Abstract

   Receptor-interacting protein kinase 1 (RIPK1) regulates cell death and
   inflammation. Here, we show that T cell–specific RIPK1 deficiency in
   mice leads to the premature senescence of T cells and induces various
   age-related diseases, resulting in premature death. RIPK1 deficiency
   causes higher basal activation of mTORC1 (mechanistic target of
   rapamycin complex 1) that drives enhanced cytokine production,
   induction of senescence-related genes, and increased activation of
   caspase-3/7, which are restored by inhibition of mTORC1. Critically,
   normal aged T cells exhibit similar phenotypes and responses.
   Mechanistically, a combined deficiency of RIPK3 and caspase-8
   inhibition restores the impaired proliferative responses; the elevated
   activation of Akt, mTORC1, extracellular signal–regulated kinase, and
   caspase-3/7; and the increased expression of senescence-related genes
   in RIPK1-deficient CD4 T cells. Last, we revealed that the senescent
   phenotype of RIPK1-deficient and aged CD4 T cells is restored in the
   normal tissue environment. Thus, we have clarified the function of
   RIPK3 and caspase-8 in inducing CD4 T cell senescence, which is
   modulated by environmental signals.
     __________________________________________________________________

   RIPK1-deficiency in T cells induces T cell senescence by activating
   RIPK3/caspase-8, leading to age-related disease development.

INTRODUCTION

   Age-associated inflammation, called “inflammaging,” is characterized by
   chronic, low-grade, systemic inflammation with aging in the absence of
   infection, which leads to tissue degeneration ([38]1). Inflammaging is
   a risk factor for type 2 diabetes, cardiovascular diseases,
   neurodegenerative diseases, cancer, and death and leads to increased
   susceptibility to infection in the elderly ([39]2).

   Age-associated inflammation, called inflammaging, is characterized by
   chronic, low-grade, systemic inflammation with aging in the absence of
   infection, which leads to tissue degeneration ([40]1). Inflammaging is
   a risk factor for type 2 diabetes, cardiovascular diseases,
   neurodegenerative diseases, cancer, and death and leads to increased
   susceptibility to infection in the elderly ([41]2). Aging also induces
   thymic involution, which results in a decrease of naïve T cells and an
   accumulation of memory T cell subsets ([42]3). The proliferative
   response of these senescent T cells upon T cell receptor (TCR)
   stimulation is impaired, leading to reduced vaccine efficacy and
   increased susceptibility to cancer and infections in elderly people
   ([43]3, [44]4). Senescent cells secrete high levels of inflammatory
   cytokines, chemokines, matrix metalloproteinases, growth factors, and
   angiogenic factors, termed the senescent-associated secretory phenotype
   (SASP) ([45]5). The SASP factors contribute to inflammaging, contribute
   to induction of insulin resistance and fibrosis, and reinforce and
   propagate senescence by both autocrine and paracrine mechanisms
   ([46]5). Thus, the SASP mediates many of the pathological effects by
   senescent cells. Mechanistically, the SASP is controlled by enhancer
   remodeling and activation of transcription factors, such as nuclear
   factor κB (NF-κB), CCAAT/enhancer binding protein β (C/EBPβ), and GATA
   binding protein 4 (GATA4), and by activation of mitogen-activating
   protein kinases (MAPKs) and mechanistic target of rapamycin (mTOR)
   ([47]5).

   The nutrient-sensing serine/threonine kinase mTOR integrates
   extracellular signals and intracellular metabolic cues and promotes
   glycolysis, growth, and proliferation. mTOR forms two distinct protein
   complexes, mTORC1 and mTORC2, and a deficiency of both complexes in T
   cells results in impaired T cell activation and differentiation ([48]6,
   [49]7). Optimal mTOR activity is required for proper T cell functions.
   T cell–specific deletion of Tuberous sclerosis 1 (TSC1), an upstream
   negative regulator of mTORC1, abrogates naïve T cell quiescence,
   increases apoptosis, and impairs antibacterial immune responses
   ([50]8–[51]10).

