Mice
Male Alb-cre-NEMO∆hepa, Alb-cre-NEMOfl/fl referred to as WT, Alb-cre-NEMO∆hepa/Nlrp6−/− and Alb-cre-NEMO∆hepa/Tlr4−/− of the C57Bl6 background were bred and housed in the central animal facility of the University hospital RWTH Aachen. NEMO∆hepa/Nlrp6−/− and NEMO∆hepa lines were generated from an initial heterozygous breeding and then separated for at least 3 generations to allow the development of the Nlrp6−/− dysbiotic microbiota community25. Subsequently, these two lines were kept strictly separate and we did not allow any exchange of mice or bedding material between the two lines as the microbiota related phenotype of these mice has been shown to be transmissible upon co-housing17.
All mice were housed in the individually ventilated cages with access to a standard chow diet and drinking water ad libitum. Upon birth, male mice were assigned to either no treatment, FMT or ABx groups and followed up until week 13. Experiments for these age progression experiments were run and analyzed in parallel. FMT or ABx was initiated in the respective groups at 7–9 weeks of age and continued until week 13. All mice were housed at a temperature of 21−23 °C with relative humidity of 35–65% and 12 h light/dark cycle. All animal experiments were approved by the appropriate German authorities (LANUV. North Rhine-Westphalia. (#AZ84-02.04.2013.A184 (C.T.), (#AZ84-02.04.2013.A260 (C.T.), #AZ84-02.04.2017.A 327 (C.T.), #AZ84-03.04.2013.A240 (C.T.)) All mice were treated in accordance to the criteria of the German administrative panels on laboratory animal care as outlined in the “Guide for the Care and Use of Laboratory Animals” prepared by the National Academy of Sciences and published by the National Institutes of Health (NIH publication 86-23 revised 1985).
Cirrhosis cohort
Human cirrhosis liver tissue specimen were taken from patients that underwent liver transplantation between 1999 and 2005 at the University Hospital Bonn (Supplemental Table 2). The human ethics committee of the University of Bonn (029/13) approved the study. Healthy surgical tissue specimen were obtained from patients who underwent clinically indicated liver resection at University Hospital Bonn or University Hospital rechts der Isar of the Technical University Munich. All patients gave written informed consent to use excess biopsy material for research purposes. The study of these pseudonymized tissue specimen has been approved by the local ethics committee RWTH Aachen University (EK 196/19).
Depletion of microbiota with broad spectrum antibiotics
For microbiota depletion, a broad-spectrum antibiotic cocktail (ampicillin 1 g/l, vancomycin 1 g/l, gentamycin 160 mg/l, metronidazole 1 g/l) was administered in the drinking water of 8-week-old NEMO∆hepa/Nlrp6−/− mice. To decrease the bitter taste of the antibiotics, 25 g glucose were added per liter. Antibiotic treatment was performed until week 13. Antibiotic water was replenished every second day.
Fecal microbiota transfer
For microbiota modulation experiments (fecal microbiota transfer, FMT), NEMO∆hepa mice were treated for 5 weeks three times/week (Monday–Wednesday and Friday) via oral gavage with 200 µl of fecal dilution. To prepare this dilution, per mouse 20 mg of freshly harvested stool (immediately upon defecation) was collected from donor mice. Stool pellets were pooled and then vortexed for 5 min in 20 mg/100 µl anaerobic PBS to homogenize it almost entirely. Next, samples were gently centrifuged for 5 min at 350 × g to allow stool particulate to settle. The supernatant was collected and diluted 1:1 in anaerobic PBS. 200 µl of this suspension was transferred by oral gavage into recipient mice.
This is Akkermansia muciniphila MucT strain was isolated in the lab of Willem de Vos28,66. It was grown as detailed by Depommier et al. Akkermansia muc. was stored in Glycerol at a concentration of 2 × 108 CFU/100 µl at −80 °C. Immediately before gavage Akkermansia was thawed and diluted 1:2 in anaerobic PBS reduced with 0.5 g/l of l-cysteine–HCl. Mice were then gavaged with either 200 µl of this solution or anaerobic PBS.
