Quantitative and Functional Assessment of the Influence of Routinely Used Cryopreservation Media on Mononuclear Leukocytes for Medical Research

Quantitative and functional analysis of mononuclear leukocyte populations is an invaluable tool to understand the role of the immune system in the pathogenesis of a disease. Cryopreservation of mononuclear cells (MNCs) is routinely used to guarantee similar experimental conditions. Immune cells react differently to cryopreservation, and populations and functions of immune cells change during the process of freeze–thawing. To allow for a setup that preserves cell number and function optimally, we tested four different cryopreservation media. MNCs from 15 human individuals were analyzed. Before freezing and after thawing, the distribution of leukocytes was quantified by flow cytometry. Cultured cells were stimulated using lipopolysaccharide, and their immune response was quantified by flow cytometry, quantitative polymerase chain reaction (qPCR), and enzyme-linked immunosorbent assay (ELISA). Ultimately, the performance of the cryopreservation media was ranked. Cell recovery and viability were different between the media. Cryopreservation led to changes in the relative number of monocytes, T cells, B cells, and their subsets. The inflammatory response of MNCs was altered by cryopreservation, enhancing the basal production of inflammatory cytokines. Different cryopreservation media induce biases, which needs to be considered when designing a study relying on cryopreservation. Here, we provide an overview of four different cryopreservation media for choosing the optimal medium for a specific task.


Introduction
The analysis of human mononuclear cells (MNCs), which comprise circulating monocytes and lymphocytes, has become an invaluable tool in medical research and nowadays even in experimental clinical cell-therapies including autologous dendritic cell and chimeric antigen receptor T (CAR-T) cell immunotherapy [1][2][3][4]. Further, the absolute amounts and the relative composition of individual leukocyte populations can be used as reliable biomarkers and predictors in different (inflammatory) diseases, i.e., an increase of circulating CD14 ++ CD16 + CCR2 + intermediate monocytes (Mon2) was found to be independently predictive for cardiovascular events or outcome in multiple cardiovascular diseases [5][6][7][8]. Functional immunological assessment of MNCs ex vivo has further important implications in basic [9] as well as in therapeutic immunological research, as has been shown, for 2 of 20 example, in the assessment of vaccine-induced cell-based immune responses during the COVID-19 pandemic [10,11].
Assessment of MNCs in research laboratories in the context of clinical trials requires personal and technical resources, i.e., instruments and standardized protocols for flow cytometry [12] and the respective technicians processing and analyzing the samples in an appropriate time. Especially in multi-centric studies, not all centers perform the laboratory analysis of their samples themselves. These pre-processing and pre-analytic steps challenge a standardized analysis, which is required to draw solid scientific conclusions [13]. Cryopreservation of MNCs has become an important method to overcome these issues, allowing transportation and batch-wise analysis of cells, which reduces inter-laboratory and intra-assay biases [14,15]. However, cryopreservation induces various quantitative and functional alterations in different immune cell populations, which needs to be considered when studies using cryopreserved MNCs are interpreted [16][17][18][19]. Due to these observations, the European Society of Cardiology recommends the use of only fresh samples for monocyte subset analysis [20].
In this study, we aimed to compare the effects of four routinely used cryopreservation media [CryoStor CS10 (Stemcell, Vancouver, Canada); Synth-a-Freeze (ThermoFisher, Waltham, MA, USA); 90% FBS + 10% DMSO (self-prepared); 70% RPMI1640 + 20% FBS + 10% DMSO (self-prepared)] on the recovery/viability and the amount/function of different leukocyte populations in detail. Therefore, leukocyte populations were quantified by flow cytometry before cryopreservation as well as directly after and after 1-h resting period at 37 • C post thawing. Further, the immunological response of the MNCs was assessed by qPCR, ELISA, and flow cytometry of cytokines and activation markers.
The choice of used cryopreservation medium affected the obtained recovery and viability after cryopreservation. Further, we were able to verify previously reported findings on relative alterations in T cell and monocyte subsets. Freeze-thawing increased the basal production of pro-inflammatory activation markers and cytokines in all cryopreservation media. By stimulating the toll-like receptor 4 (TLR4) -competent cell compartment of MNCs with lipopolysaccharide, we were able to show that the used cryopreservation medium changed the immune response significantly. Further, the expression of activated CD11b on the surface of monocytes was affected by the used cryopreservation medium. To summarize our findings, we ranked the different cryopreservation media based on the differences in conditions before freezing and the observed standard deviation for each experimental aspect and created an overview table to aid other research groups in selecting an appropriate cryopreservation medium for their future studies.

