Introduction

Sensing of “non-self” is a key feature of innate immunity and underlies the recognition of viruses, bacteria, and fungi, but also plays important roles in cancer and auto-immune diseases [1, 2]. Pattern recognition receptors (PRRs), like Toll-like receptors (TLRs), Nucleotide-binding oligomerization domain (NOD)-like receptors (NLRs), and retinoic acid-inducible gene 1 protein (RIG-I)-like receptors (RLRs) are essential components of innate immune signaling and selectively recognize pathogen-associated molecular patterns (PAMPs). Dedicated PRRs, like TLR3, RIG-I, Melanoma differentiation-associated protein 5 (MDA5), and Cyclic GMP-AMP synthase (cGAS)-Stimulator of interferon genes protein (STING) recognize viral double-stranded RNA (dsRNA) and dsDNA, and are important sensors for infections with RNA and DNA viruses, as well as infections with retroviruses [1,2,3]. While TLR3 recognizes dsRNA in endosomes, the prototypical RLRs, RIG-I and MDA5, sense cytosolic dsRNAs, and cGAS-STING detects viral dsDNA [1,2,3,4]. STING is activated either directly via viral dsDNA, through the STING agonist 2′3′-cGAMP generated by the cyclic GMP-AMP synthase cGAS upon detection of viral dsDNA, or indirectly via RIG-I and MDA5 [5]. Activated STING interacts with TANK-binding kinase 1 (TBK1) and activates interferon regulatory factor (IRF) 1, -3, and -7 and Nuclear factor-κB (NF-κB), leading to the initiation of anti-viral and inflammatory transcriptional programs, including interferon-stimulated genes (ISGs) and interferons (IFNs) [5,6,7,8].

IFNs are secreted cytokines with important roles in immunity and anti-viral responses. IFN signaling relies on Janus kinase-Signal transducer and activator of transcription (JAK-STAT) activation, phosphorylation of STAT1/2, and the induction of ISG and IFN gene expression that influence viral replication [9, 10]. Although the vast majority of cell types can be triggered to express type I (IFN-α, -β, -ε, -κ and -ω) and type III (IFN-λ1, -λ2, -λ3 and -λ4) IFNs, the expression of IFN-specific receptors is cell type-restricted and determines IFN responses. For example, type I IFN receptor (IFNAR) is ubiquitously expressed in many tissues, whereas expression of type III IFN receptor IFNLR1 is mainly limited to epithelial cells, e.g. the gastro-intestinal and respiratory epithelium [6,7,8, 11]. Although type I and type III IFNs induce similar ISG signatures, type I IFNs generally trigger a more rapid increase and decay of ISG expression [7]. In addition, IFN-λs have been described to be the first-in-line defense against viral infections and might inhibit viral spread without triggering inflammatory responses, depending on IFN-λ receptor expression [7, 12, 13].

The novel severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is the causative agent of the pandemic Coronavirus disease 2019 (COVID-19) and belongs to the human coronaviruses (HCoV) that also include SARS-CoV and MERS-CoV [14]. In many patients with severe COVID-19, SARS-CoV-2 infection induces the secretion of highly pro-inflammatory cytokines through cGAS-STING and NF-κB-mediated signaling [15, 16]. Type I and III IFNs are important regulators of host viral defense against SARS-CoV-2 [6,7,8, 11, 17, 18], but at the same time, SARS-CoV-2 evades immune recognition via IFN and ISG suppression [10, 19]. Prolonged expression of low basal levels of type I and III IFNs might prime host responses against virus infection, including SARS-CoV-2 [20,21,22,23]. Although type III IFNs restrict SARS-CoV-2 infection in intestinal and airway epithelial cells [18, 24,25,26,27,28] and STING agonism reduces SARS-CoV-2 infection [29,30,31,32], context-dependent damaging effects of type III IFNs on airway epithelia during viral infections have been described as well [33, 34].

Innate immunity, PRRs, and IFN signaling are closely regulated by ubiquitination, both by the host machinery as well as through viral E3 ligases and deubiquitinating enzymes (DUBs) that hijack the host ubiquitin machinery [35]. STING, RIG-I, TLR3, and TBK1 are positively and negatively regulated by differential modification of polyubiquitin chains, including K11-, K27-, K48- and K63-linked chains [36, 37], by a variety of E3 ligases, such as TRIM56 [38], TRIM32 [39], MUL1 [40], AMFR [41], RNF5 [42] and TRIM29 [44]. The interplay and functional consequences of ubiquitin modifications are complex and include proteasomal degradation as well as stabilization of protein-protein interactions. Importantly, IFNs and anti-viral signaling are also heavily regulated by DUBs, like USP13 [45], USP35 [46], and Cylindromatosis (CYLD) [47].