Bone marrow transplantation
Bone marrow cells from WT and Tlr4−/− donors were transplanted into 6-week-old WT, and NEMOΔhepa recipients after ablative γ-irradiation. Recipients were radiated twice with 6 Gy with an interval of 4 h. Donors were sacrificed and femur and tibia were exposed. With a fine needle the medullary canal was flushed with Hanks/FCS. After twice washing with Hanks/FCS, cells were counted, and recipients received 1 × 106 cells via tail vein injection after the second radiation. During the first four weeks mice received antibiotic water to minimize the danger of infection. Mice were sacrificed 8 weeks after transplantation.
Intestinal permeability in vivo
Isothiocyanate conjugated dextran (FITC-dextran. molecular mass 4.0 kDa. Uppsala. Sweden) dissolved in PBS at a concentration of 200 mg/ml was administered to mice (10 ml/kg body weight) by oral gavage. 4 h after gavage the mice were sacrificed under general anesthesia by isoflurane. Blood samples were collected from inferior vena cava and immediately stored at 4 °C in in the dark. Concentration of FITC in serum was determined by spectrophotofluorometry at an excitation wavelength of 485 nm (20 nm band width) and an emission wavelength of 528 nm (20 nm band width). Relative induction of FITC signal relative to age-matched WT control mice was calculated.
H&E—histology
Hematoxylin and eosin (H&E) staining was performed as previously described18. Briefly, tissue sections fixed in 4% paraformaledehyde (PFA) were cut into 2 µm sections. Tissue sections were deparaffinized and rehydrated. Next samples were stained with Mayer’s Hematoxylin solution for 1 min. Samples were rinsed in tap water for 15 min, placed in distilled water for 30 s, placed in 95% alcohol for 30 s and next counterstained in Eosin solution for 1 min. Finally, samples were dehydrated and mounted with coverslips using the the Roti® Histokit.
Sirius Red staining
Liver fibrosis development was studied using the following protocol. First, tissue sections embedded in paraffin were stained with Sirius red. For this purpose, tissue sections were deparaffinized by heating the slides at 65 °C for 15 min, followed by 2 × 5 min in xylene, and rehydration by introducing a descending concentration of ethanol (100% ethanol and 96% ethanol, 5 min in 70% ethanol and distilled water). Tissue sections were then placed for 45 min in a 0.1% Sirius red solution, followed by 2 × 15 s incubation in 0.5% glacial acetic acid. Finally, sections were dehydrated by ascending alcohol incubations (2 min 96%, 2 × 5 min 100% ethanol and 2 × 5 min xylene). Mounting of Tissue sections was performed with coverslips using the Roti® Histokit.
Immunohistochemistry staining
Five µm thick formalin-fixed, paraffin-embedded liver tissue sections were used to perform immunohistochemical stainings. First, the tissue sections were deparaffinized and rehydrated. For Antigen recovery, sections were heated in a pressure cooker in citrate buffer (pH 6.0). The tissue sections were then immersed in H 2 O 2 solution (0.3% in methanol) for 10 min to block the endogenous peroxidases. To further block unspecific binding, the tissue sections were incubated in 1% bovine serum albumin in PBS for 2 h. Blocking was followed by incubation of the tissue sections overnight with the primary antibodies (Supplementary Table 6) at 4 °C in a humid chamber. After primary antibody incubations tisue sections were washed thoroughly in PBS. Next, the tissue sections were incubated with appropriate horseradish peroxidase-conjugated secondary antibodies (Supplementary Table 6) in a humid chamber at room temperature. Visualized of target signals was achieved by staining with 3,3′-diaminobenzidine solution (Vector Laboratories, Burlingame, CA, USA) for 2–5 min under the microscope. The nuclei were visualized by hematoxylin counterstaining. Finally, the stained sections were dehydrated in increasing concentrations of ethanol and mounted in Entellan.
Immunofluorescence staining
After collection tissue specimens were immediately embedded in Tissue-Tek. Using a cryotome, tissues were cut into 5 µm-thick sections and stored at −80 °C. Slides were air-dried for 30 min at RT followed by 4% PFA fixation. Next, tissue samples were encircled using a hydrophobic pen and blocked with 5% goat serum for 1 h at RT in a humidity chamber.
