The study was approved by the IRB committee at Sheba Medical Center and Rabin Medical Center. Informed written consent was obtained from all individual participants included in the study, including RIPK1-deficient patients, their parents, control subjects and patients with active CD (see below). Whole exome sequencing (WES) was performed for the two index patients using an Agilent v5 Sureselect capture kit and Illumina 2500 sequencing technology. Paired end reads (2X100 bp) were obtained, processed and mapped to the genome. The average sequencing depth of the target region is 92X with 95.63% of bases reached at least 10X coverage. We used the BWA mem algorithm (version 0.7.12) (12) for alignment of the sequence reads to the human reference genome (hg19). The HaplotypeCaller algorithm of GATK version 3.4 was applied for variant calling, as recommended in the best practice pipeline (13). KGG-seq v.08 was used for annotation of identified variants (14) and in house scripts were applied for filtering based on family pedigree and local dataset of variants detected in previous sequencing projects. Likely pathogenicity was assessed if the variant was truncating (splicing or non-sense) or missense; in-frame indels were considered if they were predicted to be pathogenic by online prediction tools including PolyPhen-2, SIFT, CADD and MutationAssessor.

Cell surface markers of peripheral blood mononuclear cells (PBMCs) were determined by immunofluorescent staining using flow cytometry (Navios, Beckman Coulter, Brea, CA, US) with antibodies purchased from Beckman Coulter. T cell receptor excision circles (TREC) analysis was performed using DNA extracted from the patient’s PBMCs. The amount of signal joint TREC copies per DNA content was determined by real-time quantitative PCR as previously described (15).

Exons 4 and 11 of RIPK1 gene were amplified and sequenced by the Sanger method. Briefly, polymerase chain reaction (PCR) amplification was performed using three sets of the following primers:

exon 4-Fw: CAGAATTTCATGTGAACGTTTCCT

exon 4-Rw: GGCTAAGTCCTCACAAGCAGAA

exon 11-Fw: CTGCCAGTGCATCAACAGCTA

exon 11 Rw: CCCATTCTCCAGCTATGAAGTACA

The PCR reaction took place in a 25-μL volume containing 50 ng of DNA, 10 ng of each primer, 1.5 mM dNTPs, in 1.5 mM MgCl2, PCR buffer, with 1.2 units of Taq polymerase (Bio-Line, London, UK). After an initial denaturation of 5 min at 95°C, 30 cycles were performed (94°C for 30 s, 60°C for 30 s, and 72°C for 30 s), followed by a final extension of 10 min at 72°C. PCR amplicons were sequenced in both directions using a commercial sequencing service [Hy Laboratories, Ltd. (Hylabs) Park Tamar Rehovot, Israel].

We obtained formalin-fixed, paraffin-embedded (FFPE) rectal mucosa specimens from the two RIPK1-deficient patients. As controls we used samples from two control subjects with normal colonoscopies, without a history of IBD (Control 1, 16-years-old male evalulated for bloody stools; Control 2, 14-years-old female with recurrent diarrhea), and two patients with CD (Patient 1, 15-years-old male with active ileo-colonic disease; Patient 2, 10-years-old female with active colonic disease). Rectal FFPE specimens were cut into 5 μm-thick sequential sections, that were then dewaxed in xylene and rehydrated stepwise in descending ethanol series followed by antigen retrieval with Antigen Unmasking Solution, Citrate-Based (Vector laboratories, H-3300) at 96°C for 10 min. The slides were then washed with PBS followed by permeabilization and blocking with PBS with 1% BSA, 3%NDS, 0.05% Tween, 0.025% Triton for 30 minutes at room temperature. The sections were then incubated with primary rabbit anti-human RIPK1 monoclonal antibody D94C12 (3493, Cell signaling technology, dilution 1:50), overnight at 4°C, followed by staining with the corresponding secondary antibodies for 1 hour at room temperature and counterstaining with DAPI, which was included in the mounting medium (GBI Labs, E-19-18, Mukilteo, WA). Samples were visualized by confocal microscope (LSM 800, Zeiss, Oberkochen, Germany). All images within each experiment were acquired under the same conditions.

Visualization of the structural consequences of the mutations to RIPK1 was performed using PyMOL (Version 2.4.1) by using the mutagenesis wizard on the indicated experimentally determined structures.

Peripheral blood mononuclear cells (PBMCs) from a RIPK1-deficient patient, both parents, three controls and three patients with active CD were stained with a panel of metal-chylated surface antibodies targeting markers of major immune cell lineages ( Supplementary Table 1 ) per previously published protocol (16). For controls we used PBMCs from 2 females and 1 male aged 14-17 years, with normal endoscopic evaluation and no concern of IBD, while for the CD group we obtained PBMCs from 3 males, aged 11-17 years, one treated with infliximab and two newly-diagnosed treatment-naïve.