Ubiquitin-specific peptidase 22 (USP22) is a DUB that is part of the deubiquitination module of the Spt-Ada-Gcn5-acetyltransferase (SAGA) complex, through which it regulates transcription via the control of histone H2A K119 and H2B K120 monoubiquitination (H2AK119ub1 and H2BK120ub1, respectively) [48,49,50]. Recently, additional USP22 substrates have emerged, with important roles in cell fate regulation and programmed cell death [51,52,53]. Interestingly, USP22 has mostly been associated with IFN signaling and ISG expression upon viral infection [

Results

Profiling USP22-mediated gene expression in HT-29 hIECs

Substrate-specific deubiquitination is a crucial determinant of ubiquitin homeostasis and regulates receptor activation and internalization, proteasomal degradation, and transcription. For the ubiquitin-specific protease USP22, both transcriptional and extranuclear targets have been identified. As part of the DUB module of the SAGA complex, USP22 regulates transcriptional elongation via H2AK119ub1 and H2BK120ub1 [48,49,50]. Up till now, the spectrum of target genes regulated by USP22 remains largely unclear, partially due to organism-, cell- and context-dependent redundancy in alternative DUBs that might compensate for loss of USP22 [56]. We previously reported that CRISPR/Cas9-mediated knockout (KO) of USP22 in the human colon carcinoma cell line HT-29 affects RIPK3 ubiquitination during necroptosis without inducing major changes in RIPK1, RIPK3, and MLKL gene expression [53], suggesting gene-specific regulation of USP22. To identify the spectrum of USP22-regulated genes, we profiled USP22-dependent changes in gene expression in the hIEC line HT-29. Quantification of alterations in gene expression in two independent HT-29 USP22 KO single cell clones revealed a marked alteration in gene expression, with 401 genes upregulated and 182 downregulated (Fig. 1A and Supplemental Fig. 1A). Loss of USP22 expression was accompanied by changes in H2Bub1, but not H2Aub1 (Supplemental Fig. 1B, C). Among the top-50 differentially regulated genes, 30 were up- and 20 downregulated, with an adjusted P-value of <0.05 (Fig. 1B). Genes upregulated in both USP22 KO clones (#16 and #62) compared to control (non-human target: NHT) HT-29 cells include genes that encode for proteins involved in growth and differentiation, like Transforming Growth Factor β-1 (TGFB1), Tumor-associated calcium signal transducer 2 (TACSTD2) and Tyrosine-protein kinase Mer (MERTK) and the cytosolic RNA- and DNA sensor DExD/H-Box Helicase 60 (DDX60). Downregulated genes include USP22, mitochondrial adenylate kinase 4 (AK4) that is involved in the regulation of mitochondrial function and ATP production [57], and regenerating islet-derived protein 4 (REG4), a carbohydrate-binding lectin that has been identified as marker for deep crypt secretory cells (DSCs) that acts as niche for Lgr5-positive stem cells in the colon [58]. Differential regulation of gene expression, as well as loss of USP22 expression, was also demonstrated by independent qRT-PCR of the USP22-dependent upregulated genes TGFB1, SLFN5, TGM2, and DDX60, as well as downregulation of USP22, CXCR4, and AK4 (Fig. 1C), confirming the quality of the microarray.

Fig. 1: Profiling USP22-mediated gene expression in HT-29 hIECs.
figure 1

A. Volcano plot showing the differential gene expression patterns of two independent single-cell HT-29 USP22 CRISPR/Cas9 KO clones (#16 and #62) compared to CRISPR/Cas9 control (NHT) HT-29 cells. Color code represents the log2 fold change compared to NHT. B Heatmap of the top-50 differentially regulated genes between HT-29 USP22 KO single clones #16 and #62 and the NHT control. Color coding represents the row-wise scaled (Z-score) RNA intensities. Genes are sorted according to their log2 fold change, compared to NHT. C Basal mRNA expression levels of the indicated genes were determined in control and two independent USP22 KO HT-29 single clones using qRT-PCR. Gene expression was normalized against 28S mRNA and is presented as x-fold mRNA expression compared to NHT. Mean and SD of three independent experiments in triplicate are shown. *P < 0.05; **P < 0.01, ***P < 0.001.