After blocking, samples were incubated with the primary antibodies (Supplementary Table 6) at 4 °C in a humidity chamber overnight. Samples were washed thoroughly in PBS and then incubated with the secondary antibodies (Supplementary Table 6) for 1 h in a humidity chamber. After incubation, sections were washed again thoroughly in PBS. Finally, sections were mounted in a DAPI (Vector Laboratories, Burlingame, CA, USA) aqueous medium to counterstain nuclei. Staining of mucus and gut bacteria was performed according to an established protocol72. Briefly, colon tissue sections containing feces were fixed using the Carnoy fixation method (60% absolute methanol, 30% chloroform, 10% glacial acetic acid). After paraffin embedding, mucus and gut bacteria were stained with an anti-Muc2 primary antibody and a fluorescence in situ hybridization (FISH) probe against eubacteria (16S rRNA: 5′-GCTGCCTCCCGTAGGAGT-3′).
Flow cytometry analysis of intrahepatic leukocytes
Same amounts of livers were digested by collagenase type IV for 1 h at 37 °C (Worthington Biochemical Corporation, Lakewood, NJ, USA) and intrahepatic immune cells were isolated by multiple differential centrifugation steps as detailed73. Cell isolates were incubated with blocking buffer for 30 min to block the unspecific binding sites of cell surface, then divided into two subgroups and stained with fluorochrome-conjugated antibodies either for myeloid cells FITC Rat anti-Mouse Ly-6G (561105; BD bioscience, Heidelberg, Germany), CD11b Monoclonal Antibody (M1/70), PE (12-0112-82, Thermo Fisher Scientific, Waltham, MA, USA), APC anti-mouse CD11c (117310, Biolegend, San Diego, CA, USA), F4/80 Monoclonal Antibody (BM8), PE-Cyanine7 (25-4801-82, Thermo Fisher Scientific, Waltham, MA, USA), PerCP-Cy™5.5 Rat Anti-Mouse Ly-6G and Ly-6C (552093, BD bioscience, Heidelberg, Germany), APC-Cy™7 Rat Anti-Mouse CD45 (557659, BD bioscience, Heidelberg, Germany) (1:200) or lymphocytes CD3e Monoclonal Antibody (145-2C11), APC (17-0031-83, Thermo Fisher Scientific, Waltham, MA, USA), CD4 Monoclonal Antibody (GK1.5), PE (12-0041-83, Thermo Fisher Scientific, Waltham, MA, USA) CD8a Monoclonal Antibody (53-6.7), FITC (11-0081-85, Thermo Fisher Scientific, Waltham, MA, USA), PerCP-Cy™5.5 Rat Anti-Mouse CD19 (551001, BD bioscience, Flow cytometry measurements were performed on a FACS Fortessa or FACS Canto instrument (BD, bioscience, Heidelberg, Germany). Data were analyzed with the FlowJo software (Ashland, OR, USA).
DNA Isolation and 16S rRNA amplicon sequencing
For 16 S rRNA gene sequencing, DNA was isolated from fecal samples using an established protocol74. Briefly, each sample (around 200 mg) was resuspended in 500 µl of extraction buffer (200 mM Tris, 20 mM EDTA, 200 mM NaCl, pH 8.0). 200 µl of 20% SDS. 500 µl of phenol:chloroform:isoamyl alcohol (24:24:1) and 100 µl of zirconia/silica beads (0.1 mm diameter). Samples were homogenized twice with a bead beater (BioSpec, Bartlesville, OK, USA) for 2 min. After precipitation of DNA, crude DNA extracts were resuspended in TE Buffer with 100 µg/ml RNase I and column purified to remove PCR inhibitors.
Amplification of the V4 region (F515/R806) of the 16S rRNA gene was performed according to previously described protocols75. Briefly, for 16S rRNA amplicon sequencing 25 ng of DNA were used per PCR reaction (30 µl). The PCR conditions consisted of initial denaturation for 30 s at 98 °C, followed by 25 cycles (10 s at 98 °C, 20 s at 55 °C, and 20 s at 72 °C. Each sample was amplified in triplicates and subsequently pooled. After normalization PCR amplicons were sequenced on an Illumina MiSeq platform (PE250).
