Our previous work shows that following TNF stimulation in T cells, phosphorylation of RIPK1 within complex I is dependent upon IKK (8). We therefore asked whether Ser 25 of RIPK1 was phosphorylated in mouse T cells and if it was dependent on IKK activity. Mouse thymocytes were stimulated with TNF and then cytosolic extracts analysed by immunoblotting. Shortly after TNF stimulation and mouse thymocytes, RIPK1 is recruited to TNFR1 complex I and becomes modified with extensive ubiquitin chain additions and phosphorylation residues. To detect pS25 RIPK1, a phospho-specific anti-RIPK1 pSer25 anti-sera was used to immunoprecipitate pS25 RIPK1 from cell extracts (16). Immunoprecipitates were subsequently treated to remove ubiquitin and phospho groups and then total RIPK1 detected by immunoblotting. Following TNF stimulation of WT thymocytes, pS25 RIPK1 was readily detectable ( Figure 1A ). Pretreatment of cells with pan-IKK inhibitor, IKK16, prior to TNF stimulation, prevented detection of pS25 RIPK1, demonstrating that pS25 was IKK dependent. To confirm the requirement for IKK genetically, we wished to assess TNF stimulation of thymocytes lacking IKK expression. However, since IKK deletion results in a complete loss of mature thymocytes, we analysed Chuk^fx/fx Ikbkb^fx/fx huCD2^iCre (IKK1/2^ΔCD2) mice that express a kinase dead RIPK1^D138N mutant. Inactivation of RIPK1 in vivo in this way in IKK1/2^ΔCD2 RIPK1^D138N mice rescues thymocytes from cell death, and permits analysis of RIPK1 recruitment to TNF induced complex I (8). Following TNF stimulation, pS25 RIPK1 was readily detectable in thymocytes from Cre -ve litter mate controls. In contrast, no pS25 RIPK1 was detected following TNF stimulation of IKK1/2 deficient thymocytes ( Figure 1B ). Together, these data demonstrate that Ser25 of RIPK1 is a target of phosphorylation following TNF stimulation of thymocytes, and that phosphorylation is strictly IKK dependent.

We next wanted to ask whether IKK dependent phospshorylation of RIPK1Ser25 was sufficient for repression of TNF induced cell death by IKK. To do this, we took advantage of a recently described knockin mutant mouse strain that expresses an RIPK1^S25D mutant (16). In this strain, Ser25 has been mutated to aspartic acid. Aspartic acid and phosphorylated serine closely resemble one another chemically, and this mutation has been shown to have the predicted phosphomimetic properties (16). We therefore asked whether this phosphomimetic of RIPK1 Ser25 was sufficient to prevent RIPK1 induced cell death in the absence of IKK activity in T cells. Since IKK1/2^ΔCD2 mice lack mature thymocytes and have no peripheral T cells, we took a combined genetic and pharmacological approach to test the efficacy of RIPK1^S25D in blocking cell death. We generated IKK1^ΔCD2 RIPK1^S25D that specifically lack IKK1 but express IKK2. Redundancy between IKK1 and IKK2 permits near normal development and generation of T cells in the absence of IKK1 expression (3, 8, 26). Application of the highly specific IKK2 inhibitor, Bl605906, (27) to IKK1 deficient T cells results in a robust blockade of IKK activity.

We first tested the capacity of RIPK1^S25D to inhibit TNF induced apoptosis in thymocytes in vitro. Our previous work shows that thymocytes become progressively more dependent upon IKK signalling to promote survival as they transition through development, correlating with induction of RIPK1 protein expression (8). CD4⁺CD8⁺ double positive (DP) thymocytes and immature HSA^hi CD4⁺ single positive (CD4iSP) thymocytes do not require IKK for survival in vivo. In contrast, immature HSA^hi CD8⁺ SP (CD8iSP) thymocytes and both mature HSA^lo CD4⁺ (CD4mSP) and HSA^lo CD8⁺ SP (CD8mSP) thymocytes are all highly dependent on IKK expression for their survival in vivo (3). To assess TNF induced death, thymocytes from WT, IKK1^ΔCD2 and IKK1^ΔCD2RIPK1^S25D mice were cultured overnight with different doses of TNF in the presence and absence of IKK2 inhibitor and cell death assessed by flow cytometry. IKK2 inhibition alone did not render any subset of thymocytes susceptible to TNF induced cell death ( Figure 2A ). In contrast, in the presence of IKK2 inhibitor, CD8iSP, CD8mSP and CD4mSP underwent increasing extents of cell death with increasing doses of TNF. Cell death was RIPK1 kinase dependent, since further application of RIPK1 inhibitor, Nec1, completely blocked cell death in cultures. In contrast, no TNF inducible cell death was observed in cultures of thymocytes from IKK1^ΔCD2 RIPK1^S25D mice with IKK2 inhibitor. We presume that cell death was mediated by apoptosis rather than necroptosis, since thymocytes do not express MLKL (8), and previous studies show that T cells become susceptible to necroptosis only following activation (28).

