A list of all reagents and resources with the source and identifier is provided in Supplementary Table 3.
C57BL/6J, C57BL/6J (CD45.1), PF4-cre (C57BL/6-Tg(Pf4-icre)Q3Rsko/J)56, Rosa26-iDTRflox (C57BL/6Gt(ROSA)26Sortm1(HBEGF)Awai/J)57, Ifnar−/− (B6.129S2-Ifnar1tm1Agt/Mmjax)58, Ifnar1flox (B6(Cg)-Ifnar1tm1.1Ees/J)59, RS26-creERT2 (B6.129-Gt(ROSA)26Sortm1(cre/ERT2)Tyj/J)60, Myd88−/−(B6.129P2(SJL)-Myd88tm1.1Defr/J)61, CD11b-DTR (B6.FVB-Tg(ITGAM-HBEGF/EGFP)34Lan/J)62, LysM-cre (B6.129P2-Lyz2tm1(cre)Ifo/J)63, Mcl-1fl/fl (B6;129-Mcl1tm3Sjk/J)64 and BDCA2-DTR (C57BL/6-Tg(CLEC4C-HBEGF)956Cln/J)24 mice were purchased from The Jackson Laboratory. Vwf-cre mice were generated by W. Aird and were described previously65. Vwf-eGFP mice were generated by C. Nerlov and described previously10. Tcf4fl/fl (C57BL/6N-Tcf4tm1c(EUCOMM)Wtsi/WtsiH)66 mice were obtained from Wellcome Sanger Institute and INFRAFRONTTIER/EMMA partner (Vienna) from which the mouse was received. PF4-cre mice were crossed with Rosa26-iDTR mice to induce MK cell death in vivo (PF4-cre; RS26-iDTR)52. PF4-cre;RS26-iDTR mice were crossed with Vwf-eGFP mice to visualize the megakaryocytic lineage after induction of MK cell death. Vwf-cre mice were crossed with IFNαR1fl/fl mice to conditionally delete Ifnar in the megakaryocytic lineage. BDCA2-DTR and Ifnar−/− were cross bred to achieve pDC depletion in Ifnar−/− animals (BDCA2-DTR;Ifnar−/−). RS26-creERT2 mice were cross bred with Tcf4fl/fl (C57BL/6N-Tcf4tm1c(EUCOMM)Wtsi/WtsiH) mice to constitutively reduce pDC numbers. FVB-K18-hACE2 expressing humanized ACE2 were bred in the Iannacone laboratory67.
Both male and female mice were used in this study. Unless otherwise stated, mice of the control and experimental group were sex- matched and age-matched (6–12 weeks). Animals were bred and maintained in the animal facilities of the Walter-Brendel Zentrum (Wbex), the Zentrum für Neuropathologie und Prionforschung (ZNP) or the Biomedical center of the LMU Munich, Germany or IRCCS San Raffaele Scientific Institute, Italy or Institute of Science and Technology Austria, Austria. All mice live in standardized conditions in which temperature, humidity and hours of light and darkness are maintained at a constant level all year round. The housing of laboratory mice was in accordance with European and German animal welfare legislations (5.1-231 5682/LMU/BMC/CAM/), Wbex and ZNP. Room temperature and relative humidity ranged from 20 to 22 °C to 45 to 55%. The light cycle was adjusted to a 12 h–12 h light–dark period. Room air was exchanged 11 times per hour and filtered with HEPA-systems. All of the mice were housed in individually ventilated cages (Typ II long, Tecniplast) under specified-pathogen-free conditions. Hygiene monitoring was performed every 3 months based on the recommendations of the FELASA-14 working group. All of the animals had free access to water and food (irradiated, 10 mm pellet; 1314P, Altromin). The cages were equipped with nesting material (5 × 5 cm, Nestlet, Datesand), a red corner house (Tecniplast) and a rodent play tunnel (7.5 × 3.0 cm, Datesand). Soiled bedding (LASbedding, 3–6 mm, PG3, LASvendi) was removed every 7 days. All of the animal experiments were performed in compliance with all relevant ethical regulations for studies involving mice and were approved by the local legislation on protection of animals (Regierung von Oberbayern, Munich; ROB 55.2-1-54-2532-190-2015; ROB 55.2-2532; Vet 02-17-194).
If not stated otherwise, anaesthesia was performed by isoflurane induction, followed by intraperitoneal injection of medetomidine (0.5 mg per kg body weight), midazolam (5 mg per kg body weight) and fentanyl (0.05 mg per kg body weight). Toe pinching reflexes and breathing pattern were used to determine the adequate depth of anaesthesia. Core body temperature was maintained by heating pads, and narcosis was maintained by repetitive injections of 50% of the induction dose, if necessary.