16S rRNA analysis was conducted based on a previously described computational workflow76. In brief, obtained reads were assembled, quality controlled and clustered using Usearch8.1 (http://www.drive5.com/usearch/). Next, reads were merged using -fastq_mergepairs –with fastq_maxdiffs 30 and quality controlled with fastq_filter (-fastq_maxee 1), minimum read length 200 bp. The OTU and representative sequences were determined using the UPARSE algorithm77, followed by taxonomy assignment using a curated Silva database v12878 and the RDP Classifier79 with a bootstrap confidence cutoff of 80%. The OTU absolute abundance table and mapping file were used for statistical analyses and data visualization in the R statistical programming environment (http://www.rproject.org) package phyloseq80. The permutational multivariate ANOVA (ADONIS test) was performed in R. Factors with p value < 0.05 were considered as significant. Differential abundance analysis (DAA) was performed using a consensus approach based on multiple methods (DESeq2, LefSE, and ALDEx2) to help ensure robust biological interpretation81. DESeq2 was performed using the parameters, test = “Wald”, fitType = “parametric”, alpha = 0.01)82. OTUs were considered significantly DA between genotypes if their adjusted p-value was <0.05 and if the estimated 2-fold change was >2 (Love et al., 2014, McMurdie and Holmes, 2014). LefSe was performed using the R wrapper lefser (Khleborodova A 2021) with the following parameters kruskal.threshold = “0.05”, wilcox.threshold = “0.05”, lda.threshold = “2.5”. ALDEX283 Was performed using default settings, OTUs were considered significantly DA between contrasts if (we.eBH Expected Benjamini–Hochberg corrected p value of Welch’s t test) or (wi.eBH Expected Benjamini–Hochberg corrected p value of Wilcoxon test) was <0.05.
16S rDNA quantitation and taxonomic profiling in liver tissue
Microbial DNA was isolated from frozen liver biopsies with a protocol designed to minimize the risk of contamination between samples, by the environment or experimenters as previously described32. Negative controls consisting of molecular grade water were placed in separate isolation tubes during the isolation process and processed simultaneously throughout the protocol. DNA was amplified using real-time polymerase chain reaction (qPCR) amplification using universal 16S primers targeting the hypervariable V3–V4 region of the bacterial 16s ribosomal RNA gene. qPCR was performed on a ViiA 7® PCR system (Life Technologies, Carlsbad, CA, USA) using Sybr Green technology. Quality control and quantification of the extracted nucleic acids were performed based on gel electrophoresis (1% w/w agarose in TBE 0.5x) and absorption spectroscopy with a NanoDrop 2000 UV spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). High-throughput next-generation sequencing of microbial rDNA was performed using Illumina MiSeq technology as previously described84. Next, (a) The last 20 bases of reads R1 were removed; (b) the last 40 bases of reads R2 were removed; (c) amplicons <350 or >500 nucleotides in length were removed; (d) OTUs with a frequency <0.005% of the total record frequency have been removed; (e) Total Sum Scaling (TSS) normalization was used to normalize OTU read counts to relative frequencies. Because the number of sequences per sample was high and fairly constant between samples (Supplementary Fig. 10a), we chose not to rarefy the data in order to normalize the number of sequences in each sample.
Numerous controls both in vitro and in silico were included to ensure the absence of artifacts related to non-specific amplification of eukaryotic DNA or reagent contamination33. Negative controls and liver samples were compared based on qPCR and beta diversity analyses and showed a clear separation (Supplementary Fig. 10b,c).
In line with our previous data, these numerous quality controls demonstrate that potential bacterial contamination was well contained and had a negligible impact on the taxonomic profiles of the samples in our study33,85,86.
qRT-PCR
Frozen tissue samples from liver or intestine were homogenized in 1 ml Trizol Reagent (Life Technologies, Carlsbad, CA, USA). 200 µl chloroform were added to separate the phases, the upper aqueous phase was transferred into a new collection tube. 500 µl isopropanol were added and the samples remained at RT for 15 min. Afterwards, the RNA was pelleted by centrifugation at 13,000 × g for 10 min at 4 °C, the supernatant was discarded, and the pellets were washed twice with ethanol 70%. Next pellets were air dried and 300 µl DEPC water was used for resuspension. For transcription 1 µg of the isolated mRNA were used and reverse transcription into cDNA was performed using Omniscript® RT Kit (Cat. No. 205113. Qiagen, Venlo, The Netherlands) according to the manufacturer’s protocol. Real-time PCR reactions were performed with Real-Time PCR System Quant studio 6 Flex (Thermo Fisher Scientific, Waltham, MA, USA) and Fast SYBR® GreenER Master Mix (Thermo Fisher Scientific, Waltham, MA, USA) according to manufacturer’s recommendations. The primers were diluted 1:10 fold or 1:50 respectively. All primer sequences are listed (Supplemental Table 7). The Quant Studio Flex software (Thermo Fisher Scientific, Waltham, MA, USA) was used for analysis. In the following the relative mRNA expression was calculated with the 2−ΔΔCT method comparing target gene expression to the GAPDH house-keeping gene.
