Comparative analysis of monocyte distribution width alterations in Escherichia coli sepsis: insights from in vivo and ex vivo models

Sample collection, hematological analysis, and digital morphology

Peripheral venous blood was collected in EDTA-K₂ tubes (BD Vacutainer^®, 6 mL) from healthy adult subjects (n=24 for E. coli studies and n=10 for LPS curve studies) recruited as anonymous volunteers at the Blood Transfusion Centre of the Hospital Santa Maria della Misericordia of Urbino during their periodic blood donation session. Healthy donors fulfill all the requirements needed for blood donation. In detail, the inclusion criteria were donors of all genders, aged 18–65 years, ≥50 kg, BP≤180/100; HR=50/100 bpm; Hb=13.5/12.5 g/dL (M/F), no known hematological, oncological, and infective diseases. Exclusion criteria included: positive tests for HbsAg, anti-HCV antibodies, anti-HIV 1–2 antibodies and HIV 1–2 antigens, anti-Treponema Pallidum (TP) antibodies. All donors signed an informed consent form. Samples were processed within 4 h from blood collection at the Clinical Biochemistry Laboratories of the University of Urbino.

Routine complete blood cell counts and MDW were analysed on a UniCell DxH 690T Hematology Analyzer (Beckman Coulter, Inc., Brea, USA). The instrument performance has been daily validated by using three different commercially available internal quality controls: COULTER 6C Plus Cell Control (three concentration levels) to monitor the DxH 690T system performance for CBC, DIFF and NRBC parameters, COULTER Retic-X Cell Control (three concentration levels) for monitoring system performance of the reticulocyte parameters, and COULTER LATRON CP-X Control, a suspension of latex particles of standardized size used to monitor VCSn technology (volume, conductivity, and light scatter measurements). Each sample has been assayed once, except for the samples treated with the highest E. coli concentrations, which required multiple measurements due to the appearance of instrumental flags.

Manually prepared air-dried smears were stained through the automated slide preparation system (Sysmex SP-50) with May-Grunwald-Giemsa-stain for digital cell morphology analyses with the CellaVision DM software (DI-60).

This observational non-interventional ex vivo study was approved by the local Ethical Committee (on the basis of an official document of accordance with the Blood Transfusion Center of “S. Maria della Misericordia” Hospital in Urbino (PU), Italy) and all investigations have been conducted according to the Declaration of Helsinki principles as revised in 2013.

Bacterial strain, preparation of bacterial suspensions, and growth condition

The bacterial strain used in this study was E. coli ATCC 25922. Bacteria were routinely grown at 37 °C on Muller–Hinton (MH) agar.

For each experiment, one colony obtained from a fresh culture was transferred in 20 mL of MH broth and incubated at 37 °C for 14–18 h (usually overnight) with the cap slightly loosened.

The next day, the overnight culture was harvested, washed, resuspended twice in saline solution (NaCl 0.9 %), and then adjusted by spectrophotometry at optical densities at 600 nm (OD₆₀₀), corresponding to a concentration of 10¹⁰ colony forming unit (CFU)/mL. Then, from this bacterial suspension, 10-fold serial dilutions in saline solution (NaCl 0.9 %), from 10⁻¹ to 10⁻⁴ were prepared in 10 mL final volumes. Aliquots (1 mL) of this dilution were placed into a microcentrifuge tube and centrifuged for 15 min at 6000×g, to form a pellet, and the supernatant was carefully removed with a pipette to ensure the pellet was not dislodged; blood was immediately added to the pellet. Amounts equivalent to a final concentration of 10⁶, 10⁸, and 10¹⁰ CFU/mL were used as stimuli for the human whole blood model.

Ten microliters of each dilution of E. coli suspension were plated in triplicate on tryptone soya agar (TSA) plates and incubated at 37 °C for 18–24 h for the final enumeration of CFU/mL.

All culture materials were purchased from Thermo-Fisher Scientific (Waltham, MA USA).

