Low-Density Granulocyte Contamination From Peripheral Blood Mononuclear Cells of Patients With Sepsis and How to Remove It – A Technical Report

Elucidating the mechanisms contributing to the dysregulated host response to infection as part of the syndrome is a current challenge in sepsis research. Peripheral blood mononuclear cells are widely used in immunological studies. Density gradient centrifugation, a common method, is of limited use for blood drawn from patients with sepsis. A significant number of low-density granulocytes co-purify contributing to low purity of isolated peripheral blood mononuclear cells. Whole blood anticoagulated with lithium heparin was drawn from patients with sepsis (n=14) and healthy volunteers (n=11). Immediately after drawing, the plasma fraction was removed and PBMC were isolated from the cellular fraction by density gradient centrifugation. Samples derived from patients with sepsis were subsequently incubated with cluster of differentiation 15 MicroBeads and granulocytes were depleted using magnetic-activated cell sorting. Core cellular functions as antigen presentation and cytokine secretion were analyzed in cells isolated from healthy volunteers (n=3) before and after depletion to confirm consistent functionality. We report here that depleting CD15+ cells after density gradient centrifugation is a feasible way to get rid of the low-density granulocyte contamination. Afterwards, the purity of isolated, functionally intact peripheral blood mononuclear cells is comparable to healthy volunteers. Information on the isolation purity and identification of the containing cell types are necessary for good comparability between different studies. Depletion of CD15+ cells after density gradient centrifugation is an easy but highly efficient way to gain a higher quality and more reliability in studies using peripheral blood mononuclear cells from septic patients without affecting the functionality of the cells.


INTRODUCTION
The currently valid sepsis definition, published in 2016, focuses on the dysregulated host response to infection leading to a lifethreatening organ dysfunction (1). Despite great efforts in recent years, many of the mechanisms contributing to this dysregulation are still not properly understood. A deeper insight into the underlying immunopathology is of utmost importance to develop new therapy strategies and consequently a targeted patient care (2). In addition, it is also necessary to determine sepsis sub-groups (endotypes) according to their clinical and outcome characteristics within the very heterogeneous syndrome (3). Currently, different sepsis endotypes are identified through clinical variables or blood gene expression profiles (4)(5)(6). Expanding this concept, a further characterization of these already identified endotypes based on immunological parameters as well as the identification of new endotypes based on such parameters is possible. This ensures that misperceptions made in the past based on trials conducted on unstratified patient cohorts are not repeated (7), but patients benefit from precision medicine.
Within this framework, a substantial amount of research is conducted using peripheral blood mononuclear cells (PBMC), defined as lymphocytes and monocytes, from septic patients' blood due to the ease of isolation by density gradient centrifugation. However, even the best designed studies lose their biological validity if based on wrong methodological assumptions. In this article we would therefore like to draw attention to a phenomenon that has been known for a long time but has received little attention: In certain pathological conditions apart from sepsis (8) including burn patients (9,10), chronical infections (11,12), and autoimmune diseases (13), but also in healthy pregnancies (14), increased amounts of granulocytes with a lower density compared to normal neutrophils can be found. The detailed elucidation of their origin, function, and role in these conditions is part of current research efforts (15). As their density resembles those of lymphocytes and monocytes, classical density gradient centrifugation with the aim of isolating PBMC results in a relevant low-density granulocyte contamination, which is expected to strongly bias down-stream analytical methods. Here, we provide a simple, yet highly effective solution to improve preanalytical isolation of PBMC from patients with sepsis.