   Recent studies revealed that dominant-active gain-of-function mutations
   in phosphatidylinositol 3-kinase (PI3K) result in hyperactivation of
   Akt and mTOR in T cells, leading to T cell senescence and
   immunodeficiency in humans ([52]11). Senescent T cells also exhibit
   hyperactivation of MAPK cascades, which is mediated by a complex of
   MAPKs and stress-sensing proteins, sestrins ([53]12). However, the
   molecular mechanisms of hyperactivation of mTOR and MAPKs in senescent
   T cells and the pathological role of senescent T cells remain unclear.

   RIPK1 is a key signaling molecule controlling cell death and
   inflammation through both its scaffold and kinase functions ([54]13,
   [55]14). The kinase function of RIPK1 induces caspase-8–dependent
   apoptosis and RIPK3-dependent necroptosis ([56]15–[57]18). On the other
   hand, RIPK1 inhibits apoptosis and necroptosis through its scaffold
   function, which is critical for blocking early postnatal lethality
   mediated by caspase-8 and RIPK3 ([58]19–[59]21).

   Recent studies characterized a complete RIPK1 deficiency in patients
   with biallelic loss-of-function mutations in the RIPK1 gene ([60]22,
   [61]23). The patients suffered from recurrent infections, early onset
   inflammatory bowel disease, and progressive polyarthritis ([62]22,
   [63]23). The patients had lymphopenia, suggesting the possibility of a
   functional importance of RIPK1 in T cells ([64]22). Consistent with the
   human data, a recent study demonstrated that T cell–specific deletion
   of RIPK1 leads to severe T lymphopenia in mice, whereas the immature T
   cell compartment in the thymus was intact ([65]24). Moreover,
   RIPK1-deficient [knockout (KO)] T cells were defective in TCR-induced
   proliferation and exhibited elevated caspase activation independently
   of kinase activity ([66]24). However, whether RIPK1-KO in T cells
   results in immunodeficiency and inflammatory diseases remains unknown.

   In this study, we found that T cell–specific RIPK1 deficiency leads to
   the premature senescence in T cells, which is accompanied by the SASP.
   T cell–specific RIPK1-KO (tKO) mice develop chronic inflammatory
   diseases with aging, resulting in premature death. We demonstrate that
   basal activation of Akt, mTORC1, and extracellular signal–regulated
   kinase (ERK) is enhanced in RIPK1-KO T cells and that inhibition of
   both RIPK3 and caspace-8 or mTORC1 inhibits T cell senescence, SASP,
   and caspase-3/7 activation. These basal activation phenotypes were
   similarly observed in aged wild-type (WT) T cells. We further
   demonstrated that the senescent phenotype of aged and RIPK1-KO CD4 T
   cells is restored in the normal tissue environment.

RESULTS

RIPK1 deficiency in T cells induces inflammatory diseases

   To elucidate the physiological functions of RIPK1 in T cells, we
   specifically deleted Ripk1 in T cells by crossing Ripk1^fl/fl mice with
   CD4-Cre Tg mice (designated as RIPK1-tKO). Consistent with a previous
   report ([67]24), RIPK1-tKO mice look normal until 4 months of age.
   However, they develop inflammatory diseases characterized by a reduced
   body weight, eye opacity, and infiltration of leukocytes in the
   salivary gland and lung from 5 months of age ([68]Fig. 1, A to F). We
   also observed an increased inflammatory cell infiltration in the
   bronchioalveolar lavage (BAL) of RIPK1-tKO mice ([69]Fig. 1G). Despite
   these inflammatory manifestations, RIPK1-tKO mice showed no elevation
   of serum double-stranded DNA (dsDNA) autoantibody levels as
   representative of generalized autoimmunity (fig. S1A). RIPK1-tKO mice
   also showed hindlimb paralysis ([70]Fig. 1, H and I), which was
   characterized by clenching of the hindlimbs to the body when suspended
   by the tail, a common neuropathic phenotype. Because axon abnormalities
   and subsequent degeneration of the sciatic nerve are common causes of
   hindlimb paralysis, the sciatic nerve morphology was examined. Semithin
   sections of sciatic nerve showed a remarkable decrease in myelinated
   axons and axon diameter in RIPK1-tKO mice compared to control mice
   ([71]Fig. 1, J and K). Accordingly, the infiltration of leukocytes and
   the increased expression of Ifng, Cxcl10, and Cd3e were observed in the
   sciatic nerve of RIPK1-tKO mice ([72]Fig. 1L). Last, RIPK1-tKO mice
   showed premature death; more than half of mice died by 8 months
   ([73]Fig. 1M). These results demonstrated that RIPK1 deficiency in T
   cells causes multiorgan inflammation and eventually results in
   premature death.