Whole blood in vitro stimulation with LPS or living E. coli

Aliquots of 1 mL of EDTA-K₂ whole-blood from each volunteer were immediately incubated with 0.1, 0.5, 1.0, 5.0, and 10 μg/mL of lipopolysaccharide (LPS from E. coli O127:B8, Sigma-Aldrich cod. L3129) or living E. coli suspension to a final concentration of 10⁶, 10⁸, and 10¹⁰ CFU/mL at RT (about 25 °C).

CBC results, including MDW, were measured at 0, 30, 60, and 180 min after careful mixing to avoid blood cell sedimentation. After 3 h, all blood samples were centrifuged (2000×g, 15 min) to obtain plasma for further analyses. All tubes, tips and solutions were endotoxin free to avoid artefactual introduction of endotoxin.

Determination of the total protein concentration

Plasma samples were assayed by Bradford Protein Assay (Bio-Rad, USA) to determine the total protein concentration through Onda V-30 Scan Spectrophotometer (Giorgio Bormac, Italy).

MDW measurement in sepsis patients

MDW values and clinical data of sepsis patients (n=23; age range: 44–87 years; age mean: 67 years) diagnosed with E. coli bloodstream infection were retrospectively extracted by data archives of the Laboratory Medicine Department, Laboratori Clínic Metropolitana Nord (LCMN), University Hospital Germans Trias i Pujol, Badalona, Spain. At the same time, MDW values and clinical data from unaffected controls (n=135; age range: 19–80 years; age mean: 54 years) were collected to overcome the possible confounding factor due to the different anticoagulants used (EDTA-K₃) compared with the ex vivo study.Data from septic patients and control subjects were collected from two different studies carried out in the Emergency Department and Intensive Care Unit of Germans Trias i Pujol University Hospital. Both studies were carried out according to the basic ethical principles of the Helsinki Declaration (Fortaleza, October 2013). Both in vivo studies were reviewed and approved by the Clinical Research Ethics Committee of the Hospital (PI-18-249 and PI-18-140). All patients included in these studies have signed an informed consent previously approved by the CEIC (Ethics Committee for Clinical Investigation).

Cytokine determination by multiplex immunomagnetic assay

After 3 h of treatments, all plasma samples obtained were assayed to evaluate a panel of 27 inflammatory biomarkers through the Pro™ Human Cytokine 27-plex assay (including: IL-1β, IL-1ra, IL-2, IL-4, IL-5, IL-6, IL-7, IL-8/CXCL8, IL-9, IL-10, IL-12(p70), IL-13, IL-15, IL-17, Eotaxin/CCL11, bFGF, G-CSF, GM-CSF, IFN-γ, IP-10/CXCL10, MCP-1/CCL2, MIP-1α/CCL3, MIP-1β/CCL4, PDGFbb, RANTES, TNF-α, and VEGF) according to the manufacturer’s instructions (BioPlex, Bio-Rad Labs) as previously reported [22]. The cytokine concentrations (pg/mL) were calculated through a 8-point standard curve performed in duplicate.

Statistical analysis

Regarding CBC parameters, differences among groups and times were determined using a mixed-effects model followed by Tukey’s multiple comparison test. The Mann-Whitney test was selected for comparing MDW differences between two groups and Kruskal–Wallis with Dunn’s Multiple comparison test was chosen for comparing differences among multiple groups. Concerning inflammatory mediators, differences among groups were determined using a mixed-effects model followed by Dunnet’s multiple comparison test. Values are expressed as mean±standard deviation (SD) unless otherwise specified, and p-values <0.05 were considered significant. All statistical tests were performed using GraphPad Prism 9.0.

Digital morphology analyses

The treatment of whole blood aliquots with increasing concentrations of E. coli and LPS was associated with the early appearance of monocyte anisocytosis. Monocytes lose their round shape with a reniform nucleus (Figure 1A) in favour of an increased cell volume with intracellular vacuolization and nuclear alterations (Figure 1B–G) and clear evidence of phagocytosis of live E. coli bacteria (Figure 1B–E).

