Cytokine responses to LPS in reprogrammed monocytes are associated with the transcription factor PU.1

Abstract Myeloid‐derived suppressor cells (MDSCs) are functionally immunosuppressive cells that arise and expand during extensive inflammatory conditions by increased hematopoietic output or reprogramming of immune cells. In sepsis, an increase of circulating MDSCs is associated with adverse outcomes, but unique traits that can be used to identify increased activity of MDSCs are lacking. By using endotoxin tolerance as a model of sepsis‐induced monocytic MDSCs (M‐MDSC‐like cells), this study aims to identify the mediator and transcriptional regulator profile associated with M‐MDSC activity. After analyzing 180 inflammation‐associated proteins, a profile of differentially expressed cytokines was found in M‐MDSC‐like cells versus normal monocytes stimulated with LPS. These cytokines were associated with 5 candidate transcription factors, where particularly PU.1 showed differential expression on both transcriptional and protein levels in M‐MDSC‐like cells. Furthermore, inhibition of PU.1 led to increased production of CXCL5 and CCL8 in M‐MDSC‐like cells indicating its role in regulating the ability of M‐MDSC‐like cells to recruit other immune cells. Taken together, the study identifies a unique profile in the pattern of immune mediators defining M‐MDSC activity upon LPS stimulation, which offers a functional link to their contribution to immunosuppression.


Summary sentence
Differential cytokine response in endotoxin induced M-MDSC-like cells and their associated regulators. INTRODUCTION Immunosuppressive mechanisms are essential for regulation of the immune responses in order to prevent excessive tissue damage during an inflammatory event. The cell-mediated suppression is largely mediated through a number of specific immune cells, for example, regulatory T cells and myeloid-derived suppressor cells (MDSCs). 1,2 Although these cells are important for immunologic homeostasis, they are also found to be involved in the pathology of diseases, like sepsis and cancer, if not effectively regulated. [3][4][5] MDCSs constitute a heterogeneous population of myeloid cells with immunosuppressive function that expand during infection, inflammation, and cancer, by increased hematopoietic output or by reprogramming of immune cells. 6,7 MDSCs can be divided into 2 main subtypes: granulocytic-MDSCs or polymorphonuclear-MDSCs (G-MDSCs or PMN-MDSCs) and monocytic MDSCs (M-MDSCs). 8 The reprogramming of monocytes into M-MDSC is initiated by exposure to damageassociated molecular patterns and pathogen-associated molecular patterns. 9 In the case of LPS-mediated reprogramming, this phenomenon is also known as endotoxin tolerance. 4,6,7,10 The reprogrammed immunosuppressive monocytes in experimental models, referred to as M-MDSC-like cells, 6 display similar properties as M-MDSCs studied in cancer 6 and sepsis, 11,12 including a drastic reduction of proinflammatory TNF production and an impaired antigen presenting capacity. The latter is indicated by decreased surface expression of MHC class II molecules, such as HLA-DR. 10,13 MDSCs were first discovered as "suppressor cells" in tumors 14 but recently MDSCs have been investigated in other pathologic conditions, including sepsis, 5 trauma, 15 and most recently in COVID- 19. 16 In sepsis, a correlation between low HLA-DR expression and impaired TNF production in septic monocytes upon LPS stimulation has been demonstrated, 17 and this endotoxin tolerance has been suggested to be associated with sepsis severity. 12 Immunosuppression further contributes to mortality, in particular at later stages of sepsis, where the numbers of MDSCs in the circulation increase. 11,18 Although MDSCs have gained attention as critical regulators of the inflammatory response during sepsis, there is still a lack of unique traits that can be used to identify increased activity of MDSCs, and to potentially identify septic patients in an early phase of immunosuppression. The aim of this study was to identify mediators, associated with M-MDSC activity and their potential transcriptional regulators upon LPS stimulation, to provide a profile specific for M-MDSC identification.

Cell culture
Primary human monocytes were isolated from buffy coats collected from anonymous healthy blood donors by the Department of Transfusion Medicine at Örebro University Hospital, Sweden. The study was conducted in compliance with the ethical guidelines of Helsinki as well as the ethical policy at Örebro University Hospital. All donors had signed a consent allowing the blood to be used for research purposes.
The blood samples were anonymized by the Department of Transfusion Medicine at Örebro University Hospital, and no personal information can thereby be tracked back. As the buffy coats were prepared in connection to a regular blood donation, the donors were not exposed to any additional harm or risk. According to the paragraph 4 of Swedish

Generation of M-MDSC-like cells and LPS challenge
To induce M-MDSC-like cells, the isolated primary monocytes from each donor were divided into 2 fractions and seeded (1. were also purchased from Invivogen. Following a medium change, both F I G U R E 1 A representation of used cell model. Isolated primary CD14 + cells were seeded (1.5 × 10 6 /cm 2 ) into culture flasks, either left untreated (monocytes) or stimulated with LPS (10 ng/ml) for 20 h (M-MDSC-like cells). Following a medium change and a 40 h rest, the cells were reseeded and subjected to an initial challenge with LPS (10 ng/ml, normal cells) or rechallenged with LPS (10 ng/ml, M-MDSC-like cells) fractions of cells were then allowed to rest for 40 h. Following the resting period, the cells were reseeded and were challenged with 10 ng/ml LPS. A schematic presentation of the timeline can be seen in Figure 1.