The Choice of Cryopreservation Medium Affects Recovery and Viability of MNCs after Thawing
Four cryopreservation media (CryoStor CS10™ = CS; Synth-a-Freeze™ = SF; 90%FBS/10%DMSO = FBS; 70%RPMI/20%FBS/10%DMSO = RPMI), which are routinely used in medical research studies, were examined for their effects on MNCs. MNCs were isolated from leukocyte reduction chambers, which were stored overnight at 4 • C before isolation. Overnight storage did not alter the viability relative distribution and activatability ( Figure S1). The experimental workflow for sample processing in this study is given in Figure 1A. To determine whether there were differences in recovery and viability between the individual cryopreservation media, MNCs were counted after thawing on a LUNA Automated Cell Counter using 0.4% Trypan Blue Solution to assess viability before they were further processed for flow cytometry and cell culture experiments.  Figure 1B).   showed the lowest viability in this validation experiment ( Figure 1D). Taken together, we found significant differences between the individual cryopreservation media. CS and FBS seem to be superior to RPMI and SF in terms of cell recovery and viability.

Cryopreservation Alters the Relative Distribution of Leukocyte Populations
Next, we asked how the different cryopreservation media would affect the relative distribution of leukocyte populations after freeze-thawing, as it was shown before that cryopreservation can affect individual populations differently [21]. Therefore, MNCs were stained with three antibody panels and analyzed by flow cytometry to quantify the major leukocyte populations (gating is depicted in Figure S2). After cryopreservation, the relative number of monocytes and T cells compared to all CD45+ leukocytes, was drastically altered ( Figure 2A). We observed increased amounts of monocytes after thawing in all used media. CS (28.5% [21.7-36. Next, we wanted to know whether the measured counts of the individual leukocyte populations would correlate before and after cryopreservation, which would be important for longitudinal studies comparing changes in individual patients rather than changes in the means of whole populations ( Figure 2B). The individual linear regressions for each leukocyte population are given in Figure S3A. For monocytes, the overall correlation before and after freezing was very weak (R 2 values: CS 0.2675, SF 0.2564, FBS 0.1760, RPMI 0.1633) and only significant for CS-stored monocytes. T cells (R 2 values: CS 0.7589, SF 0.7543, FBS 0.6852, RPMI 0.7203) and NK cells (R 2 values: CS 0.7308, SF 0.5601, FBS 0.5890, RPMI 0.6731) correlated significantly in all cryopreservation media, with CS showing steeper linear regression lines than the other cryopreservation media. Importantly and unpredicted, B cell amounts before and after cryopreservation did not correlate at all in any medium. As recovery and viability were different between the cryopreservation media, we also assessed the absolute amount of dead cells for each leukocyte population. As shown in Figure 2B,C and NK cells were less affected by cryopreservation, with significant differences before and after cryopreservation only found for SF and RPMI. A significant number of monocytes and T cells underwent cell death after thawing compared to fresh samples (0.2% [0. 13- . In summary, cryopreservation affected leukocyte populations differently, with some media performing better in maintaining the relative distribution of the MNC populations before and after freeze-thawing. Further, correlation before and after cryopreservation was very different between the individual leukocyte populations.

Analysis of the Effect of Cryopreservation Media on the Relative Frequency of Individual Leukocyte Subsets
Often, investigators are only interested in distinct leukocyte subsets such as CD4+ T cells in HIV [22] or relative subset distributions such as monocyte subsets in critically ill patients [23]. We analyzed the distribution of human monocyte subsets (classical CD14++CD16-= Mon1, intermediate CD14++CD16+ = Mon2 and non-classical CD14+CD16++ = Mon3) before and after cryopreservation. As shown above and similar to