LMU Munich: BM samples of five patients with clinically proven ITP, of five patients with non-Hodgkin lymphoma without BM involvement and 12 patients who died of COVID-19 were analysed. The samples of patients with ITP and lymphoma were archived material, and the COVID-19 specimens were taken during autopsy. Clinical details are provided in Supplementary Table 2. The study was approved by and conducted according to requirements of the ethics committees at the Ludwig Maximilians University of Munich (20-1039). There was no commercial support for this study. University Clinic Aachen: we included 6 consecutive clinical autopsies of patients who were positive for COVID-19 between 9 March 2020 and 5 May 2020 performed at the Institute of Pathology of the University Clinic Aachen. Each patient had a positive clinical SARS-CoV-2 PCR test from upper or lower respiratory tract before autopsy, confirmed by post-mortem PCR with reverse transcription (RT–PCR). Consent to autopsy was obtained by the legal representatives of the deceased patients. The study was approved by the local ethics committee (EK 304/20, EK 119/20 and EK 092/20). BM samples were obtained using an electric autopsy saw (Medezine 5000, Medezine) from the vertebral bodies. The autopsies were performed in two steps according to a modified standard protocol to further increase employee safety and sample acquisition (developed in the frame of the German Registry of COVID-19 autopsies, www.DeRegCOVID.ukaachen.de). The samples were decalcified in formic acid or EDTA before dehydration and embedding in paraffin. Formalin-fixed, paraffin-embedded BM blocks were cut on a microtome at 1–3 µm thickness and decalcified again in EDTA if necessary.
DT was purchased from Sigma-Aldrich (322326) and was intraperitoneally injected into CD11b-DTR mice as a single dose of 25 ng per g for 2 days and BDCA2-DTR and BDCA2-DTR;Ifnar−/− mice with a dose of 8 ng per g per day for consecutive 3 days. A single dose was injected into Pf4-cre;iDTRfl/fl mice 24 h before the experiment. Platelet-depleting antibodies (R300, anti-GPIbα) and isotype control (C301) were purchased from Emfret and used according to the manufacturer’s protocol. pDC-depleting antibodies (ultra-LEAF purified anti-PDCA-1, 927, BioLegend) were injected intraperitoneally for up to 3 consecutive days at a concentration of 150 µg per mouse at day 1 and 100 µg per mouse on the following days. The isotype control (ultra-LEAF purified rat IgG2bk isotype control, RTK4530, BioLegend) was injected accordingly. Type I IFNα was applied by injecting universal IFNα (PBL, assay science) with 5000 U per mouse intraperitoneally in 200 µl PBS. For macrophage ablation, wild-type mice were feed with PLX 5622 chow (D19101002i, AIN-76A), or control chow (D10001i, AIN-76A) from Research Diets, for 7 consecutive days.
Cre-recombinase in RS26creERT2/WT;Tcf4fl/fl mice was induced by intraperitoneal injection of tamoxifen (Sigma-Aldrich, 10540-29-1) dissolved in corn oil (Sigma-Aldrich C8267) three times every other day (1 mg per day), and the mice were analysed 10 days after the first administration26.
Lin−Sca-1+KIT+ (LSK) cells were isolated and sorted from BM of Ifnar−/−, RS26creERT2/WT;Tcf4fl/fl and control mice (Lin-Pacific Blue (Ter-119, CD3, CD8a, CD45R, CD11b, Ly-6G), Sca-1–PE–Cy7, KIT–APC, all purchased from BioLegend, 1:100). A total of 8 × 103 LSK cells was intravenously injected into lethally irradiated C57BL/6J female mice (CD45.1) (two doses of 6.5 Gy with a time interval of 8 h). The BM of chimeras was analysed 8 weeks after the transplantation.
B6.Cg-Tg(K18-ACE2)2Prlmn/ J mice (on the C57BL/6 background) were purchased from The Jackson Laboratory and bred against FVB mice to obtain C57BL/6 × FVB F1 hybrids. Mice were housed under specific-pathogen-free conditions and heterozygous mice were used at 6–10 weeks of age. All of the experimental animal procedures were approved by the Institutional Animal Committee of the San Raffaele Scientific Institute and all infectious work was performed in designed BSL-3 workspaces. Mice were infected intranasal with 105 TCID50 of SARS-CoV-2/human/ITA/Milan-UNIMI-1/2020 (GenBank: MT748758.1) in 25 μl. Then, 5 days after infection, the mice were perfused fixed with 4% PFA and the femurs were embedded in Tissue Tek (also see below). The frozen femurs were cut until the marrow was exposed. The femurs were rinsed with PBS and post-fixed with 4% PFA for 15 min at room temperature. The femurs were washed with PBS and incubated with 10% goat serum for 1–2 h at room temperature. BM was stained with anti-mouse CD41 (for MK/MKP), anti-mouse BST2 (for pDC) and DAPI for nucleus staining. In selected experiments, K18-hACE2 mice were injected intraperitoneally with 2 mg per mouse of anti-IFNAR1 blocking antibody (BioXcell, BE0241, MAR1-5A3) 1 day before infection. All the COVID-19 mouse infection experiments were approved by the Authorization no 270/2022-PR (6EEAF.228).
