Endothelial growth media components alter SARS-CoV-2 spike-directed growth kinetics

Abstract

Direct SARS-CoV-2 infection of endothelial cells is challenging to study in vitro. To examine whether endothelial cell culture conditions impact the ability of SARS-CoV-2 to infect cells, we evaluated the effects of commercial cell culture media composition on SARS-CoV-2 Spike-directed viral infection. In African Green Monkey kidney epithelial cells (VeroE6), we found that commercial cell culture media (EGM2) produced inhibitory effects on recombinant vesicular stomatitis virus (rVSV-SARS-CoV-2) growth that is not seen in Dulbecco’s Modified Eagle Medium (DMEM). We identified Lonza’s GA-1000 supplement as a potential SARS-CoV-2 inhibitor. We then compared how titrations of GA-1000 influenced rVSV-SARS-CoV-2 and SARS-CoV-2 infection. The determined IC50 of GA-1000 is observed at a 1:450 dilution, about 2-fold higher concentration than the typical cell culture concentration, against rVSV-SARS-CoV-2. In contrast, the inhibition against SARS-CoV-2 infection is modest. This work highlights the importance of cell culture medium selection for in vitro models of infections and demonstrates a method to characterize and ameliorate media formulations for infection studies.

1. Introduction

Although COVID-19 is often considered a respiratory illness, the pandemic has revealed that the pathophysiology of SARS-CoV-2 extends beyond the lung. Endothelial dysfunction, or endotheliopathy, was a prevalent feature of severe COVID-19 early in the pandemic. Multiple efforts have attempted to clarify the mechanisms of how the endothelium is disrupted in SARS-CoV-2 infections (Schimmel et al., 2021; Zhang et al., 2020; Knight et al., 2022; Whiteley and Wood, 2022; Singh et al., 2023; Barrett et al., 2021; Nappi and Avtaar Singh, 2022). Despite the frequency of the clinical presentation of endothelial dysfunction in patients with severe COVID-19, the mechanism underlying this pathology is unknown. Recent in vitro studies support the idea that endothelial cells are indirectly damaged by SARS-CoV-2 infection while other studies fail to capture infection in vitro. Lack of infection has been attributed to low expression of critical host factors such as, the angiotensin converting enzyme 2 (ACE2) receptor (Schimmel et al., 2021; Bugatti et al., 2022; McCracken et al., 2021). However, aside from host cell properties, there are other extrinsic factors in cell culture that may have influenced the lack of infection found in these studies.

A critical and often overlooked aspect of in vitro model systems is cell culture growth medium selection and optimization. For example, serum is a common additive in cell culture media and has been shown to occlude the furin cleavage site within Spike which is necessary for directing membrane fusion in viral entry (Varricchio et al., 2024; Lamers et al., 2021). In the case of endothelial cells, previous studies have yielded conflicting data and inconsistent reports of direct SARS-CoV-2 infection. In addition to differences in the endothelial cell types evaluated, studies also differed by the growth media used in in vitro studies (Table 1). Therefore, it is possible that components, such as serum or other supplements, in endothelial cell growth media could be a contributing factor to the discrepancies observed in direct and productive SARS-CoV-2 infection in endothelial cells.

Table 1.

Impact of Growth Media Selection on SARS-CoV-2 infection of Endothelial Cells.

Cell Culture Media	Infection (Yes/No/Not Productive)	Includes Gentamicin and Amphotericin	Reference
EGM2 (Case 1)	No	Yes	(Schimmel et al., 2021; Bugatti et al., 2022)
EGM2 (Case 2)	Not productive	Yes	(Nascimento Conde et al., 2020)
DMEM	Yes	No	(Amraei et al., 2022)
OptiMEM	Yes	No	(Liu et al., 2021)
RPMI	Yes	No	(Monteil et al., 2020)
ECM	No	No	(Rauti et al., 2021)
HuMedia-EG2	Not productive	Yes	(Urata et al., 2022)
M199	No	No	(Colunga Biancatelli et al., 2021)

Infection: Yes = Direct and productive infection. No = No direct infection. Not productive = Direct infection but no generation of progeny virions. Includes Gentamicin and Amphotericin: Yes = Antibiotic/Antifungal was added. No = Antibiotic/Antifungal was added or not mentioned in the study.