Figure 1:

Digital light microscopy images of peripheral blood monocytes representative of untreated whole blood controls (A), and after 3 h of treatments with E. coli 10⁶ CFU/mL (B, C), E. coli 10⁸ CFU/mL (D, E) and LPS 1 μg/mL (F, G). (May-Grunwald-Giemsa Staining, ×100).

Complete blood count and MDW determination in LPS-stimulated blood samples

CBCs, including MDW, were anonymously determined in 10 healthy donors (6 M, 4 F; age range: 31–65; mean age: 46.1 years) before and after stimulation with increasing LPS doses. MDW values in controls reached a mean ± SD value of 17.4 ± 1.6 at T0, 17.8 ± 1.3 at T30, 17.7 ± 1.9 at T60, and 18.4 ± 1.3 at T180, without significant differences among all times considered.

The in vitro stimulation of whole blood with increasing concentrations of LPS from E. coli induced a significant time- and dose-dependent MDW increase without affecting the monocyte count (Figure 2A–C). In detail, MDW reached values of 20.4 ± 1.6, 21.4 ± 1.4, 23.7 ± 1.7, 27.5 ± 4.1, and 26.9 ± 4.1 (mean ± SD) after 3 h of treatment with 0.1, 0.5, 1.0, 5.0, and 10.0 μg/mL of bacterial LPS. The two highest concentrations of LPS showed overlapping MDW modifications at each time point explored, highlighting a possible plateau effect (Figure 2A). A representative time- and dose-dependent two-dimensional data plots of all WBC subpopulations separated according to volume and RLSs values is reported in Figure 2D, where the green cloud represents the monocyte population; a focus on the monocyte population is also reported as surface plots (Figure 2E).

Figure 2:

Time- and dose-dependent MDW variations (A) and monocyte absolute (B) and percentage (C) counts in whole blood stimulated with increasing concentrations of LPS from E. coli (0.1, 0.5, 1.0, 5.0, 10 μg/mL). Values are expressed as mean±SD. The table under Figure A reports the significance values resulting from the statistical analysis (the comparisons were calculated among treatments and respective controls at each time). Representative time- and dose-dependent two-dimensional data plots (D) and monocyte surface plots (E) resulting from whole blood stimulated with increasing concentrations of LPS from E. coli (0.1, 0.5, 1.0, 5.0, 10 μg/mL). In (D) all WBC subpopulations were separated according to volume and RLSs values; in (E) volume and RLS of monocytes are plotted (RLS, Rotated Light Scatter); in (D) and (E) the monocyte population is represented in green.

Complete blood count and MDW determination in E. coli-stimulated blood samples

CBCs and MDW were evaluated in 24 healthy donors (18 M, 6 F; age range: 18–65; mean age: 42.5 years) before and after stimulation with increasing concentrations of live E. coli. MDW values in controls reached a mean ± SD value of 17.2 ± 1.4 at T0 and T30, 17.8 ± 1.6 at T60, and 18.8 ± 1.4 at T180, without significant differences among all times considered. The treatment of whole blood with living E. coli determined a time- and dose-dependent MDW increase (Figure 3A). In detail, E. coli 10⁶ CFU/mL induced a significant time-dependent increase of MDW at 30, 60, and 180 min (mean±SD) vs. respective controls (19.2 ± 2.1, p=0.0078; 20.5 ± 2.2, p<0.0001; 21.7 ± 3.7, p=0.0005, respectively) (Figure 3A).

Figure 3:

Time- and dose-dependent MDW variations (A) and monocyte absolute (B) and percentage (C) count in whole blood stimulated with increasing concentrations of living E. coli (10⁶, 10⁸, 10¹⁰ CFU/mL). Values are expressed as mean±SD. The dotted line indicates normal values. The comparisons were calculated among treatments and respective controls at each time. (∗∗p=0.001–0.01; ∗∗∗p=0.0001–0.001; ∗∗∗∗p<0.0001). Representative time- and dose-dependent two-dimensional data plots (D) and monocyte surface plots (E) resulting from whole blood stimulated with increasing concentrations of live E. coli (10⁶, 10⁸, 10¹⁰ CFU/mL). In (D) all WBC subpopulations were separated according to volume and RLSs values; in (E) volume and RLS of monocytes are plotted (RLS, Rotated Light Scatter); in (D) and (E) the monocyte population is represented in green.