BM biopsies of five patients with confirmed immune thrombocytopenia and platelet counts <30 × 109 per l were compared with age-matched controls (normal BM biopsies performed for lymphoma staging). Tissue was fixed for 12 h in 4% formalin and embedded in paraffin. For immunohistochemistry, 1.5 µm sections were used. Multiplex immunofluorescence or confocal laser-scanning microscopy imaging were performed after antigen retrieval with epitope retrieval buffer (PerkinElmer). Slides were incubated sequentially for 1 h using the following antibodies: pDCs (anti-human CD123, ab257307, Abcam, 1:100); and MKs (anti-human CD41, ab134131, Abcam, 1:100, or MCA467G Bio-Rad, 1:100) and detection was performed using by TSA-Opal620 (PerkinElmer) and TSA-Opal650 (PerkinElmer). Multispectral imaging was performed using the PerkinElmer Vectra Polaris platform. Images were analysed using HALO (Indica labs) software. Furthermore, the samples were imaged on the LSM 880 confocal microscope using the Airyscan module (Carl Zeiss), Plan-Apo ×20/0.8 or ×63/1.46 objectives and analysed using Zen Blue (v.2.3; Carl Zeiss). The study was approved by and conducted according to requirements of the ethics committees at the Ludwig Maximilians University of Munich (20-1039) and the local ethics committee (EK 304/20, EK 119/20 and EK 092/20).
BM autopsies from patients with COVID-19 (embedded in paraffin) were deparaffinized with xilol twice for 5 min, ethanol (100%) twice for 2 min, ethanol (96%) once for 3 min, ethanol (70%) once for 2 min, and submitted for antigen retrieval with Tris-EDTA pH 9 for 20 min, washed once in 0.5% BSA-PBS-Tween-20 (0.1%) for 5 min. The samples were blocked in 10% donkey serum with 0.5% saponin for 1 h at room temperature. To monitor pDC activation and IFNα production, the following primary antibodies were used: mouse CD69 anti-human (MA5-15612, Thermo Fisher Scientific, 1:200), IFNα rabbit polyclonal (PA5-115430 Thermo Fisher Scientific, 1:50). Anti-CD123 goat polyclonal (ab257307, Abcam, 1:100) was used to label pDCs. Primary antibodies were incubated at 4 °C overnight and samples were subsequently washed three times with 0.5% BSA-PBS-Tween-20 (0.1%) for 5 min before adding secondary antibodies. Secondary antibodies (1:200) were as follows: donkey anti-rabbit-AF488 (A-21206), donkey anti-mouse-AF555 (A-31570) and donkey anti-goat-AF647 (A-21447), all from Thermo Fisher Scientific. DAPI (1:1,000) was used for nucleus staining (15 min at room temperature). The samples were washed three times with PBS 5 min before mounting with DAKO (S3023, DAKO) mounting medium.
Mice were euthanized and bones (sternum, femur and tibiae) were collected and post-fixed in 4% PFA for 1 h at room temperature, and incubated in 15% sucrose for 2 h at 4 °C and in 30% sucrose at 4 °C overnight. Next, the bones were embedded in Tissue-Tek O.C.T. Compound and frozen and stored at −80 °C. Frozen bones were cut on the Histo Serve NX70 cryostat until the exposure of the BM. The sternum was cut as sagittal section. The femurs and tibiae were cut as coronal section or cross-section, according to purpose. Bones were carefully removed from O.C.T. and gently washed in 1× PBS. For whole-mount staining, the cut bones were fixed again in 4% PFA for 15 min at room temperature, washed in PBS and incubated in 10% normal goat serum (Thermo Fisher Scientific) for at least 45 min at room temperature (blocking/permeabilization). The bones were then incubated with primary antibodies at room temperature overnight and washed with PBS before adding secondary antibodies for 2 h at room temperature. Labelling of MKs/MKPs was as follows: primary antibodies: CD41–FITC+ (BioLegend, 133903, 1:100) and CD42-purified hamster anti-mouse (BioLegend, 148501, (1:100)); secondary antibodies: goat anti-hamster Alexa Fluor 647 (Abcam, ab173004, 1:100). Labelling of vessels was as follows: primary antibodies: anti-VE-cadherin (CD144) biotin purified (eBioacience, 13-1441-82, 1:100); secondary antibodies: streptavidin-PE (eBioscience, 12-4317-87, 1:200). Labelling of pDCs was as follows: primary antibodies: anti-SIGLECH-PE or FITC (BioLegend, 129606 or 129603, 1:100) or BST2 (CD317/PDCA-1, Thermo Fisher Scientific, PA5-120152, or eBioscience, 16-3172-81, 1:100) anti-mouse purified, anti-mouse rabbit polyclonal; secondary antibodies: goat anti-rat Alexa Fluor 647 (Abcam, ab150159) or goat anti-rabbit Alexa Fluor 594 (Thermo Fisher Scientific, A-11012) all at a dilution of 1:200. To label the nucleus, Hoechst 33342 or DAPI (Thermo Fisher Scientific, 1:1,000) was used. Lineage-biotin antibodies (Ter-119, CD3e, CD45R, CD11b, Ly-6G) and streptavidin-PE were used at a dilution of 1:200; all antibodies were purchased from eBioscience (San Diego). After staining, bone samples were imaged using the multiphoton LaVision Biotech TrimScope II system connected to an upright Olympus microscope, equipped with a Ti;Sa Chameleon Ultra II laser (Coherent) tunable in the range of 680 to 1,080 nm and a ×16 water-immersion objective (numerical aperture 0.8, Nikon). Single images were acquired at a depth of 50–80 μm, with a z interval of 2 μm. The signal was detected by photomultipliers (G6780-20, Hamamatsu Photonics, Hamamatsu). ImSpector Pro 275 (LaVision) was used as acquisition software. Alternatively, a LSM 880 laser-scanning confocal microscope equipped with an Aryscan module (Carl Zeiss), and the Zen Black acquisition software v.2.3 was used. The images were acquired using the Plan-Apo ×20/0.8 or ×63/1.46 objectives, z-step size of 2 µm, range in z-stack of 40 µm.