Full size image

Loss of USP22 specifically enriches for genes involved in IFN signaling and response to viral infection

Next, gene-set enrichment analysis was performed on USP22-regulated genes to investigate if certain gene ontology (GO) terms are specifically regulated by USP22. Interestingly, GO analysis revealed an enrichment of genes linked to type I and II IFN signaling, as well as regulation of viral genome replication and several other viral processes, such as the regulation of viral genome replication, response to virus, response to IFN-γ-, and IFN-γ-mediated signaling pathways in USP22 KO HT-29 cells as compared to control NHT HT-29 cells (Fig. 2A). Interestingly, the GO terms of genes that are strongly downregulated are enriched in mitochondrial translation and gene expression, ribosomal and ribonucleoprotein complex biogenesis and the processing of tRNA, rRNA, and ncRNA (Fig. 2A).

Fig. 2: Loss of USP22 specifically enriches for genes involved in IFN signaling and response to viral infection.
figure 2

A Bar plot showing the top-20 regulated GO terms in two independent single-cell HT-29 USP22 CRISPR/Cas9 KO clones (#16 and #62) compared to control (NHT) HT-29 cells. Color code represents the number of annotated genes within each gene set. B Heatmap of differentially expressed genes contributing to the GO terms response to type I IFN (left) and type II signaling (right). Color code represents the log2 fold change compared to NHT. Note that due to lack of annotation and overlap** ISGs between type I/II and type III IFNs, response to type III IFN as GO term was not included. C Basal mRNA expression levels of GO- enriched genes related to IFN signaling in control (NHT) and two independent USP22 KO HT-29 single clones using qRT-PCR. Gene expression was normalized against 28S mRNA and is presented as x-fold mRNA expression compared to NHT. Mean and SD of three independent experiments in triplicate are shown. *P < 0.05; **P < 0.01, ***P < 0.001, n.s. not significant. D Western blot analysis of basal MX1, IRF9, ISG56, ISG20, and USP22 expression levels in control and USP22 KO HT-29 cells (clone USP22 KO #62). GAPDH served as loading control. Representative blots of at least two different independent experiments are shown.

Full size image

Since previous studies suggest controversial roles of USP22 in IFN signaling [79], differences in GO term enrichment might be functionally irrelevant. IFN-λ mostly exhibits long-term signaling effects and plays important roles in SARS-CoV-2 infection in airway epithelial and gastro-intestinal cells and organoids and has been shown to critically control antiviral defense [18, 24,25,26,27,28].

The susceptibility towards SARS-CoV-2-infections is determined by USP22-mediated regulation of STING. STING is described as a mediator of IFN-λ1 production in HT-29 cells, and during viral infection in primary human macrophages in a Ku70-dependent manner [80, 81]. We furthermore demonstrate for the first time that in the absence of viral infections or exogenous IFN, loss of USP22 expression resulted in basal and 2’3’-cGAMP-induced STING ubiquitination in hIECs. In addition, loss of STING expression decreased IFN/ISG signaling that occurred under conditions of USP22 deficiency, suggesting that STING acts as a physical scaffold for USP22-dependent ubiquitin modifications. STING ubiquitination serves different physiological roles, including determining protein stability, mediating protein-protein interactions, and cellular localization [38,39,40,41,42,44,45,46,47]. Recently, cGAS-STING activity has emerged as regulator of immunopathology in COVID-19, highlighting the relevance of adequate STING regulation [82]. STING ubiquitination enables the STING-TBK1 interaction upon cGAS-mediated recognition of cytosolic DNA and is generally associated with activation of ISG expression [71]. Until now, USP22-mediated STING ubiquitination has only been described upon viral infection and upon ectopic overexpression. For example, overexpressed USP22 modifies ectopically expressed STING with HA-tagged K27 ubiquitin upon HSV-1 infection in HEK293T cells [83]. At least three independent experiments in duplicates are shown. All primers were purchased at Eurofins (Hamburg, Germany). Primer sequences are shown in Supplementary Table 1.