Analysis of the Effect of Cryopreservation Media on the Relative Frequency of Individual Leukocyte Subsets
Often, investigators are only interested in distinct leukocyte subsets such as CD4+ T cells in HIV [22] or relative subset distributions such as monocyte subsets in critically ill patients [23]. We analyzed the distribution of human monocyte subsets (classical CD14++CD16-= Mon1, intermediate CD14++CD16+ = Mon2 and non-classical CD14+CD16++ = Mon3) before and after cryopreservation. As shown above and similar to published data [16], the relative amount of monocytes in the CD45+ leukocyte compartment was significantly increased due to the death of other leukocyte populations. Interestingly, we were able to show that this increase was not evenly dispersed over all monocyte subsets. Cryopreservation led to a significant increase in total classical (Fresh 10.7% [8.8-17 Figure 3A). When we compared the relative subset distribution, we found a significant increase in classical monocytes ( Figure 3B). By calculating the ratio of the relative change of the individual subset to the relative change of all monocytes overall, we were able to show that cryopreservation led to an overrepresentation of classical monocytes and a loss of non-classical ones ( Figure 3C). These alterations were significantly less prominent in samples preserved in CS than in other cryopreservation media. We correlated the relative fraction of the individual subsets before and after cryopreservation.  Figure 3D). The individual linear regressions for each monocyte subset are given in Figure S3B.
Next, we focused on the different T cell subsets. We did not observe significant differences in the relative amount of CD8+ ( Figure 3E). These observations were also evident when we compared the correlations of the individual subsets. R 2 values were prominently lower for DP+ T cells and NKT cells than for the other subsets ( Figure 3F). Loss of CD28 expression is associated with aging [24], and increased amounts of CD28-T cells were shown to be associated with the development of atrial fibrillation after cardiac surgery as well as with increased mortality in patients with known atrial fibrillation [25,26]. When we compared the effects of different cryopreservation media with the absolute amount and relative fraction of CD8+ and CD4+ T cells lacking CD28 expression (CD28-), we only observed small changes in the relative fraction of the CD4 and CD8 subsets in samples frozen in media containing FBS ( Figure 3G). These findings show that cryopreservation does affect individual leukocyte subsets to a varying degree, which needs to be considered when studies using cryopreserved leukocyte populations are compared to ones using fresh MNCs.

A 1-h Resting Period after Thawing Does Not Dramatically Alter Leukocyte Population Amounts
Lemieux et al. showed that a resting period of 1 h at 37 • C in an incubator after cryopreservation could refine the expression of some surface markers, improving the relative distribution of leukocyte populations after thawing [19]. Therefore, we assessed a part of the cells after a 1-h resting period. However, in our experimental setup, we did not observe such prominent differences between unrested and rested MNCs.
Compared to the analysis directly after thawing, B cell amounts were not significantly increased after a 1-h resting period (CS 8.1% For monocytes and T cells, we did not observe any significant differences comparing a resting period of 1 h to no resting ( Figure 4A). Although the amount of total monocytes did not differ between unrested or rested MNCs, we found significant changes in the relative amounts of the individual subsets for CS, SF, and FBS. The relative proportion of classical monocytes was further significantly increased compared to unrested conditions ( Figure 4B). These changes were not observed for MNCs stored in RPMI. In summary, a 1-h resting period at 37 • C in an incubator led to significant changes only in the amount of B cells and the relative distribution of the individual monocyte subsets in our experiments.

Changes in the Immunological Response to Bacterial Lipopolysaccharide after Storage in Different Cryopreservation Media
Cryopreservation does not only affect the survival and relative distribution of different leukocyte populations, it can also interfere with the cells' ability to respond to different inflammatory stimuli needed to fulfil their role in clearing altered or infected cells such as in modern chimeric antigen receptor (CAR) T cell therapy [27]. Therefore, we evaluated the effects of cryopreservation of MNC samples on the expression of different inflammatory cytokines and compared the alterations to samples before freeze-thawing. LPS was used to activate cellular inflammatory pathways in TLR4expressing cells, evaluating the effects of different cryopreservation media on the reactivity of these MNCs. First, we measured mRNA levels of the four inflammatory cytokines interferon γ (IFNG), interleukin-1α (IL1A), interleukin-6 (IL6), and tumor-necrosis factor α (TNFA). MNCs showed an increase in the unstimulated transcription of IFNG after cryopreservation compared to the fresh condition ( Figure 5A).   For IL1A, we also observed a significantly increased basal production after cryopreservation compared to fresh levels (CS 3.0× [1. 7-6.  For IL1A, we also observed a significantly increased basal production after cryopreservation compared to fresh levels (CS 3.0× [1. 7-6. Figure 5D).
We aimed to verify our findings on protein level after 24 h of stimulation. Basal IL-6 production was indeed increased after thawing in unstimulated MNCs ( Figure 5E). Although TNFA expression was higher after thawing at a transcriptional level, basal TNF-α protein production was not significantly increased. LPS stimulation led to a significant increased production of TNF-α under all conditions and further, cryopreserved MNCs produced significantly higher amounts compared to fresh ones (Fresh 1049 pg/mL [543.  Figure 5F). Thus, although MNCs preserved their ability to react to inflammatory stimuli after freeze-thawing, cryopreservation of MNCs can affect the inflammatory readouts observed. This impacts the usage of MNCs for functional assessment after cryopreservation, as the cellular composition and different pathways and cellular functions seem to be affected unequally.