This study evaluated the effects of the Lonza EGM2 Endothelial Cell Growth Medium-2 Kit (Catalog No. CC-3162), including base media and supplements, on SARS-CoV-2 infection in a permissive cell line. For this study, we infected African Green Monkey cells (VeroE6) with a reporter tagged, recombinant vesicular stomatitis virus that expressed the SARS-CoV-2 Spike protein (rVSV-SARS-CoV-2-GFP). From this assessment, we identified GA-1000, a combination of both gentamicin and amphotericin, as a supplement with antiviral effects on Spike-directed viral entry. We show that GA-1000 generates variable inhibitory effects on chimeric VSV and SAR-CoV-2 growth kinetics, but its variability is controllable under specific base media and multiplicity of infections (MOIs).

2. Materials & methods

2.1. Cell lines and media

VeroE6 cells (African green monkey kidney cells) were obtained from ATCC (Catalog No. CRL-1586). Dulbecco’s Modified Eagle Medium (DMEM) (Catalog No. 11965092) and fetal bovine serum (FBS) were obtained from ThermoFisher Scientific. EGM2 Endothelial Cell Growth Medium-2 BulletKit^™ (Catalog No. CC-3162) was procured from Lonza and included EBM-2 Basal Medium (CC-3156) and EGMTM-2 Single-Quots Supplements (CC-4176).

2.2. Methods

2.2.1. Cell and virus culture

VeroE6 cells were maintained in DMEM containing 7 % FBS and 5 mM L-glutamine. Two virus stocks were used in infection growth kinetics: recombinant vesicular stomatitis virus (rVSV-SARS-CoV-2-GFP) and SARS-CoV-2-mCherry (USA-WA1/2020 isolate (NR-52281)). The rVSV-CoV-2-GFP were chimeric VSV virions which possessed the SARS-CoV-2 S protein and included a green fluorescence protein (GFP) reporter gene (Dieterle et al., 2020). SAR-CoV-2 also possessed a mCherry reporter gene and was a gift from the laboratory of Luis Martinez-Sobrido (Chiem et al., 2021). Virus stocks were grown in VeroE6 cells and virus titers were determined by plaque assays (AU - Baer and AU - Kehn-Hall, 2014). All cell and virus cultures were grown at 37°C and 5 % CO_2.

2.2.2. Comparison of chimeric VSV growth in endothelial growth media

VeroE6 cells were seeded into 12-well plates in complete DMEM. After 24 h, the cell culture medium was replaced with either complete DMEM or EGM2 (both containing 7 % FBS). Cells were infected with chimeric VSV at MOI 0.1 for 48 h. At 48 h post infection, the 12-well plate was live imaged to observe GFP positive cells for comparison of the amount of virus growth in each media condition. The amount of virus growth was quantified by the percentage of area of image that was GFP positive.

2.2.3. Evaluation of chimeric VSV infection inhibition of individual factors in EGM2

For screening of EGM2 kit supplements, VeroE6 cells were infected with rVSV-CoV-2-GFP at MOI 0.1 for 48 h in complete DMEM, complete EGM2, or EGM2 basal medium (EBM2) with an individual supplement. The supplements screened included hydrocortisone, vascular endothelial growth factor (VEGF), human fibroblastic growth factor (h-FGF-ß), insulin-like growth factor 3 (R3-IGF-1), ascorbic acid, human endothelial growth factor (hEGF), GA-1000, and heparin. After 48 h, media was aspirated and replaced with 4 % paraformaldehyde (PFA). After 15 minutes of fixation, cells were imaged for GFP positive cells for comparison of the amount of virus growth in each media condition. The amount of virus growth in each imaged was quantified as previously mentioned.