Gene expression profiling

To quantify global changes in gene expression, RNA was isolated as described above, followed by a DNase digest upon RNA binding using the peqGOLD DNase Digest Kit, according to the manufacturer’s instructions. Samples were processed and gene expression was profiled at the DKFZ Genomics and Proteomics Core Facility (Heidelberg, Germany) using the Affymetrix human Clariom S array.

Gene expression profiling analysis

Raw.CEL files were processed with the oligo R package [84] and normalized intensities were obtained after RMA normalization. Genes with differential expression between NHT control and USP22 KO have been identified using the linear model-based approach limma R package [85]. An adjusted P-value <0.05 was considered significant. Gene-set enrichment analysis was performed with gage R package [86] using the MSigDB [87] as gene set repository. An adjusted P-value < 0.05 was considered significant.

Multiplex quantification of cytokine secretion

Cells were seeded in 2 ml cell culture medium and supernatant was collected after 66 h, centrifuged at 300 x g, 4 °C for 5 min. and frozen in liquid nitrogen. Samples were analyzed using the LEGENDplex™ Human Anti-Virus Response Panel multiplex assay (BioLegend, San Diego, CA, USA) following the manufacturer’s protocol. The analysis was performed with the BD FACSVerse™ flow cytometer (BD Biosciences, San Jose, CA, USA). At least 300 events were acquired per analyte. Data were analyzed with the LEGENDplex v.8 software (BioLegend).

IFN ELISA

Appropriate cells were seeded two days before sample collection. Cell culture media were replaced and incubated for 24 h after which supernatants were collected, centrifuged at 300 x g, 4 °C for 5 min. and frozen in liquid nitrogen. For quantification of IFN-β, VeriKine-HSTM Human IFN Beta Serum ELISA Kit was used according to the manufacturer’s protocol (PBL Assay Science, Piscataway, United States). IFN-β concentrations were calculated using a 4-parameter fit.

Detection of type I and type III IFN production

Quantification of biologically active type I and type III IFNs in cell culture supernatants was done as previously described [88]. Briefly, HEK-Blue IFN-α/β or HEK-Blue IFN-λ (Invivogen) reporter cells were seeded (50,000 cells/well) in the presence of 20 µl of supernatants from indicated cell lines and incubated overnight. Reporter expression was detected with Quanti-Blue Solution (Invivogen) according to the manufacturer’s protocol.

Western Blot analysis

The indicated cell lines were seeded two days before lysis and treated as indicated, or left untreated. Lysis was done on ice using RIPA lysis buffer (50 mM Tris–HCl pH 8, 150 mM NaCl, 1% Nonidet P-40 (NP-40), 150 mM MgCl2, 0.5% sodium deoxycholate), with phosphatase inhibitors (1 mM sodium orthovanadate, 1 mM β-glycerophosphate, 5 mM sodium fluoride), protease inhibitor cocktail (Roche, Grenzach, Germany), 0.1% sodium dodecyl sulfate (SDS) and Pierce Universal Nuclease (Thermo Fisher Scientific) for 30 min., followed by centrifugation at 18000 x g for 25 min. at 4 °C. Protein concentrations of the cell lysates were measured using the BCA Protein Assay Kit from Pierce™, according to the manufacturer’s instructions. For Western blot detection, 20-40 µg of the lysates were boiled in Laemmli loading buffer (6x Laemmli: 360 nM Tris Base pH 6.8, 30% glycerol, 120 mg/ml SDS, 93 mg/ml dithiothreitol (DTT), 12 mg/ml bromophenol blue) at 95 °C for 5 min., followed by Western blot analysis. The following antibodies were used: rabbit anti-STING (13647S, Cell Signaling Beverly, MA, USA), rabbit anti-phospho-STAT1 (9167L, Cell Signaling), mouse anti-STAT1 (9176S, Cell signaling), rabbit anti-USP22 (#ab195298, Abcam), mouse anti-glyceraldehyde 3-phosphate dehydrogenase (GAPDH) (5G4cc, HyTest, Turku, Finland), mouse anti-Vinculin (#V9131-100UL, Merck), rabbit anti-TBK1 (ab40676, Abcam), rabbit anti-phospho-TBK1 (ab109272, Abcam), rabbit anti-Histone H2B (#07-371, Merck), mouse anti-Ubiquityl-Histone H2B (#05-1312, Merck), rabbit anti-p65 (sc-372X, Santa Cruz Biotechnologies, Santa Cruz, CA, USA), rabbit anti-phospho-p65 (3033 S, Cell Signaling), mouse anti-IRF3 (sc-33641, Santa Cruz), rabbit anti-phospho-IRF3 (4947S, Cell Signaling), rabbit anti-RIG-I (3743S, Cell Signaling), rabbit anti-MDA5 (5321S, Cell Signaling), rabbit anti-TLR3 (6961S, Cell Signaling), mouse anti-ISG56 (PA3-848, Thermo scientific), rabbit anti-MX1 (37849S, Cell Signaling), rabbit anti-IRF9 (76684S, Cell Signaling), rabbit anti-ISG20 (PA5-30073, Thermo scientific), rabbit anti-γ-H2AX (phospho Ser139) (NB100-384, Novus Biologicals) and mouse anti-NF-κB p52 (05-361, Millipore). Secondary antibodies labeled with horseradish peroxidase (HRP) were used for detection with enhanced chemiluminescence (Amersham Bioscience, Freiburg, Germany). HRP-conjugated goat anti-mouse IgG (ab6789, Abcam) was diluted at 1:10000 and HRP-conjugated goat anti-rabbit IgG (ab6721, Abcam) was diluted 1:30,000 in 5% milk powder in PBS with 0.2% Tween 20 (PBS-T). When necessary, membranes were stripped using 0.4 M NaOH for 10 min., followed by 1 h of blocking and incubation with a second primary antibody. Representative blots of at least two independent experiments are shown. When detected on separate membranes, only one representative loading control is shown for clarity. Scans of uncropped Western blots are shown in the Supplementary Information.