Study Cohort
The study protocol of the clinical study was assessed and positively evaluated by the ethics committee of the Medical Faculty of the Heidelberg University, Germany (S-003/2018). The study was registered in the German Clinical Trials Register (ID: DRKS00018867) and conducted in the interdisciplinary surgical intensive care unit of the Heidelberg University Hospital. This publication reports on a subcohort of the full study. All patients with sepsis and healthy volunteers were of legal age and gave written informed consent. If the patients were incapable, it was obtained from their legal designees. For enrolment of patients, at least two SIRS criteria (16) had to be fulfilled, clinical or microbiological proof of infection as well as sepsis-associated organ dysfunction (1) (defined as a change in the Sequential Organ Failure Assessment (SOFA) score by at least two points compared to the previous day) were needed ( Table 1). Exclusion criteria were: sepsis diagnosis >24h, pregnancy, enrolment in an interventional study, immunosuppression, or viral infections. Blood was drawn at enrolment [sepsis onset (d1)] and, if possible, at day 8 (d8).
For quantification of human leukocyte antigen-DR (HLA-DR) expression on CD14 + monocytes, 2x10 5 cells were stained with 20µL anti-HLA-DR/anti-Monocyte PerCP-Cy5.5 reagent (clone: L243/MwP9) (BD Biosciences, Franklin Lakes, USA) for 30min in darkness and measured directly afterwards. BD Quantibrite PE tubes (BD Biosciences, Franklin Lakes, USA) were used for quantifying the average number of HLA-DR molecules per monocyte as indicated by the manufacturer.
For measurement of mitochondrial reactive oxygen species (ROS) production 1x10 5 cells were washed with 2mL prewarmed Hank's Balanced Salt Solution (HBSS) and incubated with 5µM MitoSOX ™ Red mitochondrial superoxide indicator diluted in HBSS (both from Thermo Fisher Scientific, Waltham, USA) for 10min at 37°C in darkness. Subsequently cells were washed with 2mL pre-warmed HBSS and resuspended in HBSS for measurement.
A FACSVerse ™ flow cytometer (BD Biosciences, Franklin Lakes, USA) was used for all measurement. Results were analyzed using BD FACSuite ™ software (BD Biosciences, Franklin Lakes, USA) and are shown using density plots.

Cell Counting
Cells were counted after density gradient centrifugation and after depletion of low-density granulocytes by impedance-based particle detection using a Scepter ™ 2.0 Cell Counter (Merck Millipore, Burlington, USA).
The total number of cells removed by the depletion process was calculated by subtracting the cell count after depletion (pure PBMC fraction) from the cell count before depletion (PBMC and low-density granulocytes).

Statistical Analysis
Statistical analyses were performed in GraphPad Prism (V 8.0.2 for Windows, GraphPad Software, La Jolla, USA). Results are visualized as scatter plots. Horizontal lines depict the median. Statistical comparison between two groups was done using the Mann-Whitney test. Statistical significance was considered with P ≤ 0.05.