IFNα staining was as follows: primary antibodies: IFNα polyclonal antibody (PA5115430, Thermo Fisher Scientific, 1:100); secondary antibodies: goat-anti-rabbit 594 (Thermo Fisher Scientific, 1:200). Macrophage staining was performed as follows: primary antibodies: anti-CD68 monoclonal (Bio-Rad, MCA1957GA, 1:50); secondary goat-anti-rat Alexa 647 (Abcam, 1:200). IFNAR staining was performed as follows: primary antibody: IFNAR1 anti-mouse (BioLegend, 127302, 1:100); secondary antibodies: goat-anti mouse Alexa 555 (Thermo Fisher Scientific, 1:200). Bones were imaged using the LSM 880 confocal microscopy using the Airyscan module, objective Plan-Apo ×20 objective NA, 0.8 or with ×63/1.46 oil Plan-Apo. Images were taken with a z step size of 2 µm, range in z stack of 40 µm and analysed using Zen Blue v.2.3. 3D projections and rendering were performed using Imaris v.9.2 (Oxford Instruments/Imaris).
Anaesthetized mice were placed onto a metal stage with a warming pad to maintain the body temperature. The hair over the skull was carefully removed using an electric hair clipper. The skin on the skull was then cut in the midline to expose the frontal bone. For short-term imaging (<4 h), a custom-built metal ring was glued directly onto the centre of the skull, and the mouse’s head was immobilized by fixing the ring on a stereotactic metal stage. After imaging, the mice were euthanized by cervical dislocation. For long-term (chronic) imaging, a chronic window was implanted on the skull. In brief, a round cover glass (diameter: 6 mm) was centred on top of the frontal bone with sterile saline in between glass and the bone surface. The surrounding area of the glass was then filled with dental glue (Cyano veneer) and a custom plastic ring with inner diameter 8 mm was carefully centred on the frontal bone, with the glass exactly in the middle of the ring. The ring was further immobilized by applying the glue in the gap between the outer edge of the glass and the inner edge of the ring, as well as the gap between the outer edge of the ring and the tissue. Surgery was performed under sterile conditions. The mouse calvarium was imaged using a multiphoton LaVision Biotech TrimScope II system connected to an upright Olympus microscope, equipped with a Ti;Sa Chameleon Ultra II laser (Coherent) tunable in the range of 680 to 1,080 nm and additionally an optical parametric oscillator (OPO) compact to support the range of 1,000 to 1,600 nm and a ×16 water-immersion objective (NA 0.8, Nikon). Time-lapse videos of 3D stacks were recorded within 30 μm to 40 μm depth, with a z interval of 2 or 3 μm and a frame rate of 1 min. Chronic imaging was performed at frame rates of <6 h. Blood vessels and bone structure were taken as landmarks to retrieve the same imaging area of the BM. 3D z stacks were acquired with a z interval of 2 μm; 870 nm or 900 nm was used as an excitation wavelength. The signal was detected by Photomultipliers (G6780-20, Hamamatsu Photonics, Hamamatsu). ImSpector Pro 275 (LaVision) was used as acquisition software. Imaging was performed at 37 °C using a customized incubator. Blood vessels were visualized by intravenous injection of dextran tetramethylrhodamine 500,000 Da (TRITC-dextran, 100 μg in 100 μl solution, D7136, Thermo Fisher Scientific) or Dextran Cascade Blue 10,000 Da molecular mass (D1976, Thermo Fisher Scientific) before imaging. Vwf-eGFP mice were used to visualize the megakaryocytic lineage; pDCs were labelled with SIGLECH-PE antibody (BioLegend, 129606) injected intravenously 20 min before imaging (20 µl diluted with 100 µl NaCl).
Videos and images were analysed using Imaris v.9.2 (Oxford Instruments/Imaris) or ZEN Blue software v.2.3 (Carl Zeiss) or FIJI68. Image denoising using Noise2Void69 was performed in representative micrographs shown in Fig. 1c and Extended Data Fig. 2c. Mosaic images were stitched in Imaris. The numbers of MKs, MKPs and pDCs were quantified in the whole mosaic images and normalized to the total volume of the BM in the image. The cell distance to vessels and/or endosteal surface was measured manually in Imaris Slice mode or by using ZEN Blue (v.2.3). The mean diameter of an MKP or MK was calculated by the average of the longest and shortest axis of the cell. Cell volumes of 3D-rendered BM stacks were measured automatically in Imaris. Cell migration was analysed in 3D time-lapse videos by tracking the cell at every timepoint using Imaris. The cell speed was calculated by dividing the track length with the track duration. The distance of migrating pDCs to MK surfaces was measured and compared to computed random spots using Imaris v.10.9 (Oxford Instruments/Imaris).
