Despite its dedicated design for vertebrate cell culture at lower temperatures, a range of important components are lacking in L-15 to facilitate cell proliferation and hence must traditionally have stemmed from FBS. One component is ITS, which is considered universally essential for vertebrate cell cultures (van der Valk et al., 2010). Four of the other components missing from the L-15 medium are the vitamins biotin, B12, E and C. Biotin is considered an essential water-soluble vitamin for fish; deficiency leads to poor fish growth and degeneration of the gills in rainbow trout (He et al., 2020; Castledine et al., 1978). In cultured mammalian cells, biotin has also been shown to play multiple roles, such as de novo synthesis of fatty acids (Ham, 1981). Vitamin B12 has been postulated as essential for various fish species' growth and health (Khan and Khan, 2020), although only one type of extracellular cobalamin-binding protein was found in rainbow trout (Greibe et al., 2012). Vitamin B12 has been part of some culture media mostly designed for neuronal and hybridoma cells, but also other types of mammalian cells because it plays a role in cell survival, proliferation and genetic stability (Yamamoto and Niwa, 1993; Battaglia-Hsu et al., 2009). Vitamin E was included in our serum-free medium as part of the lipid mixture. This decision was influenced by its well-known antioxidant properties, which are beneficial for maintaining cell health and viability (Arigony et al., 2013). Vitamin C was deliberately incorporated into the trace element mixture due to its potential for enhanced stability when combined with certain elements like manganese and selenium (Dolińska et al., 2012). Furthermore, when included in culture medium, Vitamin C acted as a proliferative factor, promoting the growth and expansion of primary myoblast cell cultures isolated from the fast-twitch muscle of pacu fish (Piaractus mesopotamicus) (Duran et al., 2019).

Trace elements are known to play a pivotal role in cell proliferation and genome stability across various organisms (Arigony et al., 2013; Prabhu and Gadgil, 2021), though their particular role in fish cell culture is understudied. Research on fish cells thus far predominantly focused on the cytotoxicity of individual metals (Bopp et al., 2008; Tan et al., 2008), rather than on exploring their nutritional aspect. Studies on an American eel brain endothelial-like cell line showed that the effects of selenium supplements vary with exposure medium (L-15/ex vs. L-15/FBS) and selenium speciation: e.g., selenite exhibited cytotoxicity in both media; however, the effects were more pronounced in L-15/ex. This suggests that FBS in the medium may mitigate some of the cytotoxic effects (Bloch et al., 2017). An in vitro cell system derived from gilthead seabream vertebra, capable of mineralization, was used to confirm that B vitamins (thiamin and pyridoxine) and trace minerals (copper, manganese, and zinc in sulfated and chelated forms) are needed for cell proliferation and extracellular matrix mineralization (ECM) (Arigony et al., 2013). Our targeted analysis of trace elements confirmed their absence in L-15 while their presence in FBS yielded valuable insights into the types and quantities of trace elements in serum (SI Appendix,Table S3). We adopted the trace element concentrations to those found in the serum (SI Appendix,Table S4). Future experiments on the stability and potential pro-oxidative roles of trace elements in the medium may provide additional insights and aid in further fine-tuning trace element levels. For example, vitamin B12 is instable in presence of ascorbic acid (vitamin C), which is particularly exacerbated in the presence of Cu²⁺ salts (Schnellbaecher et al., 2019).