2.2.4. Inhibition assay with GA-1000 supplement

VeroE6 cells were seeded into 96-well plates 24 h prior to infection in complete DMEM. Increasing concentrations of the GA-1000 (1:100,000 to 1:10 dilution) were added prior to infection with rVSV-CoV-2-GFP or SARS-CoV-2-mCherry at MOI 0.1. After 72 h, the plates were fixed with 4 % PFA or 10 % formalin. Fixed plates were analyzed for virus growth through quantification of fluorescence intensity. Mock condition was used to measure the average background of the assay. The average positive infected signal (GFP or mCherry) was calculated by taking the average positive fluorescence signal from n = 3 wells infected without GA-1000 supplementation and subtracting the average background fluorescence signal. The normalized fluorescence intensity at each GA-1000 concentration was determined by subtracting the average background signal from the experimental fluorescence signal and dividing it by the average signal of the positive controls. To determine the IC50, a nonlinear regression was performed to fit the normalized dose-response data to a four-parameter sigmoidal equation (GraphPad Prism 10.1.0). Data on the IC50 curves were presented as averages ± standard deviation.

2.2.5. Imaging

Live imaging of infection plates was performed using fluorescence microscopy at a 10X magnification to evaluate virus positive cells and cytopathic effect. Fixed plates were imaged using Biotek Cytation 1 at 4X magnification to image virus positive cells and DAPI-stained cells.

2.2.6. Statistical analyses

All statistical analyses were performed on GraphPad Prism (Version 10.1.0). Statistical analysis for growth media comparisons between two media conditions used an unpaired, two-tailed Student’s t-test. Each condition included at least n = 3 technical replicates, and normal distribution was assumed. The difference between the media condition and control condition was considered statistically significance at p < 0.05. Data was summarized in bar plots and are represented as average ± standard deviation. Statistical analysis for more than two media conditions used one-way ANOVA. Tukey’s multiple comparisons test was performed to analyze statistical significance at p < 0.05 between conditions.

3. Results

3.1. Chimeric VSV growth in endothelial growth media

First, we compared rVSV-SARS-CoV-2-GFP virus growth in complete DMEM and complete EGM2. We hypothesized EGM2 has an inhibitory effect on infection. With respect to DMEM, there was about a 4-fold decrease in GFP-signal, in complete EGM2 after 48 h post infection (Fig. 1A). In addition, standard deviation among the EGM2 condition was ~2-fold higher in EGM2 than in DMEM. Therefore, EGM2 produced inhibitory and more variability in rVSV-SARS-CoV-2 infection in VeroE6 cells.

Fig. 1.

Live Infection of VeroE6 by a chimeric VSV expressing SARS-CoV-2 Spike. (A) Live imaging of virus growth in complete DMEM and EGM2. (B) Comparison of Infection in EBM2 and EGM2. (C) Quantification of amount GFP in live imaging in DMEM and EGM2 with n = 3 per condition and (D) Quantification of relative amount of GFP signal to DAPI signal in EGM2 media with and without supplementation, n = 3 per condition. Unpaired, two-tailed Student’s t-test was used to determine statistical significance (**p < 0.01, ***p < 0.0001). Data are represented as averages ± standard deviation. Scale bar = 500 μm.

To understand how factors in complete EGM2 impact virus growth, we then compared the effects of complete EGM2 to effects EBM2 on rVSV-SARS-CoV-2-GFP infection. We hypothesized that supplements in EGM2 were contributing factors on the observed inhibitory properties in in vitro virus growth. Each media condition was infected with chimeric VSV at MOI 0.1 for 48 h and then imaged and evaluated for GFP-positive cells. Similarly, there was a marked reduction, ~4-fold decrease, in GFP-positive cells when was grown in complete EGM2 than EBM2, devoid of supplements (Fig. 1B). From this finding, we determined that supplementation of EBM2 led to rVSV-CoV-2-GFP inhibition.