Mice were anaesthetized and euthanized by cervical dislocation. Long bones (femurs, tibiae, humerus) were collected into ice-cold sterile PBS. Bones were flushed with PBS + 2% FCS using a 26-Gauge needle and the BM suspension was further filtered through a 70 μm or 100 μm cell strainer (Miltenyi Biotec) and pelleted at 4 °C and 300g for 5 min. The supernatant was discarded and cells were resuspended and incubated in red blood cell lysis buffer for 5 min. Lysis was terminated by adding 30 ml PBS + 2 mM Ethylenediaminetetraacetic acid (EDTA, Sigma-Aldrich), followed by centrifugation at 4 °C and 300g for 5 min. Cells were resuspended with PBS + 0.5% BSA (Carl Roth).
BM isolated cells (as described above) were enriched by removing CD19+ and CD11b+ cells by negative selection using the EasySep selection kit II (StemCell Technologies) for the cell sorting experiments. Cells were incubated with mouse CD16/CD32 (BD Pharmingen (Fc block) before staining (1:100). The following antibodies were used to identify MKs: 1:100 anti-mouse CD41-FITC+ and anti-mouse CD42d-APC+ (BioLegend, 1:100); and MKPs: anti-mouse CD41-FITC+, Pacific Blue Lin− (Ter-119−CD3e−CD45R−CD11b−Ly-6G−), anti-mouse CD105-PE/PercCy7−, CD150-Brillant violet 510+ and anti-CD9-PercCy5.5+ (BioLegend) (all 1:100). We identified pDCs using the following antibodies: anti-mouse SIGLECH-FITC+, CD11b-PE-Cy7− and B220-APC+ from BioLegend (1:100). pDC activation: anti-mouse CD69-FITC (1:200), CD86-PE (1:400), CD11b-APC-Cy7 (1:200), CD317-APC (1:100), SiglecH-PercCy5.5 (1:100) antibodies all from BioLegend and Life/Dead fixable Aqua dead marker (405 nm excitation; Thermo Fisher Scientific, 10 μg ml−1); macrophages: anti-CD45.2+ (BioLegend, PE/Cyanine 7, 1:200), anti-CD45.1 (BioLegend, FITC 1:100), anti-F4/80+ (BioLegend, PerCP/Cyanine 5.5, 1:100), anti-CD64+ (BioLegend, APC, 1:100); anti-CD115− (BioLegend, Brilliant Violet 421, 1:100); neutrophils: CD11b+ (BioLegend, APC/Cyanine 7, 1:200), Ly6G/G1+ (BioLegend, PE/Cyanine 7, 1:200), CD115− (BioLegend, Brilliant Violet 421, 1:100); p-IRF7 expression by pDCs: after staining for pDC surface markers (see above), cells were fixed with PFA and methanol and stained with anti-mouse rabbit monoclonal phospho-IRF7 antibody (Ser437/438, Cell Signaling, 1:100) in Perm buffer III (BD) as previously described42 followed by secondary goat anti-rabbit-APC antibodies (Thermo Fisher Scientific, 1:200). Before loading the samples, 10 μg ml−1 Sytox Orange for the live/dead cell gating and counting beads (1,2,3count beads, Thermo Fisher Scientific), were added to the cell suspension, with exception of the p-IRF7 stain. Apoptosis was measured using Apotracker Green (BioLegend) according to the manufacturer’s instructions. For reticulated platelet staining, 2 µl of blood was fixed with PFA 1%. The blood samples were stained with anti-CD42d-APC (1:100) and thiazole orange (TO) (1 μg ml−1) (Sigma-Aldrich) for 25 min at room temperature in the dark and submitted to flow cytometry analyses70. MK ploidy was quantified after propidium iodide staining in MKs. Measurements were performed on the FACS Canto II cell analyzer equipped with FACSDiva software v.6.0 (BD Biosciences) or on the Cytoflex-S system with CytExpert acquisition software v.2.3 (Beckman Coulter). FACS data were analysed using FlowJo v.10.6.2 or v.10.9. The gating strategies for all FACS data are shown in Supplementary Data 1.
For RNA-seq analysis, BM cells were isolated by flushing the long bones with FACS buffer (2 mM EDTA, 1% FCS, PBS) and treated with Pharm Lyse buffer (BD). Cells were enriched by magnetic removal of CD11b+ and CD19+ cells (EasySep, Stem Cell Technologies). The negative fraction was stained for B220-BV421, SIGLECH-PE, CD9-PerCP-Cy5.5, CD41-FITC, CD42-APC, KIT-APC-Cy7 (all from BioLegend, (1:100). A total of 2,000 cells was sorted using the BD FACS ARIA III Cell sorter (FACSDiva acquisition software v.7.0), into NEB-lysis buffer and processed for sequencing using the NEBNext Single Cell/Low Input RNA Library Kit according to the manufacturer’s protocol (at IMGM). Libraries were pooled in equimolar amounts and sequenced on the NovaSeq 6000 (Illumina) system in a single-end 75-nucleotide run, yielding between 15 and 25 million reads per sample. Reads were mapped against GRCm38.p4 using CLC Genomics Workbench (Qiagen) with the following parameters: mismatch cost 2; insertion/deletion cost, 3; length fraction, 0.8; similarity fraction 0.8; global alignment “no”; strand specific “both”; maximum number of hits per read 5. CLC Genomics Workbench was also used to generate gene expression matrices.