Our findings suggest that selecting appropriate albumin supplementation may be more complex than initially perceived. The significantly better cellular proliferation and greater cell metabolic activity of the serum-free media containing albumin (BSAreg) underscore the complex role of the protein. A major positive influence of albumin in culture medium lacking FBS has been extensively documented in mammalian cells, such as antioxidant properties and facilitated transport of molecules across the cell membrane (Francis, 2010). On the contrary, the influence of albumin in FBS-free media on fish cells has rarely been reported: 5 % BSA (fatty acid free) in L-15 medium was able to support proliferation of the Chinook salmon embryo (CHSE- 214) cell line over seven passages, while cells of a rainbow trout gonadal cell line (RTG-2) could be passaged only twice (Barlian et al., 1993; Barlian and Bols, 1991). Ultimately, an alternative to animal-derived albumin must be part of chemically defined culture media. Most serum-free media published today report the use of albumin from human serum (Rafnsdóttir et al., 2023; Stout et al., 2022). Human serum albumin (HSA) is, however, like BSA, undefined and variable, hence not meeting the criteria posed for a chemically defined medium (Raoufinia et al., 2016). The limited availability of animal-origin-free albumin in the market poses a significant obstacle. Easily available recombinant albumin, such as rHSA, appears promising for more defined conditions for cell culture. Yet, our testing revealed that rHSA has a rather negative impact on cell viability of RTgill-W1 cells. This was observed with both rHSA sourced from Pichia pastoris and from Oryza sativa, respectively, though the negative impacts were more pronounced with the latter. Our findings align with a report on inhibitory effects of rHSA towards a kidney cell line exhibiting epithelial morphology (NRK cells) (Keenan et al., 2006) while rHSA (expressed in O. Sativa) was identified as one of six factors enhancing cell proliferation in primary myogenic cell lines of bovine muscle cells (Stout et al., 2022). The results presented here underscore the significance of various factors associated with recombinant protein sources, such as their bioactivity, purity, and levels of endotoxins, which could potentially contribute to the differing outcomes observed for bovine serum albumin (BSA), human serum albumin (HSA) and their recombinant counterparts (Schwarz et al., 2014). Overall, we failed to successfully adapt our serum-free cells to any of the rHSA for long-term proliferation. Protein-free formulations supplemented with dipeptides may offer a viable alternative, especially considering that synthetic oligopeptides have been shown to enhance cell viability in hybridoma cells (Franek and Katinger, 2002).

We aimed to limit the addition of medium components to those essential for cell viability and proliferation. Nonetheless, the largest proportion of the 27 components included in our serum-free formulation is attributed to a commercially available lipid mixture (13 components). This mixture is intended to reduce or replace the need for FBS for a wide range of mammalian and insect cultures (Jochems et al., 2002). Nevertheless, the question remains of how specific this mixture is for fish cells. Several fish cells cultured in FBS-supplemented media were shown to be deficient in omega-3 polyunsaturated fatty acids when compared to intact fish (Tocher et al., 1988). Indeed, our proposed medium should be seen as a backbone from which more fine-tuned media can be derived to better support fish cell-specific traits, such as lipid composition. First indications from downsizing the lipid mixture in our serum-free medium formulation seem promising: supplementing the serum-free medium with only two fatty acids and recombinant cholesterol (SI Appendix,Table S6) shows equally good cellular viability and proliferation by Day 6 compared to the serum-free medium containing the 13 component CD lipid mixture.

It is well documented that FBS is an abundant source of various growth factors and hormones, serving as a versatile and widely applicable component for the culture of diverse vertebrate cell lines (Yao and Asayama, 2017; Rizzino et al., 2009). In our pursuit of formulating minimal media conducive to cell viability and proliferation under serum-free conditions, our initial strategy focused on evaluating hydrocortisone and recombinant epidermal growth factor due to their documented positive influence on non-transformed epithelial cells (Ham, 1981; Kelly and Wood, 2001; Alexander et al., 2015). Cultivation of RTgill-W1 cells failed in the absence of these two factors, indicating their role in the promotion of cellular attachment and the overall metabolism of RTgill-W1 cells. The necessity of hormone and/or growth factor combinations varies considerably among different cell types, mainly normal versus transformed cells, where each component may exhibit synergistic or antagonistic properties when integrated into cell culture media. In fish, cortisol is considered as principal corticosteroid which regulates osmoregulation, growth and reproduction (Mommsen et al., 1999). Cortisol combined with 3,3′,5-Triiodo-L-thyronine exhibited favorable impacts on the physiological parameters of cultivated pavement cell epithelia from freshwater rainbow trout gills (Kelly and Wood, 2001). Cortisol supplemented to FBS–containing media was found to change cell morphology of the rainbow trout monocyte cell line (RTS11) at 100 and 1000 ng/mL (Pagniello et al., 2002). Despite the concentrations tested in our serum-free formulations being close to 10 ng/mL, our findings align with studies where enhanced cellular attachment and proliferation were observed upon the addition of the synthetic cortisol, hydrocortisone, particularly in adherent cell types such as primary bovine satellite cells and human mesenchymal stem cells in serum-free formulations (hMSCs) (Jung et al., 2010; Kolkmann et al., 2020).