To investigate the effects of individual supplements included in EGM2 kits, we performed various infections in DMEM, EBM2, and EBM2 with a single EGM2 kit supplement. We hypothesized that EBM2 and individual supplements in EGM2 kits inhibited rVSV-SARS-CoV-2-GFP. In this screening, VeroE6 were infected at MOI 0.1 for 48 h. All supplements significantly showed lower virus growth with respect to complete DMEM (Fig. 2B). In addition, images of virus positive cells revealed different cytopathic effect in EBM2 and in supplemented versions that was not typically observed when grown in completed DMEM (Fig. 2). Supplements such as VEGF, h-FGF-B, R3-IGF-1, h-EGF, ascorbic acid, GA-1000, and heparin showed predominantly GFP-positive cells in the form of syncytia whereas syncytia were less prominent in complete DMEM. Of note, the GA-1000 condition showed minimal GFP signal at 48 h, indicating poor rVSV-CoV-2-GFP growth (Fig. 2A). Correspondingly, viral titers derived from rVSV-SARS-CoV-2 samples grown in EBM2 yielded over a 3-fold decrease in PFU/mL. In contrast to live imaging results, there were no infectious particles rVSV-SARS-CoV-2-GFP grown in EBM with only RG-IGF-1, EGF, and GA-1000 by plaque assay (Fig. 2). However, rVSV-SARS-CoV-2-GFP grown in EBM with GA-1000 showed little to no growth both by live imaging and plaque assay. Therefore, EBM2 and EGM2 kit supplements produced antiviral effects, with GA-1000 as the most potent inhibitor.

Fig. 2.

Identification of GA-1000 as an inhibitor in complete EGM2. (A) Cytopathic effect visualized through GFP-positive cells. (B) Quantification of the variable effects of individual EGM2 supplements on rVSV-SARS-CoV-2-GFP at MOI 0.1 at 48 h, compared to complete DMEM (n = 5 per condition). (C) Viral concentrations determined through plaque assays for all media conditions. Scale bar = 500 μm.

3.2. Inhibition assay with GA-1000 supplement

Based on these findings, we determined GA-1000 inhibited SARS-CoV-2 Spike directed viral entry. We hypothesized that the IC50 of GA-1000 should be within of the range of its recommended dilution in the EGM2 kit (1:1000). To robustly assess the inhibitory potential of GA-1000, we generated dose-response curves to determine the IC50 against rVSV-SARS-CoV-2-GFP and SARS-CoV-2-mCherry. In the optimization of this assay, high variation was seen in virus growth in GA-1000 supplemented EBM2. Specifically, the variability was an MOI-dependent characteristic in which lower MOIs (≤0.5) yielded positive linear trends in virus growth with decreasing GA-1000 concentrations and higher MOIs (≥0.5) yielded nonlinear trends in virus growth with decreasing GA-1000 concentrations (Fig. 3C). Therefore, VeroE6 cells were infected at MOI 0.1 for 72 h with various concentrations of GA-1000 in EBM2 to control assay variation.

Fig. 3.

GA-1000 Dose-response inhibition curves. (A) Chimeric VSV expressing SARS-CoV-2 Spike inhibition in EGM2 with titrations of GA-1000, DMEM with titrations of GA-1000, DMEM, and EBM2, (B) SARS-CoV-2-mCherry inhibition in EBM2 supplemented with GA-1000. Data on the IC50 curves were presented as average fluorescence signal ± standard deviation (C) Images of virus-positive cells in inhibition assay with varying dilutions dilution, green = rVSV-SARS-CoV-2-GFP and red = SARS-CoV-2-mCherry. Scale bar = 500 μm.

Overall, GA-1000 had a modest inhibition effect on rVSV-CoV-2-GFP growth at the recommended dilution (1:1000). Instead, the IC50 of GA-1000 against rVSV-CoV-2-GFP was at a dilution of 1:450 in EBM2 (Fig. 3A). Therefore, under these conditions, the IC50 of GA-1000 occurs at about a 2-fold higher dilution than its recommended amount. In contrast, GA-1000 did not exhibit a measurable IC50 in DMEM within the tested concentration range (Fig. 3A). In EGM2 and DMEM devoid of GA-1000, there is no inhibition of virus growth seen after 72 h post infection. To validate these findings, we also examined these media conditions using SARS-CoV-2-mCherry. We hypothesized that other viral proteins in SARS-CoV-2 may overcome the inhibitory effects of GA-1000. While modest inhibition was present at high concentrations of GA-1000, GA-1000 did not exhibit a measurable IC50 in EGM2 within the tested concentration range against SARS-CoV-2. Therefore, lower dilutions of GA-1000 ≤ 1:450 inhibited chimeric VSV growth but had negligible effects with authentic SARS-CoV-2.