Olink Proteomics
A total of 180 immune mediators were analyzed in cell culture supernatants from 4 individual donors using a proximity extension assay (PEA, Olink Proteomics, Uppsala, Sweden) coupling oligonucleotidetagged antibodies to a qPCR reaction, from which the amplified sequences were quantified. 19 The arrays used were "immune response" and "inflammation," as available from the manufacturer.

Quantification of immune mediators
Cells of each phenotype were seeded, at 5 ×

Extraction of mRNA, conversion to cDNA and qPCR
Two hours after LPS challenge, cells from each phenotype were collected and lysed in Buffer RLT (Qiagen, Hilden, Germany). Total RNA from lysates derived from 5 ×  TaqMan Gene Expression Assays (Applied Biosystems; see Table 1 for details), according to the manufacturer's instructions to achieve a final reaction volume of 10 μl. Water was included as a negative control in every run to check for cross contamination. Pipetting of samples and reaction mixtures into 384-well plate was performed by PIRO Pipetting Robot (Dornier, Lindau, Germany). The PCR protocol started with an initial denaturation phase at 95 • C for 20 s, followed by 40 amplification cycles at 95 • C for 1 s, 60 • C for 20 s. Peptidylprolyl cis-trans isomerase B (PPIB) was determined as reference gene by using NormFinder R package (MOMA, Aarhus University Hospital, Denmark) for normalization among a total of 3 candidate reference genes. For cDNA quantification, a 6-point serially 4-fold diluted calibration curve was developed from PBMCs stimulated by 1 μg/ml LPS and cultured with RPMI complete cell culture medium. The RNA of treated PBMCs was extracted from 1 × 10 7 cells per reaction using QIAamp RNA Blood Mini Kit (Qiagen). cDNA reverse transcription was conducted by the same procedure as sample cDNA synthesis.
All samples were amplified in duplicate and the mean quantity values were obtained for further data analysis. The threshold for the CV value between duplicates was set to <0.15. The samples with higher CV values were rerun. Cycle threshold (CT) cut-off value was set to 35.
All reactions had an efficiency between 89% and 110%, which corresponds to a slope between −3.63 and −3.10.

Liquid chromatography-mass spectrometry-based proteomics
Twenty-four hours after LPS challenge, 1 × 10 6 cells from each phenotype were pelleted and flash-frozen in liquid nitrogen and stored in The fractions were analyzed on an Orbitrap Fusion Lumos Tribrid mass spectrometer interfaced with an Easy-nLC 1200 liquid chromatography system (both Thermo Fisher Scientific). Precursor ion spectra were recorded at 120,000 target resolution, most abundant precursors were fragmented by collision-induced dissociation at 35% collision energy, the TMT reporter ions were generated by higherenergy collision dissociation at 65% collision energy and the MS3 spectra were recorded at 50,000 resolution. Targeted inclusion list was prepared for the selected peptides of the proteins P25963 (NFkappa-B inhibitor alpha, NFKBIα), Q16665 (hypoxia-inducible factor 1-alpha, HIF-1α), and Q99814 (endothelial PAS domain-containing protein 1, EPAS-1) that were detectable according to the data in ProteomicsDB (https://www.ProteomicsDB.org). 21 The fractions were reanalyzed using the modified LC-MS method that only fragmented the precursor ions with the correct charge and the monoisotopic mass within 15 ppm of the theoretical mass in the inclusion list.
Proteins were identified and quantified using Proteome Discoverer

Statistical analysis
In the initial analysis of 180 immune mediators, each set of phenotype data was compared with control (unstimulated monocytes) using Student's t-test, followed by FDR correction (Benjamini-Hochberg). Initially, a low stringency (p = 0.15) was used to allow identification of a larger number of potential candidates that could be selected for sub- it should be noted that >7000 proteins were identified, but only the abundance of the selected transcription factors was used for the preplanned, down-stream statistical analyses of identified transcription factors. All analyses were done in GraphPad Prism v. 7.04.

Exploring the differences in LPS response
The    TNF, IL-10, and CCL8 also showed a significant difference between the 2 phenotypes ( Figure 3A). The remaining 4 mediators could not be quantified using electrochemiluminescence or cytometric bead assays.

Changes in gene expression and relative protein abundance of potential upstream regulators
The potential upstream regulators governing the identified differences in the responses between the 2 phenotypes were explored using IPA.
The data set obtained from PEA as well as further validation experiments was uploaded into the IPA software, and was used to curate a list of potential upstream transcription factor candidates. These candidates were further filtered based on their connections with both  Last, PU.1, the transcription factor encoded by SPI1, and that had differential expression on both gene and protein level in our study among the 2 phenotypes, is a transcription factor of the E26 transformation-specific family that is involved in both early development of immune cells and mature immune cells function. 43 CCL8, also known as MCP-2, plays an important role in recruiting a large number of different immune cells in inflammatory conditions 55 and has also been shown to attracting tumor-associated macrophages and promote metastasis in cancer. 56 In addition, increased CCL8 levels have also been shown in septic patients' plasma compared with healthy volunteers at the early stages of sepsis. 57 However, the levels of circulating CCL8 levels during later stages, when immunosuppression is more pronounced, remain unknown.
In conclusion, we have identified a profile comprised of mediators