Stimulation of STING with 2’3’-cGAMP

The indicated cell lines were seeded 24 or 48 h prior to stimulation in P/S-free cell culture medium. For stimulation, culture medium was removed and cell lines were permeabilized by incubation with digitonin buffer (50 mM HEPES, 100 mM KCl, 3 mM MgCl2, 0.1 mM dithiothreitol, 85 mM sucrose, 0.2% bovine serum albumin, 1 mM ATP, 5 µg/ml Digitonin) pH 7 in the presence or absence of 10 µg/ml 2’3’-cGAMP for 10 min. at 37 °C. After incubation, the permeabilization buffer was replaced with P/S-free cell culture medium and further incubated at 37 °C/5% CO2 for the indicated time points.

PRR stimulation with poly(I:C) and ISD

The indicated HT-29 cells were seeded 24 h prior to treatment in sterile 6-well plates (Greiner). For each well, two µg of ISD (Invivogen) were pre-mixed with OptiMEM and, after 5 min. incubation at room temperature, mixed with premixed Lipofectamine2000-OptiMEM at a ratio of 3:1, according to the manufacturer’s instructions. After incubation for 15 min. at room temperature, the indicated transfection mixes were added to the cells in P/S free medium. Cell lysis with RIPA or RNA lysis buffer was performed after 24 h. For stimulation with poly(I:C), the indicated HT-29 cells were seeded as described above and for each well, 2 µg of poly(I:C) was mixed with 20 µl LyoVec (Invivogen), incubated for 15 min. at room temperature to allow the formation of lipid-RNA complexes. The transfection mix was then added to the indicated HT-29 cells in P/S free medium at a 1:20 volume ratio and incubated for 24 h, after which cells were processed for Western blot or RNA isolation.

Blockage of type I/II IFN signaling

Indicated cell lines were seeded 24 h prior to treatment with 1 µg/ml anti-IFNAR2 antibody (clone MMHAR-2, PBL Assay Science). After 24 h, cells were analyzed using Western Blotting. As positive control, cells were pre-treated with 1 µg/ml anti-IFNAR2 antibody for 1 h, followed by 0.05 ng/ml recombinant human IFN-β for 1 h.