Fig. 1. RIPK1-tKO mice develop inflammatory disease.

   [74]Fig. 1.
   [75]Open in a new tab

   (A) Representative photograph showing the appearance of control and
   10-month-old RIPK1-tKO mice (right). (B) Body weight changes over time
   in control (n = 6) and RIPK1-tKO female mice (n = 8). (C)
   Representative photograph of a 10-month-old RIPK1-tKO mouse that has
   developed opacity of the eyes. (D) Eye opacity score of control
   (n = 10) and RIPK1-tKO (n = 13) mice. Representative hematoxylin and
   eosin (H&E)–stained sections of salivary glands (E) and lung (F). Scale
   bars, 100 μm (E) and 50 μm (F). (G) Quantification of total cells
   (left), eosinophils (middle), and neutrophils (right) in the BAL fluid
   from 7- to 10-month-old control and RIPK1-tKO mice (n = 9 per group).
   (H) Representative photograph of 8-month-old control and RIPK1-tKO mice
   suspended by the tail. RIPK1-tKO mice clench their hindlimbs to the
   body. (I) Hindlimb reflex score of control and RIPK1-tKO mice (n = 13
   per group). (J) Toluidine blue–stained semithin transverse sections of
   the sciatic nerve from 8-month-old control and RIPK1-tKO mice. Scale
   bars, 50 μm (top) and 20 μm (bottom). (K) H&E-stained semithin
   transverse sections of the sciatic nerve from 8-month-old control and
   RIPK1-tKO mice. Scale bars, 50 μm. (L) Quantitative polymerase chain
   reaction (qPCR) analysis for the expression of Ifng, Cxcl10, and Cd3e
   in sciatic nerve from 10-month-old control and RIPK1-tKO mice. (M)
   Survival of control (n = 8) and RIPK1-tKO (n = 11) mice. Error bars
   represent SEM. Data are representative of at least three independent
   experiments. *P < 0.05, **P < 0.01, ***P < 0.005, and ****P < 0.001
   using Student’s t test.

RIPK1 deficiency causes premature CD4 T cell senescence

   To understand how the RIPK1-tKO causes inflammatory diseases with
   aging, the cellularity of the spleen in young (2 to 4 months,
   apparently healthy) and middle-aged (6 to 10 months, already developed
   inflammatory diseases) RIPK1-tKO mice was first assessed. The total
   number of spleen cells was increased in middle-aged but not young
   RIPK1-tKO mice compared with control mice, whereas the numbers of both
   CD4 and CD8 T cells were markedly decreased in RIPK1-tKO mice ([76]Fig
   2A and fig. S2, A and B). Whereas the expression of RIPK1 in RIPK1
   heterozygous CD4 T cells is half of WT CD4 T cells (fig. S2C), the
   number of CD4 and CD8 T cells in the spleen from RIPK1 heterozygous
   mice was comparable to that of WT mice (fig. S2D). The percentage and
   number of CD44^hi CD62L^lo effector/memory phenotype CD4 T cells
   markedly increased, whereas that of CD44^lo CD62L^hi naïve CD4 T cells
   inversely decreased in RIPK1-tKO mice with aging ([77]Fig. 2, B and C,
   and fig. S2, E and F).

Fig. 2. RIPK1 deficiency in T cells induces CD4 T cell senescence.