Cryopreservation Affects the Activatability of the CD11b Surface Integrin
Bacterial LPS also induces the expression of CD11b in myeloid immune cells, leading to increased adhesion and further boosting TLR4 induced LPS signaling via a positive feedback loop in microbial infections [28,29]. Stimulation of monocytes in vitro using LPS alters the relative amount of different monocyte subsets by inducing shedding of membrane receptors used to identify CD16+ monocytes through A disintegrin and metalloproteinase 17 (ADAM17) [30]. Comparing the relative distribution of the individual monocyte subsets between unstimulated and LPS-stimulated samples under fresh and cryopreserved conditions in our study, we found that the relative amount of non-classical CD16 + CCR2 − monocytes was diminished significantly, whereas the fraction of classical monocytes was increased significantly ( Figure S4A).
Previously, it has been shown that the individual monocyte subsets express different amounts of activated CD11b and that LPS mainly induces the expression of CD11b in intermediate monocytes [31]. We evaluated the effect of different cryopreservation media on the activatability of the surface integrin molecule CD11b by flow cytometry. MNCs were stimulated for 4 h with 100 ng/mL LPS, and the CD11b expression was analyzed by measuring the median fluorescence intensity in unstimulated/stimulated monocyte subsets before and after cryopreservation ( Figure 6A Figure 6G). Taken together, these observations show that the activatability of CD11b upon LPS stimulation was significantly reduced in MNCs cryopreserved in FBS or RPMI and also significantly altered in MNCs stored in SF, although the relative distribution of CD11b expression between individual monocyte subsets remained intact after cryopreservation.

Comparison of the Individual Cryopreservation Media for Different Experimental Research Questions
To give an overview of the performance of the individual cryopreservation media in this study, we created a table (Table S1) comparing the means and coefficients of variation between fresh and the individual cryopreservation media for all individual experiments. The media were ranked by their differences in the mean (deltas) and by their dispersion (coefficients of variation), allowing the selection of the appropriate cryopreservation medium for an individual research question (see Table 1).     Cumulative performance (overall rank) 1 3 2 4

