Human neutrophils can mimic myeloid-derived suppressor cells (PMN-MDSC) and suppress microbead or lectin-induced T cell proliferation through artefactual mechanisms

We report that human conventional CD15+ neutrophils can be isolated in the peripheral blood mononuclear cell (PBMC) layer during Ficoll gradient separation, and that they can impair T cell proliferation in vitro without concomitant neutrophil activation and killing. This effect was observed in a total of 92 patients with organ transplants, lung cancer or anxiety/depression, and in 18 healthy donors. Although such features are typically associated in the literature with the presence of certain myeloid-derived suppressor cell (PMN-MDSC) populations, we found that commercial centrifuge tubes that contained membranes or gels for PBMC isolation led to up to 70% PBMC contamination by CD15+ neutrophils, with subsequent suppressive effects in certain cellular assays. In particular, the suppressive activity of human MDSC should not be evaluated using lectin or microbead stimulation, whereas assays involving soluble or plate-bound antibodies or MLR are unaffected. We conclude that CD15+ neutrophil contamination, and associated effects on suppressor assays, can lead to significant artefacts in studies of human PMN-MDSC.


Figure S2: Phenotype of CD15 + cells, isolated by different methods, immediately and the next day
(A) PBMC were isolated from the blood samples over Ficoll using conventional tubes ("No membrane") or with membrane-containing tubes ("Over membrane" and "From membrane"). Then cells were stained for live/dead, CD15 and activation markers, CXCR2, CD54 and CD62L. Bars illustrate an absence of activation of neutrophils, isolated with membrane-containing tubes, in comparison with conventional Ficoll isolation. (B) PBMC were isolated from healthy donors ("Donors") or from patients listed for lung transplant ("LT"), immediately or the next day. In comparison with donors, LT neutrophils tend to express more CD54. At the second day blood, neutrophils tend to moderately decrease CD62L and CXCR2 expression, but no apparent signs of activation were observed. At least 3 samples in each group were evaluated in 14 independent experiments. Medians and individual values are shown.

Figure S3: No substantial changes in phenotype of CD15 + cells after overnight incubation of the whole blood at 4˚C
PBMC were isolated from the same blood sample within 3 hours of drawing (top) or on next day (bottom) after incubation at 4˚C, and an absence of activation of neutrophils was confirmed by flow cytometry: CD15 + cells continued to express CD62L and CXCL2 and did not upregulate CD54. One experiment out of three is shown. Presented dot-plots consist of 11,321 events (top) and 32,185 events (bottom).

Figure S4: CD15 + cells from healthy donors are suppressive
PBMC were isolated with SepMate tubes, CD15+ cells were isolated with magnetic beads, and incubated with healthy donor (#390) responder cells in suppression assay. One representative experiment out of six is shown, and all tested healthy donor CD15 + cells were suppressive. Four of them had their PBMC isolated the same day with SepMate or Accuspin tubes, and in 2 cases PBMC were isolated one the day after acquisition. Presented histograms consist of 1,201 ± 46 (Mean ± SEM) events. antibodies and bead-to-cell ratios were tested, and suppression was only seen in microbead-stimulated assays. Presented histograms consist of 1,256 ± 40 (Mean ± SEM) events.

Figure S6: Prolonged incubation of PBMC on ice and use of different concentrations of platebound CD3 mAb in suppression assay
CD15 + cells were isolated with magnetic beads, from 1 st day PBMC of lung transplant patient, at room temperature or on ice, and incubated with healthy donor (#390) responder cells in suppression assay.
Responders were stimulated with indicated concentrations of plate-bound CD3 mAbs or by CD3 microbeads (as shown at Figure 4D). One experiment from two is shown, and both had the same results.

Figure S7: Comparison of CD3 microbead stimulation versus plate-bound CD3/28 mAb stimulation using CD4depleted cells as suppressor cells
CD4depleted cells were isolated using second day blood from PBMC of two anxiety/depression patients (samples #1 and #2), and from tumor (sample #3) and PBMC (sample #4) of a lung cancer patient.
Isolated CD4cells were incubated with healthy donor #390 responder cells in suppression assay.
Responders were stimulated with CD3 microbeads or with plate-bound CD3/28 mAbs. One experiment from two is shown. In total, 7 samples, including healthy donors, were evaluated with the same results.

Figure S8: Comparison of CD3 microbead stimulation vs. plate-bound CD3 mAb stimulation using CD4depleted cells as suppressors
CD4depleted cells were isolated from the first day PBMC of 2 healthy donors, and incubated with healthy donor #390 responder cells in suppression assay. Responders were stimulated with CD3 microbeads or with plate-bound CD3 mAbs in indicated concentrations. In total, 7 samples, including healthy donors, were evaluated in 3 experiments with the same results. Presented histograms consist of 1,520± 71 (Mean ± SEM) events.  Bars illustrate an absence of apparent activation of neutrophils and the well-preserved viability, although both lectins tend to decrease CD62L expression on neutrophils.
(B) CD15 + cells were isolated with magnetic beads and incubated in 1:1 ratios with CFSE-labeled healthy donor responder PBMC for 4 days. PBMC were stimulated as indicated. As expected, neutrophils suppressed T cell proliferation induced by CD3 microbeads stimulation.Surprisingly, neutrophils also suppressed PHA and Con A -stimulated PBMC in most cases (top row). However, when responder PBMC cells were pre-activated with the same mitogens overnight, and then washed and mixed 1:1 with neutrophils, we observed an absence of any suppression and even, to some extent, an activation of lectinsstimulated PBMC division by neutrophils (bottom row). Importantly that CD3 microbeads were pelleted down along with PBMC, and therefore neutrophils added the next day into suppression assay were still able to disrupt beads-activated divisions of T cells (bottom row, left columns). Divisions of CD8 + cells of one healthy donor in one experiment are shown. Data of CD4 + responders from the same experiment are presented in Figure 6B, and more in depth analysis of T cells proliferation is shown at Suppl. Figure   S11). In total, 3 experiments with 4 healthy donors neutrophils and 8 healthy donors PBMC responders were performed with similar results. Presented histograms consist of 3,563 ± 194 (Mean ± SEM) events. Figure S11: In depth analysis of suppressive capabilities of neutrophils with different types of T cells stimulation CD15 + cells were isolated with magnetic beads and incubated in 1:1 ratios with CFSE-labeled healthy donor responder PBMC for 4 days. PBMC were stimulated as indicated. CFSE + CD4 + and CFSE + CD8 + responders were gated and analyzed using proliferation modeling in FlowJo v.10.4. Results are presented as Division index (top), Proliferation index (middle) and Percent divided (bottom). As expected, neutrophils suppressed T cell divisions, stimulated with CD3 microbeads. Neutrophils also suppressed PHA and Con A -stimulated PBMC in most cases ("Continuous stimulation", black). However, when