Library preparation and mRNA sequencing
After quality control with the Agilent Tape Station 4200 RNA ScreenTape Analysis and quantification with the QuantiFluor RNA System (Promega), the library preparation was done according to the manufacturer’s protocol with the Illumina TruSeq Stranded Total RNA Library Prep Gold kit with IDT for Illumina—TruSeq RNA UD Indexes. Sequencing of the library pool was done on one lane using the Illumina NovaSeq 6000 S4 Reagent Kit (200 cycles) with the NovaSeq Xp 4-Lane Kit.
mRNA sequencing analysis
Pre-processing and normalization of RNA-seq data
FASTq files were aligned against the reference genome using the web application BioJupies. The count data were normalized using the Bioconductor package edgeR (version 3.30.0) that filters for lowly expressed genes and corrects for differences in library composition87. Using the Bioconductor package limma (version 3.44.1) we transformed the normalized data to log2-counts per million88.
Transcription factor activity inference with DoRothEA
Transcription factor (TF) activity can be inferred from gene expression data by interrogating the expression of the respective transcriptional targets (i.e., its regulon). It has been shown that this approach is more robust and accurate than observing the expression of the TF itself. We used DoRothEA as the regulon resource as it contains signed TF-target interactions for the majority of all human (and mouse) TFs35. Internally DoRothEA uses the statistical method viper to access the TF activity from gene expression data and returns for each TF a normalized enrichment score (NES) that we consider a proxy for TF activity.
DoRothEA was applied to the normalized gene expression matrix with the following arguments: “method = ‘scale’”, “nes = T,” “minsize = 4” and “eset.filter = F”, using the Bioconductor package dorothea (version 1.0.0; https://saezlab.github.io/dorothea/).
Differences in TF activities between healthy and cirrhotic patients were computed with a t-test. To adjust p-values for multiple hypothesis testing we computed the false discovery rate (FDR).
Pathway activity inference with PROGENy
PROGENy is a tool that allows predicting pathway activities from gene expression data in human (and mouse)34. Instead of interrogating the expression of pathway members, PROGENy takes the expression of the most responsive genes of a pathway into account. These most responsive genes upon pathway perturbation are referred to as footprints (the concept of footprints is reviewed in ref. 89. With PROGENy it is possible to infer the activity of these 14 signaling pathways in human (and mouse): Androgen, EGFR, Estrogen, Hypoxia, JAK-STAT, MAPK, NFkB, PI3K, TGFb, TNFa, Trail, p53, VEGF and WNT.
We applied PROGENy to the normalized gene expression matrix with the following parameters “top = 100”, “perm = 1”, “scale = T”, using the Bioconductor package progeny (version 1.10.0; https://saezlab.github.io/progeny/).
Differences in pathway activities between healthy and cirrhotic patients were computed with a t-test. To adjust p-values for multiple hypothesis testing we computed the false discovery rate (FDR).
Cell types enrichment with xCell
xCell is a tool that performs sample-wise cell type enrichment from gene expression data36. We subsetted the collection of the original 64 immune and stromal cell types to cell types relevant for the liver and the studied phenotype (“iDC”, “ImmuneScore”, “CD8+ T-cells”, “Tregs”, “Epithelial cells”, “NKT”, “MicroenvironmentScore”, “Fibroblasts”, “StromaScore”, “Hepatocytes”, “Th1 cells”, “GMP”, “CD4+ Tcm”, “aDC”).
As suggested by the xCell vignette we transformed the raw counts of the gene expression data to transcripts per million (TPM). Afterward, xCell was applied to the TPM matrix using the R package xCell (version 1.1.0; https://github.com/dviraran/xCell).
Differences in cell type enrichment between healthy and cirrhotic patients were computed with a t-test. To adjust p-values for multiple hypothesis testing we computed the false discovery rate (FDR).