We next tested whether RIPK1^S25D was sufficient to also protect mature peripheral T cells from TNF induced cell death. In contrast to thymocytes, IKK2 inhibitor alone was able to render WT T cells susceptible to RIPK1 dependent TNF induced cell death at higher concentrations of TNF. In contrast, T cells from RIPK1^S25D mice expressing normal IKK, were resistant to cell death in the face of IKK2 inhibition ( Figure 2B ). Similarly, IKK2 inhibition of IKK1 deficient peripheral T cells from IKK1^ΔCD2 mice resulted in high levels of cell death, even at lower concentrations of TNF. In this IKK1 deficient setting, RIPK1^S25D rescued both CD4⁺ and CD8⁺ T cells from TNF induced cell death at all but the highest concentrations of TNF, where reproducible low levels of cell death were observed. To test whether the low level of cell death observed in cultures of T cells from IKK1^ΔCD2RIPK1^S25D mice with IKK2 inhibitor were RIPK1 dependent, we also cultured cells in the presence of Nec1. This showed that the low level of cell death of TNF stimulated IKK1^ΔCD2RIPK1^S25D was in fact RIPK1 dependent, suggesting some residual kinase activity of RIPK1^S25D.

To test the capacity of RIPK1^S25D to block cell death in vivo, we asked whether the block in T cell development in IKK1/2^ΔCD2 mice could be overcome by the phosphomimetic RIPK1 mutant. We therefore generated IKK1/2^ΔCD2RIPK1^S25D and compared their phenotype with IKK1/2^ΔCD2 mice expressing a kinase dead RIPK1^D138N mutant. As previously reported, numbers of DP and CD4iSP thymocytes are normal in the absence of IKK expression, while CD8iSP, CD8mSP and CD4mSP subsets are profoundly reduced in numbers ( Figure 3 ). Kinase dead RIPK1 rescues numbers of CD4mSP and CD8iSP to near normal levels, while CD8mSP numbers are restored to levels approximately half way between those in IKK1/2^ΔCD2 and WT (8). Analysing thymus of IKK1/2^ΔCD2RIPK1^S25D mice revealed an identical pattern of rescue to the kinase dead RIPK1 expressing mice. CD4mSP and CD8iSP numbers were indistinguishable from IKK expressing controls, while CD8mSPs underwent a clear but incomplete rescue of cellularity ( Figure 3 ). These results suggest that RIPK1^S25D mutant is sufficient to repress RIPK1 dependent cell death in vivo in the absence of IKK expression.

Earlier work shows that kinase dead RIPK1 permits a substantial rescue of CD4⁺ naive T cell compartment and a small but significant rescue of CD8⁺ naive T cells (8). We therefore compared the size and composition of the peripheral T cell compartments of IKK1/2^ΔCD2RIPK1^D138N and IKK1/2^ΔCD2RIPK1^S25D mice. Substantial naive CD4⁺ compartments were evident in both IKK1/2^ΔCD2RIPK1^D138N and IKK1/2^ΔCD2RIPK1^S25D mice ( Figure 4A ), with similar numbers of recovered in both strains ( Figure 4B ). Significant rescue of CD8⁺ naive T cells was evident, albeit at a lower level than observed in the CD4⁺ compartment, in IKK1/2^ΔCD2RIPK1^D138N mice. There was some evidence of rescue in CD8⁺ naive T cell numbers in IKK1/2^ΔCD2RIPK1^S25D mice, but not to significant levels. The various IKK1/2^ΔCD2 strains also carried a Rosa26^RYFP Cre reporter construct. This is useful in the context of gene deletions resulting in strong developmental blocks, because rare cells in which iCre fails to completely delete conditional genes, so called ‘escapants’, can have a competitive advantage over gene deleted counterparts, and may accumulate in peripheral lymphoid tissues. Indeed, there are substantial fractions of YFP-ve cells amongst the naive T cells present in periphery of IKK1/2^ΔCD2 mice, indicating the presence of such escapants. Analysing YFP expression by naive T cells in IKK1/2^ΔCD2RIPK1^D138N and IKK1/2^ΔCD2RIPK1^S25D mice revealed a strong restoration of YFP expressing cells, further indicating a degree of rescue of gene deleted populations in these mice. In contrast, amongst effector phenotype and CD25⁺ Treg populations, YFP expression remained less than 5% in all IKK deficient strains (data not shown) indicating no rescue of these subsets by either RIPK1^S25D or RIPK1^D138N mutants. Nevertheless, rescue of CD4⁺ naive T cells in both RIPK1 mutant strains was far more extensive than observed in CD8⁺ subsets.