To prepare the data for gene set enrichment analysis (GSEA), DESeq2 (v.1.30.0) analysis was performed using Galaxy with the default parameters71,72. Genes were filtered for an expression of transcripts per million (TPM) > 1 in any condition (42,868 genes) to remove non-expressed or very-low-abundance genes, and then sorted according to the log2-transformed fold change of the respective analysis. For further analysis, the tool GSEA (v.4.0.3) of UC San Diego and Broad Institute was used73,74, referring to their RNA-seq manual pages for analysis. The normalized counts of each replicate as the ranked list generated above were submitted to the GSEA tool with the following parameters: gene sets of their Molecular Signatures Database (MSigDB) in the categories ‘canonical pathways’ (C2) and ‘gene ontology’ (C5) were chosen to contain Ifna1 gene (94 gene sets). Mouse gene symbols were mapped to the human gene symbol (Chip platform: Mouse_Gene_Symbol_Remapping_Human_Orthologs_MSigDB.v7.4.chip), permutation type was set to gene set and gene set size was set to contain between 15 and 2,000 genes.
Genes were filtered for log2-transformed fold change greater or lower than 1 and submitted to the Database for Annotation, Visualization and Integrated Discovery (DAVID) v.6.8 (ref. 75). Resulting GO terms were filtered for q < 0.05. The data visualization tool ClustVis (http://biit.cs.ut.ee/clustvis) was used to generate the heat map of genes expressed in MKPs in Extended Data Fig. 7 (ref. 76). Bulk RNA-seq data are accessible at the Gene Expression Omnibus (GEO; GSE185488).
BM cells were isolated as described above, by flushing the long bones with PBS + 2% FCS, without EDTA using a 26 gauge needle. The BM suspension was further filtered through a 70 μm and 40 μm cell strainer (Miltenyi Biotec) and pelleted at 4 °C and 300g for 5 min. The pellet was resuspended with 1 ml 1× red blood cell lysis buffer and incubated at room temperature for 5 min. After incubation, 15 ml of PBS + 2% FCS without EDTA was added. The cell suspension was centrifuged at 300g for 5 min, the supernatant was discarded and the pellet was resuspended in 1 ml PBS + 2% FCS, and a negative selection kit for CD11b+ and CD19+ (StemCell Technologies) was used according ot the manufacture’s instructions to remove the CD11b+ and CD19+ cells. The final pellet was incubated with the respective TotalSeqB anti-mouse Hashtag antibody (that is, BioLegend, TotalSeq-B0301 anti-mouse Hashtag 3; of this family, Hashtags 3, 4, 5 and 10 were used). After incubation for 30 min on ice and three subsequent washing steps, cells were resuspended in FACS buffer with 2% FBS, followed by centrifugation at 300g for 5 min at 4 °C. The supernatant was discarded and the cell pellet was stained for MKPs as described above. MKPs were sorted using BD FACSMelody Cell Sorter (BD FACS Chorus acquisition software v.1.1.20.0), for 10x scRNA-seq analysis.
The Chromium Next GEM Single Cell 3′ reagent kit v3.1 (CG000206 Rev D) from 10x Genomics protocol was used for sequencing of FACS-sorted BM MKPs. To decrease batch-effect related artefacts, sample multiplexing using TotalSeqB anti-mouse Hashtag antibodies, which were included into the FACS antibody mix, was performed. Four samples were multiplexed into one library. In total, 1 × 105 cells across runs were loaded for generating gel beads in emulsion (GEMs). According to the kit protocol, first, GEMs were generated, then reverse transcription was performed, and cDNA was cleaned up, amplified and size selected. After a quality control and quantification step, gene expression libraries and cell surface libraries were subsequently constructed. The libraries were sequenced using the Illumina NovaSeq system by IMGM laboratories, as described previously77.
Sequencing reads were processed using the Cell Ranger software with the mm10 mouse reference genome index provided by 10x Genomics (https://cf.10xgenomics.com/supp/cell-exp/refdata-gex-mm10-2020-A.tar.gz). This resulted in a count matrix for 16,045 cells and 32,285 genes. The count data were analysed using Seurat78. Background contamination was removed using the soupX method, setting the contamination fraction parameter of 0.1 (ref. 79). Quality control included removal of cells with less than 250 or more than 6,000 features (expressed genes), removal of cells with total UMI counts below 400 and above 20,000, removal of cells with more than 5% of UMIs mapping to mitochondrial genes and removal of genes expressed in less than 3 cells. Furthermore, ribosomal genes were removed. Count data were size normalized to a total UMI count of 10,000 per cell and subsequently log transformed (plus one pseudocount). The top 2,000 highly variable genes were selected on the basis of VST (variance stabilizing transformation)-transformed expression values. Cell cycle scoring was performed and expression values were adjusted for the percentage of mitochondrial UMIs, the S and G2M cell cycle scores. Cells were assigned to samples by demultiplexing the Hashtag oligos, resulting in 1,918 cells for control, 3,243 cells for platelet depletion plus pDC depletion and 1,900 cells for platelet depletion. For differential gene expression analysis, expression levels per gene were centred and scaled across cells. Nearest neighbour graphs (k = 30) were built based on the first 30 principal components. On the basis of the graph, ten clusters were identified using the Leiden algorithm with a resolution of 0.25. Cluster-specific marker genes were identified using Wilcoxon tests, testing only for overexpression, requiring at least 25% of the cluster to express the marker and a log-transformed fold change of at least 0.25. Clusters were assigned to cell types based on the gene annotations of these marker genes. For each cluster, differential gene expression analysis between conditions was performed using the Wilcoxon rank-sum test (wilcox). The DE genes were then selected based on an average log2-transformed fold change cut-off of greater than 0.25 and an adjusted P-value cut-off of less than 0.05. Cell-type-specific gene expression of the gene sets defined from bulk RNA-seq analysis were summarized into gene scores (average expression across the gene set) and visualized by cell type cluster. Trajectory analysis was performed on the following cell types: metabolic MKPs, late MKP, MK-MEPs, cycling MK-MEPs and early MKPs using Monocle380,81. This assigned each cell to an estimated pseudotime along a trajectory. The graph_test function was used to determine genes with pseudotime-associated gene expression patterns (FDR < 0.05 and Moran’s I > 0.25). Gene expression values of genes with pseudotime-associated gene expression were fitted using a spline function with 3 degrees of freedom and corresponding z scores were visualized as a heat map. scRNA-seq data are accessible at the GEO (GSE261996). Code is available at GitHub (https://github.com/heiniglab/gaertner_megakaryocytes).