We tested two strategies of cell adaptation to the serum-free medium – Sink-or-Swim (SOS) and gradual weaning (W). Both strategies effectively facilitated cell adaptation, albeit with differing timeframes and requirements: the SOS cells, although seeded at a higher density, reached sufficient confluency to attempt first passaging after 13 days. The W cells were gradually adapted over three weeks without sub-culture. Subsequent adaptation, however, significantly varied: SOS cells reached P3 (Passage 3) after three months compared to 1 month for W cells. We considered P3 an important milestone for RTgill-W1-sf cells: they are likely free of FBS from carry-over and start to demonstrate growth dynamics similar to those observed in serum-supplemented media at this stage. Overall, the W method not only yielded faster adaptation compared to SOS cells but also cells of better viability post-thawing. Successful adaptation by the SOS method has rarely been reported in the past and only applied to mammalian cells (Kolkmann et al., 2020; Chary et al., 2022). A gradual approach, similar to our W method, has been found to generally yield successful adaptation in fish (Yue et al., 2016) and mammalian cells (Chary et al., 2022). However, information about the long-term culturing potential of these cells in the serum-free medium from this prior work is either missing or culturing was short-spanned: e.g. up to 10 passages for RTgill-W1 weaned into In Vitrus medium (Yue et al., 2016).

A key to the functionality of the RTgill-W1 cell line assay as described in the OECD TG 249 and the ISO Standard 21115 is use of the serum-free medium, L-15/ex, during the chemical exposure phase (Schirmer, 2006; Schirmer et al., 1997; Tanneberger et al., 2013). Yet, routine culture of the cells thus far still depends on FBS supplementation. The RTgill-W1-sf cells derived in this study passed the validity criteria set out for the RTgill-W1 cell line assay to predict acute fish toxicity of chemicals according to OECD TG 249 for the positive control, 3,4-DCA. This comparable behavior in terms of sensitivity toward a toxic chemical is not a given as cultivating cells in different media can influence their morphology as well as functional and structural properties, ultimately affecting their sensitivity to toxicants (Chary et al., 2022; Janetzki et al., 2010). More chemicals will need to be tested to confirm the agreement in sensitivity of RTgill-W1 and RTgill-W1-sf cells. Nevertheless, the excellent performance of the serum-free cultured cells in comparison to RTgill-W1 cells in FBS-supplemented medium provides impetus to move the OECD TG 249 and ISO Standard 21115 to being completely serum-free.

5. Concluding remarks

Our research strategy has led to the development of a culture medium formulation omitting the need for FBS. Our new medium allowed cells to survive and proliferate for already 35 passages. Moreover, the medium allows cells to be frozen and revived after cryopreservation without negative effect. This medium can thus serve as a versatile platform from which fine-tuning of media components can be pursued to further cater to other fish cell lines and their applications. For instance, the intestinal cell line from the rainbow trout (RTgutGC) has been established as in vitro cell culture model for nutrient absorption and pathogen defense via barrier functions, accompanied by immune-related responses (Minghetti et al., 2017; Schug et al., 2019). Following our work, these cells might benefit from growth factors additional to EGF, such as fibroblast growth factor, or short-chained fatty acids that are a major energy source for the intestine (Selvam et al., 2022). Economic concerns might prove a severe roadblock for the future: the requirement for increasing chemical definition and purity also increases the cost for putting a serum-free, and even more so, animal component-free, culture into praxis. For instance, the price per liter of ready-to-use serum-free medium developed in this study is currently about 2 times higher than FBS, and the price of the animal component-free medium would be almost 4 times higher (SI Appendix, Table S7). These increased costs should, however, be balanced against hidden costs arising, e.g., from the need to test each different batch of FBS, as well as against the ethical dilemma and the uncertainty arising from its variable, undefined nature.

Funding

This study was funded by the following agencies: Swiss 3R Competence Center grant OC-2018-001, Unilever research grant MA-2022-00768 N and Swiss National Research Foundation grant 203578.

CRediT authorship contribution statement

Barbara Jozef: Writing – review & editing, Writing – original draft, Supervision, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization. Zhao Rui Zhang: Writing – review & editing, Visualization, Investigation. Hans-Michael Kaltenbach: Writing – review & editing, Investigation, Formal analysis. Kristin Schirmer: Writing – review & editing, Writing – original draft, Supervision, Project administration, Methodology, Funding acquisition, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

Data will be made available on request.