   [78]Fig. 2.
   [79]Open in a new tab

   (A) Total cell numbers in the spleen of young (n = 9) and middle-aged
   (n = 17) control or RIPK1-tKO mice. (B and C) The percentage of CD44^hi
   CD62L^lo cells in CD4 T cells (B) and CD44^lo CD62L^hi cells in CD4 T
   cells (C) from the spleen of young (n = 4) and middle-aged (n = 6)
   control or RIPK1-tKO mice. (D) RNA-seq data showing the relative
   expression of senescence-related genes in CD4^+ CD25^− T cells of
   spleen from 10-month-old control and RIPK1-tKO mice (n = 2). (E) CD4^+
   T cells from control or RIPK1-tKO mice were stimulated with immobilized
   anti-CD3ε plus anti-CD28 (anti-CD3/CD28). Cell growth (top left) and
   cytokine production were assessed after 48 hours of stimulation by a
   water soluble tetrazolium salt-8 (WST-8) proliferation assay and
   electrochemiluminescence (ECL), respectively. OD[450], optical density
   at 450 nm; Unstim, unstimulated. (F) Serum cytokine levels in young and
   middle-aged control (n = 8) and RIPK1-tKO (n = 8) mice were determined
   by ECL. Error bars represent SEM. Data are representative of at least
   three independent experiments (E and F). *P < 0.05, **P < 0.01,
   ***P < 0.005, and ****P < 0.001 using Student’s t test.

   We next assessed the contribution of RIPK1-KO regulatory T cells
   (T[regs]) to the inflammatory phenotype of RIPK1-tKO mice using
   T[reg]-specific RIPK1-deficient mice by crossing RIPK1^fl/fl mice with
   Foxp3YFP-Cre mice (designated as Foxp3cre RIPK1^fl/fl). Foxp3cre
   RIPK1^fl/fl mice did not develop inflammatory diseases such as the eye
   opacity and hindlimb paralysis (fig. S2, G and H), indicating that the
   inflammatory phenotype of RIPK1-tKO mice was not attributed to the
   dysregulation of RIPK1-KO T[regs].

   To identify the molecular mechanism by which RIPK1-KO T cells drive
   inflammation, we sorted CD4^+ CD25^− T cells of the spleen from
   10-month-old control or RIPK1-tKO mice (fig. S2I), and RNA sequencing
   (RNA-seq) analysis was performed. We found that the expression of Spp1
   encoding osteopontin, a marker of senescent CD4 T cells, was markedly
   up-regulated in RIPK1-KO CD4 T cells ([80]Fig. 2D) ([81]25, [82]26),
   suggesting that RIPK1-KO T cells exhibit a cellular senescence
   phenotype. The expression of other cellular senescence and SASP-related
   genes represented by Gzma/b/k, Adam8/9/19, Vegfb, Pdgfb, Ctsb/e, Ccl5,
   Il1r2, Prdm1, and Pdcd1 was also up-regulated in RIPK1-KO CD4 T cells
   ([83]Fig. 2D and fig. S2J). The expression of genes encoding
   proinflammatory cytokines and chemokines (fig. S2J), proinflammatory
   transcriptional factors such as NF-κB and interferon (IFN) regulatory
   factors (fig. S2K), and cytokine and chemokine receptors (fig. S2L) was
   increased in RIPK1-KO CD4 T cells. RIPK1-KO CD4 T cells produced higher
   amounts of cytokines such as IFN-γ, interleukin-17A (IL-17A), IL-4, and
   IL-5, although their proliferative responses upon TCR stimulation were
   impaired ([84]Fig. 2E). Thus, RIPK1-KO CD4 T cells functionally exhibit
   the characteristics of senescent T cells. We also analyzed the T cells
   from the lymph nodes in RIPK1-tKO mice and found that the expression of
   senescence-related genes was increased in CD4 T cells of the lymph
   nodes from RIPK1-tKO mice (fig. S2M). Next, the serum levels of
   inflammatory cytokines were examined, and the levels of IFN-γ, IL-2,
   IL-5, IL-17A, and tumor necrosis factor–α (TNF-α) were found to be
   increased in RIPK1-tKO mice ([85]Fig. 2F). Similarly, the serum levels
   of immunoglobulin M (IgM), IgG, and IgE were also significantly
   enhanced in these mice ([86]Fig. 2F). Notably, the serum levels of
   inflammatory cytokines except for IL-17A were increased even in young
   RIPK1-tKO mice that appear healthy. These data indicate that RIPK1-KO T
   cells induce chronic inflammation from an early age, resulting in
   premature inflammatory diseases in middle-aged RIPK1-tKO mice.