Immunoblotting
The liver and intestine tissue samples were homogenized with NP-40 Buffer containing phosphatase inhibitor cocktail tables (cOmplete mini, PhosSTOP (Roche, Basel, Switzerland) for protein isolation. Protein concentrations were measured using BIO-RAD protein reagent, then adapted to 2 µg/µl, before the proteins were separated electrophoretically on pre-cast 4–12% polyacrylamide gel (Bio-Rad, Hercules, CA, USA) in SDS running buffer at 160 V. After running, the gel was immediately placed in buffer to transfer the proteins to the nitrocellulose blotting membrane with the Trans-Blot Turbo Transfer System (Bio-Rad, Hercules, CA, USA). The success of transfer was checked using Ponceau Red. Before incubating with primary antibodies, the membrane was blocked with 5% non-fat dry milk or 5% BSA diluted in TBS-Tween (TBST 0.5%) to prevent unspecific antibody binding. Subsequently, the membrane was incubated with primary antibodies diluted 1:1000 in 5% dry milk or BSA overnight at 4 °C under agitation. The horseradish peroxidase (HRP)-conjugated secondary antibodies were diluted 1:2000 in 5% dry milk and the membrane was incubated for 1 h at RT. ECL substrate (Pierce, Waltham, MA, USA) developing solution was used before image acquisition with the LAS mini 4000 developing machine (Fuji). Protein expression was quantitatively analyzed with ImageJ in relation to the expression of GAPDH. The following antibodies were used in this study: β-actin (A2066, Sigma-Aldrich, St. Louis, MO, USA), Occludin (71-1500, Thermo Fisher Scientific, Waltham, MA, USA 71-1500), p-JNK/p-SAPK (#9251S, Cell signaling, Danvers, MA, USA). JNK/SAPK (#9252S, Cell signaling, Danvers, MA, USA). GAPDH (AHP1628, Bio-Rad, Hercules, CA, USA).
In-vitro MDSC assay
MDSC isolation
MDSCS were isolated with Myeloid-Derived Suppressor Cell Isolation Kit (mouse; 130-094-538, Miltenyi, Wuppertal, Germany) from liver. After preparing a single cell suspension, the cell number was determined. Cell suspension was centrifuged at 300 × g for 10 min at 4 °C. Supernatant was aspirated completely. Cell pellet was resuspended in 350 μl of buffer per 108 total cells and 50 µl of FcR Blocking Reagent per 108 total cells were added, mixed, and incubated for 10 min in the refrigerator (2−8 °C). 100 μl of Anti-Ly-6G-Biotin (MDSC-Kit) were added, mixed, and incubated for 10 min in the refrigerator (2−8 °C). Cells were washed by adding 5−10 ml of buffer per 108 cells and centrifuged at 300 × g for 10 min at 4 °C. Supernatant was aspirated completely and up to 108 cells were resuspended in 800 μl of buffer. 200 μl of Anti-Biotin MicroBeads were added, mixed, and incubated for 15 min in the refrigerator (2−8 °C). Cells were washed by adding 10−20 ml of buffer per 108 cells and centrifuged at 300 × g for 10 min at 4 °C. Supernatant was aspirated completely and up to 108 cells were resuspended in 500 μl of buffer. LS Column was placed in the magnetic field of a suitable MACS Separator. Column was rinsed with 3 ml of buffer and cell suspension applied onto the column. Flow-through was collected which contained the unlabeled cells. Column was washed with 3 × 3 ml of buffer. The unlabeled cells which passed through were combined with the effluent from step 3; These cells represented the unlabeled pre-enriched Gr-1dimLy-6G– cell fraction. Column was removed from separator and a collection tube was placed under. 5 ml of buffer was added onto the column and the magnetically labeled cells were flushed out by firmly pushing the plunger into the column. These cells represented the labeled Gr-1highLy-6G+ cell fraction.
The unlabeled pre-enriched Gr-1dimLy-6G− cell fraction was centrifuged at 300 × g for 10 min at 4 °C. Supernatant was aspirated completely and up to 108 cells were resuspended in 400 µl buffer. 100 µl of Anti-Gr-1-Biotin per 108 cells was added, mixed, and incubated for 10 min at 4 °C. Per 108 cells 5–10 ml of buffer were added and centrifuged at 300 × g for 10 min at 4 °C. Supernatant was aspirated completely and up to 108 cells were resuspended in 900 μl of buffer. In addition, 100 µl of Streptavidin MicroBeads were added, mixed, and incubated for 15 min at 4 °C. 10–20 ml buffer per 108 cells were added and centrifuged at 300 × g for 10 min at 4 °C. Supernatant was aspirated completely and up to 108 cells were resuspended in 500 μl of buffer. MS columns were placed in the magnetic field and 500 µl of buffer were added onto the column. Cell suspension was applied onto the column and the collected and represented the unlabeled cells. The column was washed 3 × 500 µl. All flow through were collected. Column was removed from separator and a collection tube was placed under. 1 ml of buffer was added onto the column and the magnetically labeled cells were flushed out by firmly pushing the plunger into the column. These cells represented the labeled Gr-1dimLy-6G− cell fraction.