Discussion
In clinical studies, the standardized collection, handling, cryopreservation, and thawing of MNCs has become a widely used and accepted method to systematically investigate the role of different immune cells in health and disease. The objective of this study was to compare the influence of four different cryopreservation media on the amount and function of MNCs, as there have been reports suggesting that cryopreservation can influence cell populations and their function [32,33].
The choice of cryopreservation medium affected the recovery and viability of MNCs after thawing in our study. The absolute recovery was higher when MNCs were stored in CS or FBS compared to samples frozen in SF or RPMI. This was paired with a higher viability of the recovered leukocytes, as shown by permeability assessment using Trypan blue or SytoxGreen. Although all media were sufficient to obtain live MNCs after thawing, the formulations of the individual media impacted the maximum gain. Similar observations were also found when others investigated changes in formulations for freeze-thawing of progenitor and stem cells [34][35][36].
Similar to prior published works [37,38], cryopreservation introduced quantitative changes in the distribution of different leukocyte populations in this study. In general, T cells were more likely to die during the cryopreservation process than other leukocyte populations, as the relative amount of CD3+ T cells of all recovered leukocytes was significantly reduced. In contrast, monocytes better withstood the procedure, as there was a relative overrepresentation of monocytes compared to all CD45+ cells after cryopreservation compared to the distribution before freezing. The unequal recovery rates between monocytes and T cells could be due to different susceptibilities of the individual leukocyte populations to the used cryoprotectants, as, for example, increasing dosages of DMSO led to significantly reduced recovery of Treg cells [39]. Interestingly, this increase was not evenly dispersed over all monocyte subsets, as the relative frequency of "younger" classical monocytes was increased and the amount of "aged" non-classical monocytes [40] was significantly reduced after cryopreservation. Although the respective changes were observed for all cryopreservation media, MNCs stored in CS showed the least significant changes among all cryopreservation media, and correlations of fresh and cryopreserved values were highest for this cryopreservation medium. These findings need to be consid-ered in interpreting clinical studies investigating the association of circulating monocyte subsets in patients, as CD16+ monocytes were shown to correlate with disease severity and the occurrence of primary endpoints [41,42]. The unreflected usage of cryopreserved MNCs for such analyses could introduce biases, potentially leading to false associations. Overall, our findings suggest that only T cell analysis can be performed on frozen MNCs with an adequate representation of the initial distribution, as all other cell populations failed to show good correlations between cell numbers before and after freezing for each individual donor. It should be noted that due to the fact that for this study NK cells were defined as CD45+CD3-CD56+ and gating was done very restrictively on higher CD56 expression, the amount of NK cells in our study was somewhat lower than that observed in other studies that focused in more detail on NK cell subpopulations. Due to missing CD16 staining in the analysis panel, we were not able to study the CD45+CD3-CD16+CD56-NK cell population [43].
As a 1-h resting period led to a recovery of the relative distribution of CD14+ cells in the study of Lemieux et al., we also evaluated the effect of a 1-h resting period in full MNC medium on the relative distribution of different cell populations in our study [19]. However, we only observed a normalization of B cells to amounts before cryopreservation. In contrast, a resting time of 1 h in cell suspension further diminished the relative amount of intermediate and non-classical monocytes in our analysis.
It has also been stated that cryopreservation alters the capacity of MNCs to react to different stimuli when compared to fresh samples [16,27,44]. We therefore asked whether the choice of cryopreservation medium would affect the responsiveness of the MNCs to LPS and if one medium would perform better than another. We analyzed the expression and production of different cytokines at basal levels and upon LPS stimulation. LPS stimulation led to a significant increase in the production of inflammatory cytokines compared to basal values. Cryopreserved MNC samples showed an increased expression of all inflammatory cytokines compared to fresh samples. Fold-wise, this basal change ranged from a 3-up to a 100-fold change compared to fresh samples. As total 18S RNA was used to normalize the quantification data and cytokine production was probably driven mainly by myeloid cell populations in this experiment, the altered monocyte-lymphocyte ratios observed after freeze-thawing have to be taken into account too with respect to biasing the observed basal change. These findings suggest that basal inflammatory activity could be overestimated in study cohorts using cryopreserved MNCs. When analyzing fold inductions, the response profiles of the individual cytokines need to be considered, as fold changes were different between the individual cytokines determined. Interestingly, when we evaluated the response of monocyte subsets to express activated CD11b on their surface upon LPS stimulation, we observed profound differences between the individual media. Monocytes stored in medium containing FBS (FBS and RPMI groups) showed diminished expression of activated CD11b, and the induction of CD11b upon LPS was significantly lower in FBS-and RPMI-stored MNCs, with only CS-stored cells being non-significantly changed compared to fresh samples.
To conclude, we report that choosing a particular cryopreservation medium can have dramatic impacts on the obtained readouts and can introduce alterations in the recovery, survival, and distribution of different immune cell populations after thawing. For all cell types studied, alterations induced by freeze-thawing need to be considered when different studies or experimental samples are compared. We summarize and rank the observed changes for different experimental setups and for individual cell populations in an overview table, aiding other researchers in selecting an appropriate cryopreservation medium.

Study Samples
Buffy coats were obtained from leukocyte reduction system chambers (LRSC) obtained by single platelet apheresis of 15 healthy individuals of both sexes. Due to technical reasons, LRSCs used to isolate buffy coats in this study were stored overnight at 4 • C before processing. The advantage of leukapheresis compared to whole-blood samples is the large yield and fast isolation of mononuclear cells, allowing one to test for largescale experimental research setups [45]. All donors gave their written informed consent before the leukocyte reduction chambers were used for MNC isolation. The study was performed after approval by and according to recommendations of the ethical committee of the Medical University of Vienna (EK 1575/2014).