BM cells (see above) were cultured in DMEM medium containing 10% fetal bovine serum, 1% penicillin–streptomycin and 70 ng μl−1 TPO (ImmunoTools) for 5 days at 37 °C and 5% CO2. On day five, a BSA step gradient was prepared by placing PBS containing 1.5% BSA on top of PBS with 3% BSA (PAA). Cells were loaded on top of the gradient, and MKs were settled to the bottom within 30 min at 1× gravity at room temperature. Mature MKs formed a pellet at the bottom of the tube.
For pDC generation, BM cells were isolated (see above) from control and Myd88−/− mice and cultured for 7 days in RPMI-1640 GlutaMAX-I (GIBCO) supplemented with 10% FCS (GIBCO), 1 mM sodium pyruvate (GIBCO), 1% penicillin–streptomycin (Thermo Fisher Scientific), 1% MEM non-essential amino acids (GIBCO), 0.05 mM β-mercaptoethanol MeEtOH (GIBCO) and recombinant 100 ng ml−1 FLT3L (BioLegend). Cells were collected by flushing Petri dishes with cold PBS. The purity of pDCs was 70–75% as determined by FACS. MK-iDTR mice were injected with DT to induce death of MKs. Control mice received PBS. After 6 h of DT injection, mice were euthanized and femurs were flushed with DMEM medium containing 10% fetal bovine serum, 1% penicillin–streptomycin and 70 ng ml−1 thrombopoietin (TPO, ImmunoTools). MKs were isolated using a BSA gradient as described above. pDCs and MK (1:1) were incubated together for 8 h at 37 °C and 5% CO2. After incubation, the supernatant was collected and analysed for IFNα level (ELISA, see below).
For serum TPO measurement, 1 ml anti-coagulated blood was collected intracardially and kept overnight at −20 °C. The next day, the blood was centrifuged at 2,000g for 20 min and the supernatant (serum) was collected for TPO measurement using the Quantikine Mouse Thrombopoietin ELISA Kit (R&D Systems) to measure the serum TPO levels. IFNα was measured by ELISA (Mouse IFN Alpha All Subtype ELISA Kit, High Sensitivity, PBL Assay Science). Blood was left at room temperature for 20 min and, after centrifugation, the serum was frozen at −20 °C until further analysis. To measure the IFN levels in the BM, one femur was flushed with 200 µl of PBS and cells were centrifuged at 300g. The supernatants were stored at −20 °C until analysis.
PF4-cre;iDTRfl/fl mice were treated with DT for 6 h. The long bones were collected and the BM was isolated by flushing the femurs, tibias and humerus with 200 µl of PBS + 2% FCS using a 26 gauge needle. The BM suspension was further filtered through a 100 μm cell strainer (Miltenyi Biotec) and pelleted at 4 °C and 300g for 5 min. The supernatant was discarded and cells were resuspended and incubated in red blood cell lysis buffer for 5 min. The MKs were isolated as described above and cell suspension was centrifuged for 5 min at 5,000g and 4 °C, followed by 1 min at 11,000g to obtain a tight pellet. The supernatant was collected and transferred to new tubes and centrifuged for 15 min at 2,500g at room temperature (Eppendorf 5415D with the F45-24-11 rotor; Eppendorf). To obtain the MP pellet, the supernatant was transferred into new tubes (homo-polymer, Axygen) and centrifuged for 40 min at 20,000g at room temperature (Mikro200R with the 2424-B rotor; Hettich)82. The resultant MK pellet was collected and the supernatant containing exosomes/extracellular vesicles was transferred into a new tube and treated or not with DNase (1 µl ml−1; Sigma-Aldrich) at 37 °C for 20 min. The treated or non-treated supernatant was added to the pDC cell culture and incubated of 60 min at 37 °C. The pDC supernatant was collected and the IFNα levels were measured using the ELISA kit according to the manufacturer’s instructions (Mouse IFN Alpha All Subtype ELISA Kit, High Sensitivity, PBL Assay Science). The pDCs were collected and stained for FACS analysis for pDC-activation markers (anti-CD69 (1:200) and anti-CD86 (1:400)).
