Evaluating the Influence of Different Serum-Free Culture Conditions on the Production and Function of Natural Killer Cell-Derived Extracellular Vesicles

Abstract

Natural killer (NK) cells are exploited in cellular therapies for cancer. While NK cell therapies are efficient against haematological cancers, it has been difficult to target solid tumours due to low tumour infiltration and a hostile tumour microenvironment. NK-cell derived extracellular vesicles (NK-EVs) target and kill cancer cells in vitro and represent an alternative treatment strategy for solid tumours. To exploit their potential, it is necessary to standardize NK-EV production protocols. Here, we have performed a comparative analysis of EVs from the human NK-92 cell line cultured in five serum-free commercial media optimized for growth of human NK cells and one serum-free medium for growth of lymphocytes. The effect of growing the NK-92 cells in static cell cultures versus shaking flasks was compared. EVs were purified via ultracentrifugation followed by size-exclusion chromatography. We found that there were no significant differences in EV yield from NK-92 cells grown under static or dynamic conditions. However, we found clear differences between the different culture media in terms of EV purity as assessed by the enrichment of the CD63 and CD81 markers in the isolates that translated into their capacity to induce apoptosis of the colon cancer cell line HCT 116. These findings will be instructive for the design of future production protocols for therapeutic NK-cell derived EVs.

Keywords: NK‐92 cells; cancer; extracellular vesicles; serum‐free; therapy.

FIGURE 1

NK cell characteristics after different serum‐free culture conditions. (A) Workflow for static and shaking NK‐92 cultures, and purification of EVs. (B) Viability of NK‐92 cells after 48 h of culture in serum‐free media under static or shaking conditions supplemented with IL‐15 was assessed by propidium iodide by flow cytometry. Control samples were NK‐92 cells grown in Advanced RPMI medium with 20% FBS. Data are presented as the mean ± SD of four separate experiments. Statistical analysis performed using the Friedman one‐way ANOVA test indicated no statistical difference between the different culture conditions. (C) Expression levels of CD56 on NK‐92 cells after 48 h of culture in serum‐free media under static or shaking conditions. One representative experiment of three.

FIGURE 2

EV isolate concentrations obtained after NK‐92 cell cultures in different serum‐free media conditions. (A) Particle size or (B) particle concentrations measured by NTA analysis of pre‐cleared culture media. Data are presented as the mean ± SD of four separate experiments. Statistical analysis performed using the Friedman one‐way ANOVA test indicated no statistical difference between the different culture conditions. Protein concentration of EV isolates obtained after ultracentrifugation (C) or after additional purification by SEC (D) from indicated NK‐92 culture conditions measured via NanoDrop spectrophotometer or BCA assay, respectively. Data are presented as the mean ± SD of four separate experiments. Statistical analysis was performed using the Friedman one‐way ANOVA test. (E)–(F) Protein concentration measurements of 30 SEC fractions obtained when purifying ultracentrifuged EV isolates from indicated cultures. One representative experiment of four. SEC, size‐exclusion chromatography.

FIGURE 3

TEM images of purified EVs derived from different serum‐free NK cell culture media. (A) TEM images of EVs obtained from static culture conditions. (B) TEM images of EVs obtained from shaking culture conditions. Scale bar 200 nm. Images are representative of 8–10 images from each sample.

FIGURE 4

EV phenotyping and functional analysis. (A) Equal amounts of EVs were captured on aldehyde/sulphate latex beads and stained with either isotype control antibody, anti‐CD63 or anti‐CD81 antibodies, and expression levels measured by flow cytometry. Empty beads were included as a negative control (ctr). Data are presented as the mean ± SD of four separate experiments. Statistical analysis performed using the Friedman one‐way ANOVA test indicated no statistical difference between the different culture conditions. (B) Western blot analysis using 40 µg of indicated EV isolates. Representative of three independent experiments. (C) Measurement of apoptosis in HCT 116 cancer cells after 36 h as detected with Caspase 3/7 green detection reagent, using 40 µg purified EVs from the indicated NK‐92 culture conditions. Data are presented as the mean ± SD of four separate experiments. Statistical analysis performed using the Friedman one‐way ANOVA test indicated no statistical difference between the different culture conditions.