MNC Isolation
MNCs were collected by density gradient centrifugation as published before [46]; please see supplemental information. In short, the content of the leukocyte reduction chambers was diluted with phosphate-buffered saline (PBS, Sigma-Aldrich, St. Louis, MO, USA); 15 mL of Lymphocyte Separation Medium 1077 (Promocell, Heidelberg, Germany) was added to the lower chamber of SepMate-50 Tubes (Stemcell, Vancouver, Canada), and the diluted content of the leukocyte reduction chambers was added on top of the inner barrier. The special insert of these tubes allows density gradient centrifugation at 1200 g for 10 min with full break, significantly reducing the processing time of MNC isolation. The buffy coat was washed once with PBS and resuspended at 2 × 10 7 cells/mL in full MNC medium [RPMI1640 (Sigma-Aldrich, St. Louis, MO, USA) supplemented with 10% fetal bovine serum (FBS Gold; Seraglob, Switzerland), glutamine, and antibiotics (both Lonza, Basel, Switzerland)].

Cryopreservation
After an aliquot was taken to perform experiments on fresh MNCs, the remaining MNCs were split into four aliquots of 5 × 10 7 cells each for the different cryopreservation media. Freezing was performed serially for each aliquot to ensure equal stay times as cell pellets and in cryopreservation medium before freezing. Four widely used cryopreservation media were assessed, two commercially available ones and two self-prepared ones: (1) CryoStor CS10 Cell Freezing Medium (Stemcell, Vancouver, Canada); (2) Synth-a-Freeze Cryopreservation Medium (ThermoFisher, Waltham, MA, USA); (3) 90% FBS Gold (Seraglob) + 10% dimethyl sulfoxide (DMSO, >99.9% purity, Sigma-Aldrich, St. Louis, MO, USA); (4) 70% RPMI1640 + 20% FBS Gold + 10% DMSO. FBS batch-to-batch comparability was assured by usage of FBS Gold, which is a defined fetal bovine serum composed of chromatographic isolated and re-supplemented components. The MNCs were centrifuged at 500× g for 5 min and resuspended on ice in the respective cryopreservation medium at a concentration of 2 × 10 7 cells/mL. The cell suspension was split into five aliquots of 1 × 10 7 cells each, transferred into 2 mL Bio-One Cryo.s Tubes (Greiner, Kremsmuenster, Austria) and placed at −80 • C into a CoolCell reduced-scale freezing container (Corning, Corning, NY, USA), cooling the MNCs at a rate of −1 • C/min. Afterwards, the cryo-tubes were transferred into liquid nitrogen tanks and stored there until thawing.
After 6-10 weeks, the MNCs were thawed again by placing the cryo-tubes at room temperature until the outer layer of the crystal became liquid. Then, the crystal containing the MNCs was flipped into ice-cold full MNC medium. After all of the cryopreserved medium became liquid, the MNCs were washed once and resuspended in 1 mL pre-warmed full MNC medium for the assessment of recovery and viability.

Measurement of Absolute Cell Counts and Viability after Thawing
Recovery and viability of MNCs after thawing were assessed using the LUNA Automated Cell Counter (Logos Biosystems, Anyang, South Korea) together with 0.4% Trypan Blue Solution (Lonza, Basel, Switzerland). The cell suspensions were adjusted to 2 × 10 7 viable cells/mL, and the concentration was verified using a XN-350 Hematology Analyzer (Sysmex, Vienna, Austria).

In Vitro MNC Stimulation
Cell function prior to and after thawing was assessed by stimulation of MNCs with the toll-like receptor 4 agonist lipopolysaccharide (LPS, from Escherichia coli strain O111:B4, Sigma-Aldrich, St. Louis, MO, USA). MNCs were seeded at a concentration of 2 × 10 5 cells/400 µL full MNC medium and were left unstimulated or were stimulated with 100 ng / mL LPS either for 4 h for all gene expression experiments or for 24 h to collect the condition culture supernatant. For experiments involving a 1-h resting period, cells in full MNC medium were incubated for 60 min under standard cell culture conditions (humidified atmosphere, 5% CO 2 , 37 • C) before analysis and stimulation. After the indicated times, cells were harvested and conditioned culture medium was collected by taking the supernatant. Cell-free supernatant was generated by centrifugation at 500× g for 5 min.