T cell isolation
T cells were isolated with (mouse; 130-095-130, Miltenyi, Wuppertal, Germany) from spleen. After preparing a single cell suspension, cell number was determined. Up to 107 cells were resuspended in 40 µl buffer and 10 µl of biotin–antibody cocktail per 107 total cells were added, mixed, and incubated for 5 min at 4 °C. 30 µl of buffer and 20 µl of Anti-Biotin MicroBeads per 107 total cells were added, mixed, and incubated for 10 min at 4 °C. LS columns were placed in the magnetic field and 3 ml of buffer added onto the column. Cell suspension was applied onto the column and flow through collected. Column was washed 3 × 3 ml and flow through collected.
T cell CFSE labeling
T cells were centrifuged with 300 × g for 10 min at 4 °C and resuspended in 1 ml PBS/0.1% BSA. A solution of CFDA-SE (Vybrant CFDA SE Cell Tracer Kit, V12883, Thermo Fisher Scientific, Waltham, MA, USA) from DMSO Stock at 2X final labeling solution was prepared (100 µM). T cells were resuspended in 1 ml solution containing CFDA-SE dilution and incubated in the dark for 15 min at 37 °C. Cells were quenched with 4 ml ice cold T cell medium and centrifuged with 300 × g for 10 min at 4 °C. Cells were washed two times.
In vitro T cell assay
U bottom 96 wells were coated with 2 mg/ml CD3 antibody (6-0032-85 (Clone 17A2; 1 mg/ml) Thermo Fisher Scientific, Waltham, MA, USA) and incubated for 2 h at 37 °C. Plates were washed three times with PBS prior to the start of the assay. T cells (105 cells) were incubated with gMDSCs or mMDSCs in the following ratios: 1:0, 1:1, 1:2 or 1:4. Additionally, 10 µg/µl of CD28 (553294 (Clone 27.51; 1 mg/ml) BD bioscience, Heidelberg, Germany) was added per well.
Proliferation was analyzed using a FACS Fortessa (BD, Bioscience, Heidelberg, Germany). Data were analyzed with the FlowJo software (Ashland, OR, USA).
Measurement of routine serum parameters
Routine serum parameters alanine aminotransferase (ALT), aspartate aminotransferase (AST), glutamate dehydrogenase (GLDH) and alkaline phosphatase (AP) were measured in the central laboratory of clinical chemistry in RWTH Aachen University Hospital.
Quantification and statistical analyses
For comparisons of two groups, significance was tested by unpaired two-tailed Student’s t test. In case of more than two groups, we employed one-way ANOVA followed by Tukey-test with adjusted p-value for multiple comparisons. For not normally distributed data, two groups were compared using Wilcoxon–Mann–Whitney-Test and in case of more than two groups Kruskal–Wallis test with Dunn–Bonferroni-Test was used. Data were considered significant between experimental groups as: *p < 0.05. **p < 0.01 or ***p < 0.001.
Statistical analyses of 16S microbiota data was performed using R version 3.4.3 (2017-11-30) (http://www.rproject.org) and the packages ‘phyloseq’. and ‘ggplot2’80,90. The permutational multivariate analysis of variance test (ADONIS) and analysis of similarities (ANOSIM) were computed with 999 permutations. For ADONIS tests, a R2 > 0.1 (effect size 10%) and p-value < 0.05 was considered as significant. RNA Sequencing data were analyzed using R as detailed above. The clinical cirrhosis cohort was analyzed using IBM SPSS Statistics software (Version 25). For graphic representation and statistical analysis R version 3.6, Rstudio and GraphPad Prism 8.0 were used.
Reporting summary
Further information on research design is available in the Nature Research Reporting Summary linked to this article.