Tandem Ubiquitin Binding Entity (TUBE) pull-down analysis

Ubiquitinated proteins were enriched using GST-tagged tandem-repeated ubiquitin-binding entities (TUBEs) [89], as described before [53]. Briefly, the indicated cell lines were seeded 48 h prior to lysis and/or treatment, harvested in NP-40 lysis buffer (50 mM NaCl, 20 mM Tris pH 7.5, 1% NP-40, 5 mM EDTA, 10% Glycerol) supplemented with 25 mM NEM, 1 mM sodium orthovanadate, 1 mM sodium fluoride, 0.5 mM phenylmethylsulfonyl fluoride, protease inhibitor cocktail and Pierce Universal Nuclease on ice for 30 min. GST-TUBE beads were washed once with NP-40 buffer and incubated with 3 mg of protein lysate over night at 4 °C. Beads were washed four times with NP-40 buffer, followed by elution of ubiquitinated proteins by boiling in 2x Laemmli loading buffer at 96 °C for 6 min. Ubiquitinated proteins were analyzed using Western blot analysis.

SARS-CoV-2 infection

SARS-CoV-2 (strain BavPat1/2020) was obtained from the European Virology Archive and amplified in Vero E6 cells and used at passage 3. Virus titers were determined by TCID50 assay. Caco-2 cells were infected using a MOI of 1 virus particle per cell. Medium was removed from Caco-2 cells and virus was added to cells for 1 h at 37 °C. Viral supernatants were removed, infected cells were washed once with PBS, and media was added back to the cells. Virus infection was monitored 24 h post-infection.

TCID50 virus titration

Vero E6 cells were seeded (20,000 per well) in 96-well plates 24 h prior to infection. A volume of 100 µl of viral supernatant from the indicated SARS-CoV-2-infected Caco-2 cells was added to the first well. Seven 1:10 dilutions were made (all samples were performed in triplicate). Infections were allowed to proceed for 24 h. At 24 h post infection (hpi), cells were fixed in 2% paraformaldehyde (PFA) for 20 min. at room temperature. PFA was removed and cells were washed twice in PBS and permeabilized for 10 min. at room temperature in 0,5% Triton-X/PBS. Cells were blocked in a 1:2 dilution of LI-COR blocking buffer (LI-COR, Lincoln, NE, USA) for 30 min. at room temperature. Infected cells were stained with 1:1000 diluted anti-dsRNA (J2) for 1 h at room temperature, washed three times with 0.1% PBT-T, followed by incubation with secondary antibody (anti-mouse CW800) and DNA dye Draq5 (Abcam, Cambridge, UK), diluted 1:10,000 in blocking buffer and incubated for 1 h at room temperature. Cells were washed three times with 0.1% PBS-T and imaged in PBS on a LI-COR imager.

Quantification of viral RNA

At 24 hpi, RNA was extracted from infected or mock-treated Caco-2 cells using the Qiagen RNAeasy Plus Extraction Kit (Qiagen, Hilden, Germany). For quantifying the SARS-CoV-2 genome abundance in mock and infected samples, cDNA was generated using 250 ng of RNA with the iSCRIPT reverse transcriptase (BioRad, Hercules, CA, USA), according to the manufacturer’s instructions. qRT-PCR was performed using iTaq SYBR green (BioRad) following the instructions of the manufacturer and normalized on TBP. Primers were ordered at Eurofins, Luxemburg, and are listed in Supplementary Table 1.

Indirect Immunofluorescence Assay

Cells were seeded on iBIDI glass bottom 8-well chamber slides. At 24 h post-infection, cells were fixed in 4% paraformaldehyde (PFA) for 20 min. at room temperature (RT). Cells were washed and permeabilized in 0.5% Triton-X for 15 min. at RT. Primary antibody SARS-CoV NP (Sino biologicals MM05) was diluted in phosphate-buffered saline (PBS) and incubated for 1 h at RT. Cells were washed in 1X PBS three times and incubated with secondary antibodies goat-anti mouse Alexa Fluor 568 and DAPI for 45 min. at RT. Cells were washed in 1X PBS three times and maintained in PBS. Cells were imaged by epifluorescence on a Nikon Eclipse Ti-S (Nikon).

Statistical analysis

Significance was assessed using Student’s t-test (two-tailed distribution, two-sample, equal variance) using Microsoft Excel, unless indicated otherwise. P-values < 0.05 are considered significant (*P < 0.05; **P < 0.01; ***P < 0.001, n.s.: not significant).

Resource availability

Further information and requests for resources and reagents should be directed to and will be fulfilled by the corresponding author, Sjoerd J. L. van Wijk (s.wijk@kinderkrebsstiftung-frankfurt.de).

Materials availability

All unique reagents generated in this study are available from the corresponding author without restriction.