4. Discussion

4.1. Comparison of chimeric VSV growth in endothelial growth media

This study characterized the effects of different media and their components on Spike-mediated infection kinetics. Both DMEM and EGM2 are basal media frequently used to grow endothelial cells (Schimmel et al., 2021; Bugatti et al., 2022; Nascimento Conde et al., 2020; Amraei et al., 2022). However, the role of cell culture medium selection on SARS-CoV-2 infection has not been comprehensively studied. To eliminate serum as an inhibitor, media conditions in all assays of this study included 7 % FBS. We found that other factors included in the EGM2 kit produced inhibitory effects. For example, when comparing rVSV-SARS-CoV-2-GFP infection in complete DMEM and EGM2, live imaging of infected samples revealed that there were substantially fewer GFP positive cells in complete EGM2 than in DMEM (Fig. 1). These insights of the antiviral effects of EGM2 kit components shed light on the potential association between cell culture medium selection and presence of infection. Previous findings have shown conflicting observations of either direct SARS-CoV-2 infection and no infection of endothelial cells (Table 1). This association between lack of infection and cell culture medium selection can potentially explain the variability among in vitro SARS-CoV-2 infection in endothelial cells.

4.2. Identifying inhibitors in EGM2

The study showed that EGM2 contains one or more components that are antiviral, but the precise mechanisms of EGM2 kit supplements on virus growth remain unclear. We hypothesized that EGM2 contains inhibitory factors in its basal formulation (EBM2) and showed that supplementation of EBM2 also supported inhibition. For example, we found in complete EGM2, there was a reduction in the number of GFP positive cells with respect to chimeric VSV grown in EBM2 at MOI 0.1 (Fig. 1). This study provided a facile method to determine which individual supplements in EGM2 kits are potentially antiviral. However, the shortcoming to this approach is that EBM2, the basal medium, cannot be further broken down into components in this study. Consequently, it is difficult to determine whether there is a single antiviral component present that is potent or that there are multiple weak inhibitory components working synergistically to generate the overall inhibitory effect observed in EGM2.

The results also demonstrated that EGM2 can also modulate other properties, such as cytopathic effect. Individual EGM2 kit supplements also showed inhibitory effects on chimeric VSV infection (Fig. 2). Fluorescence images revealed syncytia formation rather infection of individual cells typically observed in DMEM (Fig. 2). While cell-cell fusion is expected with SARS-CoV-2 infection (Buchrieser et al., 2020; Zhao et al., 2022; Chaudhary et al., 2023), the observed patterns were primarily seen in EGM2 kit supplements. However, when the assay is scaled to a 96-well plate, this phenotype is diminished in dose-response inhibition assays (Fig. 3). The significance of syncytia formation under EGM2 conditions remains unknown but is a controllable phenotype in different assays in this study. Furthermore, plaque assays revealed that while some media supplements supported GFP-expression of the virus, virus grown in these conditions did not lead to plaque formation (Fig. 2). This result suggests that while GFP expression is shown, some of the viral particles grown in RG-IGF-1, EGF, or GA-1000 supplemented EBM may not be infectious. Therefore, these media supplements can modulate the number of infectious particles in a viral sample.

Prior to this study, there were no previous reports of specific antiviral factors found in EGM2. With exception to heparin, these supplements generally encompass a variety of growth factors, and an antibiotic-antifungal combination needed for maintaining healthy and sterile endothelial cell cultures. However, contrary to previous reports (Paiardi et al., 2022; Mendoza et al., 2020), heparin was not shown to significantly reduce chimeric VSV growth with respect to EBM2 (Fig. 2). In contrast, GA-1000 was shown to significantly inhibit virus growth and produced the highest inhibitory effect of the EGM2 supplements (Fig. 2). GA-1000 is a combination of gentamicin and amphotericin and is intended as an antibiotic and antifungal. In application, these additives are often helpful in primary cultivation of viruses to minimize or eliminate bacterial and fungal contaminations; however, the additives are not required in all applications (Belen-Apak and Sarialioglu, 2020). In contrast to its conventional application, our finding was unexpected as there had not been prior reports of gentamicin and amphotericin impacting viruses with a Spike-directed viral entry mechanism.