RESULTS
Density gradient centrifugation is a widespread standard method to isolate PBMC in high purity from whole blood. We used this method to isolate PBMC from freshly drawn blood of patients with sepsis. Using flow cytometry to determine the composition of the isolated PBMC, we found a substantial amount of lowdensity granulocyte contamination. They were identified as granulocytes due to their size, granularity, and the absence of lineage markers for T cells (CD3), B cells (CD19), and monocytes (CD14) ( Figure 1A). In three selected samples, 26.3%, 30.3%, and 62.6% of all single cells were found to be low-density granulocytes, respectively ( Figure 1B). Thus, the PBMC's purity is far from sufficient for subsequent experiments. Applying a bead-based depletion of CD15 + cells after the density gradient-based cell isolation resulted in the necessary purity of PBMC isolated from patients with sepsis. The remaining share of low-density granulocytes is even lower than in PBMC isolated from healthy controls by density gradient centrifugation only ( Figure 1C).
The low-density granulocyte contamination is also reflected in the robust increased cell yield compared to healthy individuals after density gradient centrifugation (Figure 2A). The delta between this cell count and the value after depletion ( Figure 2B) was used to computationally estimate the number of low-density granulocytes in the original sample emphasizing the high variability between different patients (d1: range 5.6x10 6 cells; d8: range: 6.6x10 6 cells) ( Figure 2C). However, the actual low-density granulocyte amount is to be assumed lower, since PBMC are also lost during the depletion process. The magnetically labelled CD15 + cells were separated from all other, unmarked cells via retention in a column placed in a strong magnetic field. Apart from the low-density granulocytes sorted out as desired, we found also small proportions of B cells, T cells, and monocytes in this fraction (Figure 3).
To investigate a possible functional impact of the depletion process on the PBMC fraction, cells from three different healthy volunteers were isolated by density gradient centrifugation but only from half of the PBMC fraction CD15 + cells were depleted. Since healthy people also possess an, albeit very low, proportion of low-density granulocytes in their blood ( Figure 4A, range 0.5% -1.2% of single cells; Figure 1C, range 0.4 -1.6% of single cells), the depletion of these cells led to slight shifts in the ratio of individual fractions to the total amount of PBMC ( Figure 4B). Low-density granulocytes were identified as CD3 -, CD14 -, CD19 -, SSC hi . (B) Granulocyte percentage in patients with sepsis after density gradient centrifugation. Density plots from three samples selected for analysis are shown. Numbers indicate the proportion of low-density granulocytes in all single cells. (C) Flow cytometric quality control of low-density granulocyte depletion. Representative density plots are shown for patients with sepsis after depletion of CD15 + cells from PBMC isolated by density gradient centrifugation respectively for healthy volunteers after density gradient centrifugation. Numbers indicate the proportion of low-density granulocytes in all single cells. For total cohort, relative percentages of low-density granulocytes after depletion of CD15 + cells are shown (in case of healthy volunteers after density gradient centrifugation). Each data point represents an individual patient (sepsis d1: n = 14; d8: n = 12/healthy: n = 11) and horizontal line marks the median. Group comparisons (sepsis vs. healthy) were performed using Mann-Whitney test (****P < 0.0001, **P ≤ 0.01). The expression of mitochondrial ROS stayed constant in comparison before and after depletion ( Figure 4C). Moreover, the number of HLA-DR molecules on the surface of CD14 + monocytes remained stable ( Figure 4D). When identical cell numbers were stimulated with LPS, flagellin, zymosan, or human T-Activator CD3/CD28, the depleted fractions secreted higher cytokine amounts than the non-depleted fractions ( Figure 4E).

