Neutrophil stages and the identification method
Mature, healthy (non-activated) neutrophils take a unique multilobed-shaped nucleus, which gives the cells high malleability and motility16. When exposed to tissue injury or infection in vivo, neutrophils are activated, responding to the lesion in different types of cell death (Fig. 1) - primarily apoptosis and NETosis7,17. Based on the mechanism of cell death, the shape of the neutrophil nuclei is distinguishably different. During apoptosis the chromatin condenses while during NETosis the chromatin decondenses18. Apoptosis and NETosis can be distinguished by analyzing the shape of nuclei with a membrane-permeable DNA stain (Hoechst)19. NETosis is known to have two distinct subtypes - suicidal NETosis (causing cell death) and vital NETosis (no immediate cell death)7. Suicidal NETosis features ruptured cytoplasmic membrane and release of diffused NETs. By contrast, vital NETosis features intact cytoplasmic membrane and release of stringy NETs. Release of NETs can be visualized by membrane-impermeable DNA stains (e.g., Sytox or propidium iodide)19. The rupture of cytoplasma membrane (in apoptosis and suicidal NETosis) can be visualized by phosphatidylserine stain (Annexin V)18.
In this work, we adopted Hoechst-Sytox-Annexin V (H-S-A) as a three-dye method. As summarized in Table 1, i) Hoechst is a DNA stain that penetrates intact plasma/nuclear membranes in presence of trace amounts of dimethyl sulfoxide (DMSO). Hoechst stains DNA of neutrophils during all stages, including healthy cells, early apoptosis, sec. necrosis, suicidal NETosis, and vital NETosis. It’s worth noting that Hoechst shows little staining on NETs (decondensed chromatin) in the presence of Sytox, likely due to the competition between the two dyes and the dissociated structure of the chromatin. Hoescht staining is normal on condensed nuclei in sec. necrosis when co-stained with Sytox. Hoechst intercalates into the double strands of DNA in living cells, which affects many aspects of cell function. To this end, Hoechst was only used for end-point staining (Supplementary Fig. 1). ii) Sytox Green is a DNA stain that is impermeable to intact plasma/nuclear membranes. When the plasma/nuclear membrane is compromised/ruptured, Sytox stains any DNA content exposed to the dye. Sytox stains the condensed chromatin of neutrophils in sec. necrosis, and NETs from neutrophils that undergo NETosis. iii) Annexin V stains phosphatidylserine (or the “Eat Me” molecules) on the plasma membrane of dying or dead cells and thus it stains neutrophils in early apoptosis, sec. necrosis, and suicidal NETosis. Sytox and Annexin V were used for real-time staining (i.e., the dye was added to the cell stock before cell seeding and present through a culture) to monitor neutrophil kinetics in culture (Supplementary Fig. 1).
Next, we optimized the H-S-A staining protocol to minimize the potential effects that fluorescence labeling has on normal cell function21. Here, we tested three different concentrations of the Sytox dye and the influence of Sytox and Annexin V on each other (Supplementary Fig. 4). For Sytox, 10000× and 5000× dilution made no significant difference on non-specific activation and neutrophil kinetics compared to the no-dye control (Supplementary Fig. 5). Sytox at 1000× dilution caused higher activation especially after 6 h from the initiation of cell culture (Supplementary Fig. 4a, b). The combined use of Annexin V (20× dilution) with Sytox (5000× dilution) led to an increased Sytox signal but interestingly, there was no difference in the Annexin V signal compared to the Annexin V only control group (Supplementary Fig. 4c, d). To minimize the artifact from real-time staining (Supplementary Fig. 1a, b), we chose to use 10000× Sytox with 40× Annexin V. Hoechst was only used for end-point staining to avoid any artifacts in cell function and prevent DNA damage22.
Isolation method
Different isolation methods have been developed and used for isolation of primary neutrophils from whole blood (e.g., gradient neutrophil isolation with RBC sedimentation or lysis, negative selection-based neutrophil isolation with or without RBC lysis). Previous work has shown that different isolation methods can affect neutrophil phenotype and activity in functional assays; therefore, we explored two commonly used methods to identify a method best suited for this study23. Here, we compared two negative selection kits [MACSxpress (with RBC lysis) versus EasySep (without RBC lysis) following the manufacturer’s instructions] (Supplementary Fig. 6a). The results showed that the non-specific activation right after isolation was comparable between the two kits but noticeably MACSxpress based isolation led to accelerated kinetics in 2D culture of the neutrophils (Supplementary Fig. 6b, c). Therefore, we chose to use EasySep negative selection without RBC lysis in this work.