The samples were run on Helios2 mass cytometer (Fluidigm, San Francisco, CA, USA). FCS files obtained were analyzed with premium Cytobank software and pre-gated on CD45⁺/viable/single/DNA⁺ events before initiating the analysis. Normalization beads were used and gated out of the analysis. The data was automatically clustered with Pheongraph through cytofkit package in R and visualized with tSNE using all leukocytes (CD45⁺ cells) as input and manually labeled based on markers expressed in the individual clusters. Cluster abundance (% of CD45⁺ viable single events) and fold change over the RIPK1 abundance were computed and plotted for comparison between groups using Prism8 software. Additionally, clusters of similar cellular subtype were combined, and similarly, the ratio of CD4/CD8 cells was calculated.

To identify dysregulation in activation and cytokine production of the immune cells in RIPK1 deficiency, we stimulated PBMCs from Patient 1 and his parents with either lipopolysaccharide (LPS) or phorbol 12-myristate 13-acetate (PMA) and Ionomycin. PBMC’s from patient and parents were thawed and washed with T cell media, then centrifuged at 300g. The supernatant was discarded and pelleted PBMC’s were resuspended in 1ml of T Cell media. GolgiStop and GolgiPlug were then added to each stimulation tube and each sample was stimulated with either LPS 1mg/ml or PMA 0.5ng/ml and Ionomycin 1mg/ml for 4 hours at 37°C. Samples were then centrifuged and washed again in T cell media then prepped for CyTOF where they were stained with heavy metal chelated antibodies ( Supplementary Table 2 ) first for surface antigens, then fixed and permeabilized and finally stained with antibodies for intracellular antigens (16). All samples were analyzed as above. In addition, mean expression of various cytokines (mean metal intensity) was calculated in the various populations indicated.

Patient 1 was referred at the age of 16 months for evaluation of IBD. He was born to a consanguineous Muslim family (parents first degree cousins) and presented at the age of 9 months with fever and diarrhea that were initially attributed to amebiasis. Despite metronidazole treatment his condition deteriorated, and he developed perianal abscesses with multiple fistulas. Blood tests were significant for anemia (hemoglobin 6.5 gram/dL), hypoalbuminemia (albumin 1.5 gr/dL) and elevated inflammatory markers (CRP 94 mg/L, normal <5 mg/L). Due to the severity of the perianal disease, the patient underwent double barrel protective ileostomy. Despite this intervention, his condition did not improve. At the age of 16 months the patient appearead cachectic. Weight was 7.4 kg (Z score -3.2) and length 73 cm (Z score -2.8). Abdominal exam showed double barrel ileostomy, with severe inflammation in the surrounding skin, and multiple perianal fistulas with deep ulcerations and perianal fissures. Colonoscopy demonstrated patchy areas of severe colonic ulcerations with pseudopolyps, with normal mucosa in between, while an upper endoscopy was unremarkable. Colonic biopsies revealed chronic active inflammation with focal cryptitis.

Following a course of broad-spectrum antibiotics and nutritional supplementation, the patient was started on metronidazole, mesalamine and azathioprine, and later began adalimumab. Nevertheless, the patient failed to respond to this TNFα antagonist despite adequate drug levels and absence of anti-drug antibodies, thus experiencing primary pharmacodynamic failure, and at the age of 2.5 years developed colo-vesical fistula that required fistulectomy. Follow up colonoscopy showed severe ulcerations at the sigmoid and ascending colon. Anakinra, an IL1 receptor antagonist, was commenced with escalation of the dose up to 3mg/kg daily, though clinical or laboratory responses were not documented. Importantly, the patient did not develop severe or atypical infections until the age of 3 years.

Patient 2 was also born to a consanguineous Muslim family (parents first degree cousins) and presented at the age of 1 month with recurrent fevers, non-bloody diarrhea, oral ulcers, poor weight gain, abdominal distention and arthritis. During infancy he developed recurrent perianal abscesses. At the age of 20 months his weight was 9.7 kg (Z score -1.4) and length 81 cm (Z score -1.1). Physical exam was noted for abdominal distention and hepatosplenomegaly, along with a draining perianal abscess. Blood tests demonstrated anemia (hemoglobin 8.8 g/dL) and mildly elevated inflammatory markers (ESR 29 mm/hour). Blood cultures were positive for Providencia stuartii and Pseudomonas aeruginosa, which were treated with Piperacillin-Tazobactam for two weeks. Colonoscopy showed patchy colitis, predominantly at the right colon, while an upper endoscopy was unremarkable. Pathologic assessment showed patchy chronic active colitis. The patient temporarily responded to mesalamine, metronidazole and azathioprine, with a 3 kg weight gain over 10 months and normalization of inflammatory markers. However, he continued to suffer from recurrent febrile episodes, oral ulcers, diarrhea, arthritis and perianal abscesses.

Immunoglobulin levels, including IgM, IgA and IgG were within normal limits for both patients ( Supplementary Table 3 ). Moreover, immunoglobulin E was also normal for Patient 1. In addition, lymphocyte subset analysis was within normal limits for both patients, beside slightly elevated CD8⁺ T cells ( Supplementary Table 3 ). Finally, TREC levels for both patients were within normal limits, reflecting intact thymic function.