DISCUSSION
We here report the efficacy of magnetic bead based CD15 + cell depletion to gain highly pure PBMC from freshly drawn blood samples of septic patients. While we used an automated cell separation device for our study, manual single-use separators might be a feasible alternative option to implement this method without extensive costs. Preobrazhensky and Bahler reported a comparable immunomagnetic approach removing granulocyte contamination from cryopreserved blood from healthy individuals (17). Additionally, CD15 is a well-established marker for neutrophils, the largest granulocyte population by distance in the blood of patients with sepsis, and described as being expressed on lowdensity granulocytes in a wide variety of diseases (11,12,14,18,19). Therefore, it seems persuasive that depletion of CD15 + cells is a successful strategy regardless of the source and function of contaminating low-density granulocytes or the underlying pathological condition.
The observed loss of PBMC might be technical. As CD15 is expressed not only on neutrophils and eosinophils but to varying degrees also on monocytes, a fraction of these cells is depleted with our approach, too. On the other hand, a disease-related interaction between column-bound granulocytes and cells contained in the PBMC fraction might prevent the latter from passing the sorting column. Moreover, the selected separation program ("DepleteS") is optimized by the manufacturer to achieve a highly pure fraction of unlabeled cells. To this end, it is accepted that a small proportion of unlabeled cells end up in   the labelled fraction and thus being lost. The PBMC loss is, however, negligible compared to the achieved high purity of PBMC, especially since the proportion of low-density granulocytes varies greatly between individual patients and is therefore a highly variable bias that is difficult to control. Nevertheless, the well-known phenomenon of sepsis-associated lymphopenia (20) is the root cause for the significantly lower PBMC yield per milliliter blood from patients with sepsis compared to healthy volunteers. The co-purifying low-density granulocytes are phenotypically different from mature neutrophils displaying characteristics of immature neutrophils (8) and possess myeloid derived suppressor functions (19). Interestingly, Darcy et al. reported that the number of these cells depends on the severity of the disease. This is an explanation for our observation, that the number of low-density granulocytes co-purifying with the PBMC fraction varies highly between patients. Although some of the studies depicting the problem of contaminating, co-purifying low-density granulocytes have been published a decade ago, it is still common practice today to use density gradient centrifugation for PBMC isolation from patients with sepsis (21)(22)(23)(24)(25)(26)(27)(28)(29)(30)(31)(32)(33). This goes along with a lack of reported information on the isolation purity and identification of the contained cell types. Partly contradictory results between different studies might be explained by clinical differences in the study populations (e.g., severity), not only leading to divergent compositions of PBMC, but primarily to a different proportion of low-density granulocyte contamination. Furthermore, this high share in samples of patients strongly limits the biological comparability with PBMC isolated from control groups (e.g., healthy individuals) without such high abundance of copurifying low-density granulocytes and, thus, a widely differing cell distribution.
By comparing cells from healthy volunteers, which have a very low proportion of low-density granulocytes in their blood, before and after depletion of CD15 + cells, we can show that the external magnetic field and the additional shear stress caused by the necessary pipetting during the depletion process have a negligible technical impact on the PBMC and do not compromise functional cellular properties. The stable expression levels of mitochondrial ROS between pre-and postdepletion indicate that the cells are not further activated during the depletion process. Antigen presentation, a key function of monocytes and essential for proper activation of the adaptive immune system, is not impaired, as shown by stable HLA-DR expression on CD14 + monocytes. As expected, the shares of individual PBMC fractions are slightly altered because the lowdensity granulocytes, although present in small numbers, are removed. This granulocyte subpopulation has myeloid derived suppressor functions, as discussed above. The depletion-related loss of this function together with the higher proportion of cytokine-producing cells after depletion, while the absolute number of stimulated cells remained the same, account for the higher cytokine amounts after depletion. These comparatively higher amounts thus arise less from altered function of the cytokine-producing cells themselves, but rather are a side effect of the deliberate loss of CD15 + cells. This uniform effect weighs small compared to the heterogenous influence of the strongly varying proportions of co-purifying granulocytes, as this high variability between individual patients impair the reproducibility and comparability between different studies and towards controls groups. However, the approach we propose here may not be suitable for all experimental settings. A limitation of our study is the lack of direct evidence that depletion of CD15 + cells does not affect the function of individual PBMC fractions even in patients with sepsis. This needs to be investigated in further studies and adapted to the particular study designs. Moreover, our experimental design does not allow us to assess whether the slight loss of PBMC during depletion is stochastically random or whether cells with specific functions that affect the interaction with low-density granulocytes are primarily lost. Notwithstanding, our results emphasize the necessity of a high degree of protocol standardization across all samples and importantly also study groups within a clinical study to keep sample-handling-induced bias uniform and as low as possible.
Combining density gradient centrifugation and CD15 + cell depletion is a simple and effective strategy to gain highly pure, functionally intact PBMC from patients with sepsis. Eliminating this inter-patient highly variable bias contributes to a higher quality and reliability of results as well as a better comparability between different studies.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by the ethics committee of the Medical Faculty of the Heidelberg University. The patients/participants provided their written informed consent to participate in this study.

AUTHOR CONTRIBUTIONS
All authors contributed to the study design. MO, SU, MW, and FU obtained ethics approval. MO, SU, and MW performed patient recruitment and informed consent. JS and FU established laboratory methodology, performed data collection and analysis, and wrote the manuscript. All authors contributed to the article and approved the submitted version.

FUNDING
The SEPSDIA project (reference number 11501) has been conducted within the framework of the European funding program "Eurostars". The German consortium partner has received financial support from the German Federal Ministry of Education and Research (BMBF), Berlin.