House-keeping factors in 2D culture of neutrophils
Different research groups may use a range of culture conditions in their studies, which likely introduces variability and impedes reproducibility/comparability of the results between studies. Here we investigate how the selection of culture conditions may influence non-specific activation and neutrophil kinetics in standard 2D monoculture ex vivo. Based on our previous experience with microscale cultures, we recognized the importance that media evaporation, incubator system, and seeding density may all play key roles in maintaining proper cell behavior.
Microscale cell culture methods possess many advantages over bulk-scale (e.g., petri dish, culture flask) methods due to their ability to better recapitulate the parameters of in vivo cellular microenvironments (e.g., structural hierarchy, fluidic dynamics, and mass transport). Due to the small volume (< 100 µl), media loss via evaporation becomes a significant variable that exerts stress to the cells in culture. Recently, under-oil open microfluidic systems (UOMS) were developed and introduced to microscale cell culture20, 24–27. UOMS enables several unique functions in microscale cell culture including i) free physical access to the culture system with minimized evaporation and sample contamination24,25,28, ii) versatile open fluidic controls26, and iii) autonomously regulated oxygen microenvironments (AROM)20. We tested the isolated neutrophils in standard 2D monoculture from nine donors with and without oil overlay (Supplementary Fig. 7a). In five of the nine randomly selected donors, oil (especially silicone oil) overlay led to less non-specific activation compared to the no-oil culture, as measured by the area fraction of the Sytox signal at the end of the stationary phase (Supplementary Fig. 7b, see the definition in Fig. 3a). For the remaining four donors, silicone oil overlay did not significantly affect the activation status of the neutrophils. There was no obvious pattern in which donor neutrophils would benefit from the silicone oil overlay. Therefore, for subsequent experiments we used a silicone oil (coded in tan color in the figures) overlay in 2D culture to get improved culture environment stability and consistency.
Next, we compared two incubator systems - a standard incubator (SI) versus an onstage incubator (OI) (Supplementary Fig. 8a). They all allow the operators to select and control the O2 level, CO2 level, temperature, and RH. Compared to a standard incubator, an onstage incubator allows optical access to the culture system on a microscope to enable real-time imaging. We interrogated two randomly selected donor neutrophils to study neutrophil activation and kinetics between these incubator conditions. Culture in the onstage incubator led to dramatically accelerated kinetics (Supplementary Fig. 8b) compared to the standard incubator. The incubation phase (see the definition in Fig. 3a) was dramatically shorter in the onstage incubator for both donors (~ 5 h) and reached the stationary phase (~ 25 h) much quicker than their standard incubator counterparts (~ 50 h) (Supplementary Fig. 8b). Therefore, we chose to use a standard incubator in this work. For experimental workflows, the culture plate was stored in a standard incubator and was only taken out for imaging in the onstage incubator.
Finally, we investigated the possible influence of cell seeding density on neutrophil kinetics. In a standard 384-well plate, 60000 (or 60K) cells per well achieves a near-confluent monolayer of neutrophils (Supplementary Fig. 9a). Neutrophil seeding densities across a broad range (10K, 30K, 60K, and 200K per well) were trialed. When seeded at 10K or 30K cells per well, neutrophils did not show any difference in normalized activation (Supplementary Fig. 9b). The highest seeding density (200K cells per well) showed higher total activation (~ 2-fold higher) compared to the 60K cells per well condition but lower activation (~ 60%) after normalization to per 10K cells (Supplementary Fig. 9c). Overall, the near and less confluent seeding densities showed no significant influence on non-specific activation of neutrophils and the seeding density of 60K cells per well allowed more even and consistent distribution of the cells in a well from cell seeding and thus enabled easier/more reliable data collection and analysis.