4.3. Inhibition assay with GA-1000 supplement

To characterize the effect of GA-1000 on Spike-mediated infection kinetics, we used an inhibition assay to test a range of concentrations of the supplement against rVSV-CoV-2-GFP and SARS-CoV-2. During the optimization of this assay, the inhibitory phenotype seen in EBM2 was a MOI-dependent behavior that is resolved when the virus is grown in DMEM, or infection is performed MOI is ≤ 0.5. This behavior suggests that endothelial cell growth media can lead to variations in SARS-CoV-2 Spike directed infections depending on the MOI used. Overall, GA-1000 had modest inhibition of rVSV-SARS-CoV-2-GFP and SARS-CoV-2-mCherry. Specifically, the IC50 was at a higher concentration (1:450) than the dilution against rVSV-SARS-CoV-2-GFP (Fig. 3). Moreover, modest inhibition (~25 %) was also seen in similar concentrations of GA-1000 in DMEM. However, the IC50 could not be determined for GA-1000 in DMEM against rVSV-CoV-2-GFP, suggesting that the supplement is inert in DMEM. Against authentic SARS-CoV-2, IC50 could not be determined for GA-1000 in EGM2. Therefore, the recommended GA-1000 dilution (1:1000) does not interfere with overall Spike-mediated infection kinetics in this assay. This observation contrasts with the preceding results and suggests that other viral proteins in SARS-CoV-2 overcome the antiviral effects of GA-1000 seen in applications using recombinant VSV. VSV is a virus commonly used for viral entry studies (Dieterle et al., 2020). Therefore, given the variable effects of media supplements on chimeric virus growth, these effects of media supplementation on recombinant VSV’s should be considered in infection studies.

Prior results showing strong inhibition at 1:1000 dilution (Fig. 1&2) could be attributed to the trend that EGM2 generally yielded variable infection results. With insight into the assay variability seen in this study and conclusions from other research studies (Table 1), it is feasible that endothelial growth media has properties that lead to considerable fluctuations in infection kinetics. It is important to note that while GA-1000 has modest inhibitory effects on its own, the supplement is typically used in combination with other additives that may mitigate or exacerbate the observed inhibition. It is possible that these components maintain a protective cell phenotype which may limit observation of infection and virus-induced endothelial dysfunction. Moreover, it is also important to consider that the SARS-CoV-2-mCherry used in this study was based on the WA1/2020 isolate which represents the one of the earliest introductions of this virus in the United States (Holshue et al., 2020). This strain circulated during early stages of COVID-19 pandemic where endotheliopathy was most prevalent in patients with severe cases. Results of this study may be more relevant to other studies using variants of concerns associated with this pathology such as Alpha (B.1.1.7), Delta (B.1.617.2), and Omicron (B1.1.529) (Perico et al., 2024). Our findings partially address whether there are supplements present in commercial endothelial cell growth which can alter SARS-CoV-2 Spike-directed virus growth. We demonstrate that there are antiviral components that influence both the amount of infection and the cytopathic effect observed. Future work should characterize whether these antiviral components found in cell culture medium are cell-type specific or virus specific. Identifying these factors in endothelial growth medium selection is critical for development of vascular models of infection.

5. Conclusion

This study addressed the impact of endothelial growth media on SARS-CoV-2 Spike driven virus growth kinetics. We identified factors that may have impacted the prior findings in COVID-19 associated endotheliopathy in vitro studies. We showed that the formulation of commercial endothelial growth media possesses additives which can generate variations in SARS-CoV-2 infection. Our study revealed GA-1000 as one potentially inhibitory compound in these infection studies. Drug inhibition assays with GA-1000 revealed its inhibitory effect on SARS-CoV-2 Spike directed infections. While IC50 values did not reflect concentrations used in EGM2, the variability seen in this assay under different MOIs highlights the complexity of tailoring cell culture conditions for infection studies. The importance of characterization of standardized media is that it enables scalable experiments and reproducible results. Overall, this research provides a framework for considering aspects of cell culture media to study infection in biologically relevant conditions without sacrificing reproducibility.

Data availability

Data will be made available on request.