Mice with conditional alleles of Ikbkb (30) and/or Chuk (31) were intercrossed with mice expressing Cre under the control of the human CD2 (huCD2^iCre ) (32), and with mice with a D138N mutation in Ripk1 (RIPK1^D138N) (33) or with a Ripk1^S25D mutation (16). Chuk^fx/fx huCD2^iCre (IKK1^ΔCD2), Chuk^fx/fx Ikbkb^fx/fx huCD2^iCre (IKK1/2^ΔCD2), Chuk^fx/fx Ikbkb^fx/fx huCD2^iCre RIPK1^D138N (IKK1/2^ΔCD2 RIPK1^D138N), Chuk^fx/fx Ikbkb^fx/fx huCD2^iCre RIPK1^S25D (IKK1/2^ΔCD2 RIPK1^D138N), Chuk^fx/fx huCD2^iCre RIPK1^S25D(IKK1^ΔCD2 RIPK1^S25D) strains were bred in the Comparative Biology Unit of the Royal Free UCL campus and at Charles River laboratories, Manston, UK. Animal experiments were performed according to institutional guidelines and Home Office regulations.

Flow cytometric analysis was performed with 2-5 x 10⁶ thymocytes, 1-5 x 10⁶ lymph node or spleen cells. Cell concentrations of thymocytes, lymph nodes (superficial cervical, mandibular, axillary, superficial inguinal, and mesenteric chain) and spleen cells were determined with a Scharf Instruments Casy Counter. Cells were incubated with saturating concentrations of antibodies in 100 μl of Dulbecco’s phosphate-buffered saline (PBS) containing bovine serum albumin (BSA, 0.1%) for 1hour at 4°C followed by two washes in PBS-BSA. Panels used the following mAb: EF450-conjugated antibody against CD25(ThermoFisher Scientific), PE-conjugated antibody against CD127 (ThermoFisher Scientific), BV785-conjugated CD44 antibody (Biolegend), BV650-conjugated antibody against CD4 (Biolegend), BUV395-conjugated antibody against CD8 (BD Biosciences), BUV737-conjugated antibody against CD24 (BD Biosciences), PerCP-cy5.5-conjugated antibody against TCR (Tonbo Biosciences). Cell viability was determined using LIVE/DEAD cell stain kit (Invitrogen Molecular Probes), following the manufacturer’s protocol. multi-color flow cytometric staining was analyzed on a LSRFortessa (Becton Dickinson) instrument, and data analysis and color compensations were performed with FlowJo V10 software (TreeStar). Naive peripheral T cells were identified by gating CD4⁺ or CD8⁺ subsets with TCR^hi CD44^lo CD25^lo. Mature CD4⁺ and CD8⁺ SP thymocytes were identified as TCR^hiCD4⁺CD8^-HSA^lo and TCR^hiCD4^-CD8⁺HSA^lo respectively.

Thymocytes and LN T cells were cultured at 37°C with 5% CO2 in RPMI-1640 (Gibco, Invitrogen Corporation, CA) supplemented with 10% (v/v) fetal bovine serum (FBS) (Gibco Invitrogen), 0.1% (v/v) 2-mercaptoethanol βME (Sigma Aldrich) and 1% (v/v) penicillin-streptomycin (Gibco Invitrogen). Recombinant TNF (Peprotech) was supplemented to cultures at 20ng/ml, unless otherwise stated, with PBS used as vehicle. Inhibitors were used at the following concentrations, unless otherwise stated: IKK2 inhibitor BI605906 (Tocris Bio-techne) (IKK2i) (10µM in 0.1% DMSO vehicle), Nec1 (10µM in 0.1% DMSO) (Insight biotechnology).

Thymocytes (2 x 10⁷/condition) were stimulated with TNF (50ng/ml) for 5 minutes, washed two times in ice-cold PBS and lysed in NP-40 lysis buffer. pSer25 RIPK1 complexes were immunoprecipitated, followed by deubiquitylated (by USP2 treatment) and dephosphorylated (by lambda protein phosphatase treatment) as described elsewhere (16). Enzymatic reactions were allowed to proceed for 30 min at 37°C and subsequently quenched by the addition of 12.5 µL 5x laemmli buffer. IPs and total cell lysates were analyzed by standard immunoblotting. Generation of anti-pSer25 RIPK1 sera and its utility for detecting pSer25 RIPK1 by immunoprecipitation is characterised and described in detail elsewhere (16).

Statistical analysis, line fitting, regression analysis, and figure preparation were performed using Graphpad Prism 8. Column data compared by unpaired Mann-Witney student’s t test. * p<0.05, ** p<0.01, *** p<0.001, **** p < 0.0001.