Following WES analysis, we identified 7,928 and 6,887 homozygous variants for Patients 1 and 2, respectively, that affect protein sequences. These numbers were reduced to 49 and 77 variants, respectively, after filtration for common variants (MAF ≥0.01) in either our local in-house exomes database (n~3500) or external databases such as 1000 Genomes Project (1 KG; https://www.internationalgenome.org/1000-genomes-browsers) or dbSNP 135 database, the NHLBI Exome Sequencing Project (ESP) (http://evs.gs.washington.edu/EVS/) or gnomAD database (https://gnomad.broadinstitute.org/). A c.1934C>T missense mutation residing in Exon 11 was identified in Patient 1, and a c.580G>A RIPK1 missense mutation residing in Exon 4 was identified in Patient 2 ( Figures 1A, B ). Both mutations were verified using Sanger sequencing. The c.1934C>T mutation has been charactarized previously as a deleterious variant (9). The c.580G>A mutation was assessed as damaging by SIFT and POLYPHEN2 as well as by other tools that predict pathogenicity and had a high CADD score over 29. Moreover the Alanine at position 194 is evolutionarily conserved across most vertebrate species which suggests that it has a role in the RIPK1 function.

To further understand how these variants affect protein function, we visualized RIPK1 structural sequences. RIPK1 is composed of two major structured domains: the kinase domain (aa 17-289) and a death-domain (aa 583-669). There are numerous x-ray structures of the kinase domain (e.g. PDB: 4ITJ) and one crystal structure of the death-domain (PDB: 6AC5). We used these structures to visualize the potential consequence of the mutations we identified.

Thr645 mediates a hydrogen bond to a structured water molecule, that is also bound by neighboring backbone carbonyl of Asn578 ( Figure 1C ). A mutation to Methionine ( Figure 1D ), as identified in Patient 1, would have to displace this water molecule, potentially disrupting the local environment. Moreover, it was shown that even conservative mutations at the death domain (K599R) are sufficient to impair its function (17). In addition, Ala194 is buried in the C’ lobe of the kinase domain ( Figure 1E ). Its mutation to Thr, as identified in Patient 2, may cause significant clashes and as such disrupt the packing and conformation of the kinase domain, which may hamper its activity ( Figure 1F ).

Finally, we performed immunofluorescence staining for RIPK1 in rectal biopsies from the two patients presented above with RIPK1 deficiency. Results were compared to staining in two non-IBD subjects (evaluated for abdominal pain and diarrhea, with normal macroscopic and histologic colonoscopy) and two patients with active CD. In comparison to control subjects, expression of RIPK1 in patients with active CD was increased in rectal tissue ( Figures 2A, B ). Among the two index patients, there was no RIPK1 expression for Patient 1 and minimal expression for Patient 2 ( Figure 2C ).

To evaluate changes in the architecture and function of peripheral immune system in one of the RIPK1-deficient patients, we performed mass cytometry (CyTOF) using a panel of 36 antibody markers, and compared the results to both maternal and paternal blood samples, non-IBD subjects (controls) and CD patients, An unbiased clustering algorithm (Phenograph) was performed on multiple immune populations. To begin, we clustered on all immune cells (CD45⁺) and were able to identify 28 unique populations ( Figures 3A–C ). When comparing abundances of major immune subsets, we identified an increased abundance of T cells, particularly CD8 effector population, in the RIPK1-deficient patient, compared to controls, along with a decrease in all other cell types including monocytes, dendritic cells and B cells ( Figures 3C–E ). Interestingly, we also observed a reversal of the CD4⁺ to CD8⁺ ratio compared to all other samples ( Figure 3F ).

To gain further insight into the functional effects of RIPK1 deficiency, we examined cytokine production by PBMCs following stimulation with either LPS or PMA/Ionomycin (PMA/I), using CyTOF, and clustering on all leukocytes (CD45⁺ cells, Supplemental Figures 1A–C ). We were able to identify 17 unique populations, again demonstrating a CD8 predominance in the patient ( Supplemental Figures 1A, B) . RIPK1-deficient patient’s immune cells exhibited decreased IL-6 production in response to LPS, but not to PMA-I stimulation, across multiple cell types including T cells, B cells and innate immune cells ( Figure 4A ). In addition, we observed an increase in IL-22 production across the majority of immune cells, particularly in response to LPS stimulation condition ( Figure 4B ).

Focusing on the individual cell populations, monocytes and dendritic cell’s ability to produce inflammatory cytokines was drastically reduced in the patient, consistent with documented reduced NFkB activity (18) ( Figure 4C ). On the other hand, patient’s T cells produced higher amounts of IFNγ than healthy controls upon PMA/I stimulation ( Figure 4D ). Interestingly, when analyzing immune populations with inhibitory potential, including myeloid derived suppressor cells (MDSC) and CD24⁺ B cells (containing the putative regulatory B cell populations), we demonstrate higher production of inflammatory cytokines (IL-23, IL1β, and IL-8), but also IL-10, compared with cells obtained from the parents ( Figure 4C ), suggesting a possible dysfunction of these cells.