Statistics of whole blood handling time
To study human neutrophils ex vivo, the typical workflow includes blood logistics (i.e., collection and transport), storage, processing, and interrogation (e.g., cell culture and assessment) (Fig. 2a). Each additional step before interrogation in the workflow introduces possible bottlenecks and delays that can contribute to the overall quality of the isolated cells used in the terminal assay. We evaluated the components of whole blood handling to identify key parameters that influence baseline neutrophil apoptosis and NETosis, defined as “non-specific activation” after transfer to standard 2D cell culture. In the 2D culture, neutrophils were cultured in a standard incubator with and without oil overlays as described above (Fig. 2a). Blood was obtained from multiple sources including draws completed in our laboratory and through a collaborator’s lab at the UW hospital (under an IRB approved protocol, see Methods). Across 36 unique blood donors, our results highlight a highly variable whole blood logistic time (tl) and whole blood processing time (tp) from case to case (Fig. 2b). tl ranged from 0.5 to 5 h depending on the efficiency of donor ID de-identification, the distance/communication between the blood collection locus and the lab performing the subsequent experiments (Fig. 2b). Similarly, tp ranged from 1 to 4.5 h depending on how many whole blood samples processed in an experiment and how many conditions in the following interrogation (Fig. 2b). On average, it took 4–5 h from blood collection to the start of neutrophil interrogation if no whole blood storage time (ts) was added to the “dead time” (i.e., tl + ts + tp) in this work performed by a single operator.
Short-time (< 8 h) whole blood storage
Clinical studies often require processing of blood samples but can be delayed due to the logistics of transport to third-party entities, so there is a need to better understand if whole blood storage time and conditions can affect the baseline characteristics of isolated cells. Recent work identified that whole blood storage time of samples drawn from pigs led to differences in ROS production, the level of NET formation, altered neutrophil survival kinetics, and modified antimicrobial activity29. Here, we look into how whole blood storage time, temperature, and method (i.e., stationary storage versus rotary storage) affect apoptosis and NETosis from samples collected from healthy human donors and patients with cancer. The analysis was broken into two distinct time periods, < 8 h (i.e., short-time storage) and > 24 h (i.e., long-time storage). Early experiments indicated that the 8 h mark constituted an inflection point at which neutrophil behavior could begin to change and coincided with a traditional 8 h work shift.
Considering the varying dead time of whole blood handling, it is important to understand first how short-time whole blood storage (with ts <8 h) may influence non-specific activation and neutrophil kinetics in standard 2D monoculture ex vivo (Fig. 3). From a whole blood sample with tl <1 h, we partitioned the blood into three groups: immediate isolation (i.e., ts = 0 h) (Fig. 3a), stationary storage (at RT for 2.5 h) (Fig. 3b), and rotary storage (on a tube rotator at RT for 6.5 h) (Fig. 3c). Comparison between the stationary storage and rotary storage is limited by the 4 h time difference in storage time. This difference is attributed to practical limitations of a single operator carrying out all the tasks and material limitations (e.g., magnetic isolation kits) available to the operator. We further analyzed the kinetic profile of the isolated neutrophils after culture in the unique culture conditions. In standard 2D monoculture, the isolated neutrophils showed three distinct phases in kinetics, which include i) incubation phase (low activation, 0–12 h), ii) exponential phase (fast activation, 12–42 h), and iii) stationary phase (saturated activation, > 42 h) (Fig. 3a). Our results indicate short whole blood storage times (< 8 h) did not increase non-specific activation or significantly alter kinetics in 2D culture (Fig. 3).
Long-time (> 24 h) whole blood storage
Conversely, long-time whole blood storage times (ts >24 h) may influence non-specific activation of neutrophils (Fig. 4). Neutrophils stored at 37 oC, in a standard 2D monoculture environment, for greater than 24 h and up to 48 h showed increased levels of Sytox staining (Fig. 4a, b), which accounts for NETosis and secondary (sec.) necrosis (Fig. 1). In contrast, neutrophils stored in whole blood at 37 oC maintain relatively low levels of Sytox staining immediately after isolation, indicating low levels of non-specific activation (Fig. 4a, b). In all conditions, Annexin V staining, which measures apoptosis, sec. necrosis, and suicidal NETosis (Fig. 1), was not appreciably different. As cells undergo apoptosis and nuclear condensation occurs, the circularity of the multilobed neutrophil nucleus increases towards 1.00 (which indicates a round nucleus) (Supplementary Fig. 2). The automated nuclear circularity analysis pipeline (Methods) developed here quantified the differences in circularity between the different storage conditions at the specified time points. Neutrophils that were immediately isolated and kept in culture at RT had increased circularity (i.e., more apoptosis) over the subsequent 48 h of culture (Fig. 4c, d). Neutrophils stored in whole blood at 37 oC and subsequently isolated at the 24 h and 48 h time points had significantly less circularity (i.e., more healthy neutrophils with multilobed nucleus) when compared to their RT counterparts (Fig. 4c, d).