MDW increases to 21.2 ± 2.6 (p<0.0001) after 30 min, 23.5 ± 3.4 (p<0.0001) after 60 min, and 26.1 ± 5.0 (p<0.0001) at 180 min were observed by treating whole blood with E. coli 10⁸ CFU/mL.

The treatment with E. coli 10¹⁰ CFU/mL promoted a MDW increase to 23.0 ± 2.1 (p<0.0001) after 30 min, 26.3 ± 2.8 (p<0.0001) after 60 min, and 29.0 ± 4.9 (p<0.0001) at 180 min (Figure 3A).

These Gram-negative related MDW modifications were not accompanied by a change in the monocyte population count, neither in the total count (Figure 3B) nor in percentage values (Figure 3C), whose mean values were distributed within the physiological ranges (0,2-1,2 × 1000/µL and 2–12 %, respectively). A representative time- and dose-dependent two-dimensional data plots of all WBC subpopulations separated according to volume and RLSs values is reported in Figure 3D, where the green cloud represents the monocyte population; a focus on the monocyte population is also reported as surface plots (Figure 3E).

MDW determination in sepsis patients

The retrospective analyses of MDW values in sepsis patients affected by E. coli infection, compared with unaffected controls recruited at the same Hospital, revealed that MDW values in sepsis patients were significantly higher compared to control subjects (MDW in sepsis – min-max: 22.4–49.3; median: 28.2; MDW in controls-min-max: 15.2–28.3; median: 20.0; p<0.0001) (Figure 4A).

Figure 4:

MDW variations in sepsis patients affected by E. coli infection and unaffected controls. In (A) MDW values are expressed as absolute values, as measured by DxH900 in EDTA-K₃ samples; in (B) MDW values are reported as percentage fold-of-change to of the values measured in vivo and, in the ex vivo model (further details in the main text). Values are expressed as box and whiskers, showing min to max values, median and 25th to 75th percentiles. Statistical tests: Mann–Whitney (A) and Kruskal–Wallis with Dunn’s Multiple comparison test (B); *p=0.01–0.05; ****p<0.0001.

Of note, these samples have been collected in EDTA-K₃, thus, to ensure a correct comparison between in vivo and ex vivo results, we normalized MDW from patients and treated samples (at 3h) for the median MDW value of the control cohort, and for the corresponding untreated controls in the ex vivo dataset, respectively. The percentage fold-of-change calculated after normalization confirmed the differences emerged in the two models between sepsis patients and controls, as well as between E. coli treated samples and controls, revealing also the absence of significant differences in MDW levels between sepsis patients and whole blood samples ex vivo triggered with E. coli for 3h (Figure 4B).

Cytokine quantification by multiplex Immunomagnetic assay

The total protein content was not significantly different between plasma from controls and E. coli/LPS-stimulated blood cells (mean mg/mL±SD; CTR: 45.8 ± 5.3; EC10⁶ CFU/mL: 45.6 ± 3.9; EC10⁸ CFU/mL: 46.2 ± 4.9; EC10¹⁰ CFU/mL: 46.6 ± 3.3; LPS: 45.6 ± 9.1).

After 3 h of treatment with E. coli 10⁶ and 10⁸ CFU/mL (due to multiple CBC readings of E. coli 10¹⁰-treated whole blood, the remaining sample volume was not sufficient to quantify the cytokine panel), we observed a dose-dependent increase of peculiar cytokines, as shown in Figure 5.