A NanoDrop spectrophotometer (Thermo Fisher Scientific, NANODROP 2000, Peqlab), was used to measure the concentration of DNA in a 2 µl drop of the MK apoptotic supernatant treated or non-treated with DNase I.
CFU assays were performed using the MegaCult kit (StemCell Technologies) according to the manufacturer’s protocol. In brief, femurs and tibias of Vwf-cre;Ifnar−/−, Vwf-cre;Ifnar−/−, Ifnar−/− and Ifnar+/+ mice were flushed with Iscove’s MDM with 2% FBS to isolate BM cells. Cells were washed in Iscove’s MDM (without FBS) before culture. Then, 2.2 × 106 cells were resuspended in cold MegaCult-C medium containing collagen, TPO 50 ng ml−1 and IFNα type 1 universal (5 U, 10 U, 100 U, 500 U or 1,000 U; PBL-Biomedical Laboratories). The final cell suspension (1.5 ml) was loaded into six-well plates and cultivated for 7 days at 37 °C under 5% CO2. After incubation, well plates were imaged using a stereo microscope (Axio Zoom v16 with Objective Plan-NEOFLUAR Z ×1.0/0.25 FWD 56 mm) and Zen Blue software (v.2.6) was used for imaging acquisition (Carl Zeiss). MK-CFUs colonies were classified according to the manufacturer’s protocol (a minimum of 3 cells in close contact).
The Click-it EdU Cell Proliferation Assay Kit (Thermo Fisher Scientific) was used to analyse the MKP proliferation. In vivo labelling of BM cells with 5-ethynyl-2′-deoxyuridine (EdU) was described previously83. In brief, Vwf-eGFP mice were intraperitoneally injected with 0.5 mg EdU in DMSO. After 4 h, mice were anaesthetized and euthanized by cervical dislocation and long bones (femurs and tibiae) were collected. BM cells were prepared as described above. The detection of EdU was performed according to the manufacturer’s protocol. In brief, cells were stained with surface marker antibodies (CD41, CD42) for 30 min at room temperature in the dark, followed by fixation for 15 min (4% PFA, provided in the kit) and permeabilization for 15 min (saponin-based permeabilization and wash reagent, provided in the kit). The samples were washed with 1% BSA between each step. The samples were then incubated for 30 min at room temperature in the dark in EdU reaction cocktail containing PBS, copper protectant, Pacific Blue picolyl azide and reaction buffer additive according to the manufacturer’s protocol. The samples were next washed and analysed by flow cytometry (LSRFortessa cell analyzer equipped with BD FACSDiva v.8.0.1, from BD Biosciences). VWF+CD41+CD42− cells were gated and EdU+ cells were measured within this population using FlowJo (v.10.6.2).
MKs and MKPs from unfractionated mouse BM cell suspensions were directly sorted into RLT buffer (Qiagen) containing 143 mM β-mercaptoethanol (Sigma Aldrich) and total RNA was isolated using the RNeasy Micro Kit (Qiagen) including an on-spin column DNase I digest to remove remaining traces of genomic DNA. First-strand cDNA was synthesized from total RNA with the High Capacity cDNA Reverse Transcription kit (Applied Biosystems) using random primers in 20 µl reaction volumes. RT–PCR was performed using the SsoAdvanced Universal SYBR Green Supermix (Bio-Rad) and the primers for murine Ifnar1 and Actb in the MyiQ Single-Colour Real-Time PCR System (Bio-Rad). Products of RT–PCR were separated by electrophoresis on a 2.5% agarose gel in 1× TBE buffer. Images were taken using a Gel iX Imager (Intas). Primers were as follows: Mm_Ifnar1 Fw, TCTCTGTCATGGTCCTTTATGC (Eurofins); Mm_Ifnar1 Rev, CTCAGCCGTCAGAAGTACAAG (Eurofins); and the Mm_Actb_1_SG primer assay (400 × 25 µl reactions; QT00095242, Qiagen).
GraphPad Prism (v.9.1.2) was used for all statistical analysis. All data were assumed to have Gaussian distribution, unless otherwise specified. Before performed the statistical analysis, the data were confirmed to have equal variance using F-tests, and Student’s unpaired t-tests were used for the comparison of two groups; otherwise, unpaired t-tests with Welch’s correction were used when variances were significantly different. For comparison of multiple groups, one-way or two-way ANOVA was used. Error bars indicate the s.d. All reported probabilities were two-sided. P < 0.05 was considered to be significant.
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
David Posner (at lectern) introduces his plan to bring students from Mexico’s Vasco de Quiroga…
Bluesky has been growing rapidly since 2023.Credit: Matteo Della Torre/NurPhoto/ShutterstockResearchers are flocking to the social-media…
Be wary of claims of alien life in samples from space, say researchers who found…
Adult chimpanzees (Pan troglodytes) play together to promote cooperation and ease tension, according to field…
A spitting cobra can deliver venom either through a bite or by ejecting it from…
Nature, Published online: 20 November 2024; doi:10.1038/d41586-024-03824-1Can jets of drugs from pressurized capsules replace needles?…