Next, we investigated the kinetic profiles of neutrophils in the different storage conditions and length of storage. For these studies, each whole blood sample, from three random donors (Prostate383, Prostate579, Healthy103), was partitioned into four groups - immediate isolation, and whole blood storage for 24 h, 48 h, and 72 h. Neutrophils in the immediate isolation condition were placed in 2D culture and incubated in RPMI + 10% FBS + 1% P/S (Fig. 4f Prostate383-i, Prostate579-i, Healthy103-i). These neutrophils were monitored for the subsequent 72 h to assess their kinetic profile. In two of the donors (Prostate383, Prostate579), the neutrophils stored in whole blood for 24 h were subsequently isolated and placed into 2D culture as described above for 72 h (Fig. 4f Prostate383-ii, Prostate579-ii) to assess their kinetic profiles. For the third donor (Healthy103), neutrophils stored in whole blood for 24 h, 48 h, and 72 h were subsequently isolated and placed into 2D culture as described above (Fig. 4f Healthy103-ii, Healthy103-iii, Healthy103-iv). For each donor, the circularity of nuclei of the neutrophils was measured to assess apoptosis (Fig. 4g) across the different time points and storage conditions. Neutrophils stored in whole blood for 24 h prior to isolation showed accelerated kinetics (Fig. 4f Prostate383-ii, Prostate579-ii, Healthy103-ii) compared to the immediate isolation control (Fig. 4f Prostate383-i, Prostate579-i, Healthy103-i) as their 0 h time point showed significantly higher levels of non-specific activation (~ 10% vs 0%). Furthermore, neutrophils from the third donor (Healthy103) that were placed into culture at 48 h and 72 h post-whole blood storage showed even greater levels of kinetic acceleration, since they reached the stationary phase of their activation profile early on in 2D culture (Fig. 4f Healthy103-iii, Healthy103-iv). However, neutrophils stored in whole blood for 24 h had lower levels of apoptosis, as measured by nuclear circularity, compared to the 24 h time point for the immediate isolation condition for all three donors (Fig. 4g). For the third donor, the neutrophils stored in whole blood for 24 h, 48 h, and 72 h all show less apoptosis compared to the 24 h time point of the immediate isolation condition (Fig. 4g). For the third donor, we compared the level of activation of the neutrophils in each storage condition with respect to total time after isolation instead of time in 2D culture. Neutrophils in either 2D culture with RPMI + 10% FBS + 1% P/S or stored in whole blood showed similar levels of non-specific activation and apoptosis up till 24 h after isolation, but there were significantly lower levels of non-specific activation and apoptosis in the whole blood storage condition between 24 to 48 h time points (Fig. 4h). By 72 h, both storage conditions showed high levels of neutrophil non-specific activation and similar levels of apoptosis (Fig. 4h).
Whole blood storage temperature
Finally, we investigated the influence of whole blood storage temperature on non-specific activation and neutrophil kinetics. Across two donors, whole blood was partitioned into immediate isolation, and either storage at RT or 37 oC. Neutrophil activation and apoptosis were measured at 0 h, 2 h, 24 h, 48 h, and 72 h for the immediate isolation condition and at 24 h and 48 h for the RT and 37 oC storage conditions. For the second donor, technical limitations prevented us from capturing the 48 h time point in the two storage temperature conditions. In the immediate isolation condition, neutrophil activation increased over the 72 h of culture with a parallel increase in Annexin V staining, which measures early apoptosis, sec. necrosis, and suicidal NETosis. The four storage conditions have higher levels of Sytox and Annexin V staining at 0 h time point compared to the immediate isolation control, however, the neutrophils stored at 37 oC for 24 h and 48 h both had lower levels of Sytox staining compared to their matched RT storage conditions (Fig. 5a, b). Annexin staining between the immediate isolation condition and both storage conditions was not consistent between the two donors. For both donors, nuclei circularity analysis showed that there was significantly more apoptosis in the 37 oC storage condition compared to the RT storage condition (Fig. 5c).