Figure 5:

Cytokine values in plasma samples (n=12) from whole blood after 3h of stimulation with increasing concentrations of living E. coli (10⁶ and 10⁸ CFU/mL) and LPS from E. coli (1 µg/mL). (Mixed-effects model followed by Dunnet’s multiple comparison test: ∗p=0.01–0.05; ∗∗p=0.001–0.01; comparisons were calculated among treatments and respective controls). Values are expressed as box and whiskers, showing min to max values, median and 25th to 75th percentiles of the pg/mg total proteins.

In detail, we highlighted that the levels of MIP-1α/CCL3 were significantly increased by all triggers (E. coli 10⁶, 10⁸ CFU/mL and LPS) compared with untreated controls, showing the strongest responses (fold of change vs. controls: E. coli 10⁶ CFU/mL: 23.4-fold, p<0.05; E. coli 10⁸ CFU/mL: 21.0-fold, p=0.001–0.01; LPS: 48.2-fold, p<0.05) (Figure 5).

The levels of IL-8/CXCL8 also showed a significant increase of about 9.1-fold (p<0.05), 17.0-fold (p=0.001–0.01), and 10.2-fold (p=0.001–0.01) after treatments with E. coli 10⁶ CFU/mL, E. coli 10⁸ CFU/mL, and LPS, respectively. Levels of TNF-α were significantly up-regulated from 5- to 10-fold vs. untreated controls, after living E. coli and LPS treatments (Figure 5).

G-CSF levels were significantly increased by about 2.9-fold (p<0.05), 3.9-fold (p=0.001–0.01), and 5.5-fold (p=0.001–0.01) after treatments with E. coli 10⁶ and 10⁸ CFU/mL, and LPS, respectively.

Plasma levels of IL-1β showed a significant increase of about 1.1-fold, 1.6-fold (p<0.05), and 1.8-fold after treatments with E. coli 10⁶ CFU/mL, E. coli 10⁸ CFU/mL, and LPS, respectively.

A minor significant up-regulation was observed for MIP-1β/CCL4 of 1.2-fold, 1.2-fold (p<0.05), and 1.7-fold after treatments with E. coli 10⁶ CFU/mL, E. coli 10⁸ CFU/mL, and LPS, respectively.

Finally, a significant decrease of Eotaxin/CCL11 was observed after E. coli 10⁶ and 10⁸ CFU/mL treatments (p=0.01–0.05); a significant down-regulation of PDGF-bb was also observed after E. coli 10⁸ CFU/mL stimulation (p=0.01–0.05) (Figure 5). According to the 8-points standard curves and the Limit of Detection (LOD) of the assay IL-2, IL-15, and VEGF were removed from the analyses due to the limited number of samples with measurable levels of cytokines.

A complete overview of cytokine behavior, according also to age and gender differences is provided in Figure 6. The heat map shows the mean value (pg/mL) of each cytokine measured in untreated controls, E. coli 10⁶ and 10⁸ CFU/mL, and LPS treated samples, obtained both in the whole cohort of donors, and classified according to gender and age (cut-off 45 years).

Figure 6:

The heat map represents the variations of cytokines, chemokines, growth factors and colony stimulating factors according to gender and age differences. On the left are indicated all the inflammatory parameters. The first group of columns (all) highlights the mean cytokine values obtained from the analyses of the entire population (n=12). The other groups of columns (gender and age) show the mean cytokine values obtained from the analyses of the population divided according to the gender (F, n=3; M, n=9) and the age (<45 years, n=7; >45 years, n=5). Cytokine concentrations (pg/mL) are depicted as colours ranging from magenta, blue, green, yellow, orange, red, brown, and grey, indicating increasing levels of concentration.

These findings suggest preliminary indications on potential age- and sex-related differences in cytokine-mediated immune responses associated with bacterial infection. Indeed, certain cytokines appeared to be differently modulated by the same trigger in subjects of different sexes (e.g., IL-8) and age ranges (e.g., MIP-1α, IL-8, and PDGFbb). Further experiments in a larger cohort of subjects are required to confirm this observation.