RIPK1-tKO mice develop age-associated diseases

   Next, we examined whether RIPK1-KO T cells induce age-associated
   diseases. Because inflammaging is a risk factor for type 2 diabetes and
   recurrent hypoglycemia is common in elderly people ([87]27), we first
   performed insulin tolerance tests (ITTs) and glucose tolerance tests
   (GTTs). Blood glucose levels were significantly decreased in RIPK1-tKO
   mice ([88]Fig 3, A and B), although food intake and serum insulin
   levels were comparable to those of control mice ([89]Fig. 3C and fig.
   S3A), suggesting that RIPK1-tKO mice develop hypoglycemia. Second,
   inflammaging plays a role in the increased risk of age-related
   sarcopenia, which is characterized by progressive loss of muscle mass,
   strength, and functions ([90]28). The grip strength test revealed a
   reduction of skeletal muscle strength in RIPK1-tKO mice ([91]Fig. 3D);
   consistently, tibialis muscle mass was markedly reduced ([92]Fig. 3, E
   and F). Further, the increased expression of Ifng, Cxcl10, and Cd3e was
   observed in the tibialis muscle of RIPK1-tKO mice ([93]Fig. 3G). These
   data indicate that RIPK1-KO T cells induced muscle inflammation and
   subsequent sarcopenia. Furthermore, RIPK1-tKO mice exhibited a marked
   reduction of adipose tissue mass and smaller adipocytes (fig. S3, B and
   C).

Fig. 3. RIPK1-tKO mice develop age-related diseases.

   [94]Fig. 3.
   [95]Open in a new tab

   (A) ITTs were performed in 6-month-old control (n = 6) and RIPK1-tKO
   (n = 3) mice. Mice were fasted for 4 hours and then injected
   intraperitoneally with insulin. Blood glucose levels were measured at
   the indicated times after insulin injection. (B) GTTs were performed in
   6-month-old control (n = 6) and RIPK1-tKO (n = 3) mice. Mice were
   fasted for 16 hours, and glucose was then injected intraperitoneally.
   Blood glucose levels were measured at the indicated time points after
   glucose injection. (C) Food intake in 6-month-old control and RIPK1-tKO
   mice was measured for 24 hours. (D) Hindlimb grip strength analysis of
   8-month-old control and RIPK1-tKO mice (n = 6 per group). (E)
   Representative photograph of the tibialis muscle from 7-month-old
   control and RIPK1-tKO mice. (F) Weight of tibialis muscle from
   7-month-old control and RIPK1-tKO mice (n = 6 per group). (G) qPCR
   analysis for the expression of Ifng, Cxcl10, and Cd3e in the tibialis
   muscle from 12-month-old control and RIPK1-tKO mice. (H) Serum
   metabolome analysis of 10-month-old control and RIPK1-tKO mice. (SDMA
   and ADMA, respectively) (n = 5 per group). (I) Blood hemoglobin levels
   of young (n = 5) and middle-aged (n = 6) control and RIPK1-tKO mice
   were determined by enzyme-linked immunosorbent assay (ELISA). (J)
   Heart/body weight ratios of 10- to 13-month-old control and RIPK1-tKO
   mice (n = 6 per group). (K) qPCR analysis for the expression of Ifng,
   Cxcl10, and Cd3e in the heart from 12-month-old control and RIPK1-tKO
   mice. (L) Quantitative measurements of β-galactosidase (β-gal) activity
   in kidney lysates of 10- to 13-month-old control and RIPK1-tKO mice by
   colorimetric assay (n = 5 per group). Error bars represent SEM. Data
   are representative of at least three independent experiments (A to G
   and I to L). *P < 0.05, **P < 0.01, ***P < 0.005, and ****P < 0.001
   using Student’s t test.

   To evaluate the morbidity of RIPK1-tKO mice, serum metabolomic
   profiling was performed. (i) The serum kynurenine/tryptophan ratio as a
   biomarker of inflammaging and cancer ([96]29, [97]30) increased in both
   young and middle-aged RIPK1-tKO mice ([98]Fig. 3H), confirming that
   RIPK1-KO T cells cause chronic inflammation from a young age. (ii)
   Because anemia of inflammation is prevalent in patients with
   inflammation, such as infections, autoimmune diseases, and cancer
   ([99]31), the blood hemoglobin levels were assessed. We observed that
   middle-aged but not young RIPK1-tKO mice developed anemia
   ([100]Fig. 3I). (iii) Serum-oxidized glutathione and ophthalmic acid,
   biomarkers of oxidative stress, which are thought to drive multiple
   age-related diseases ([101]32–[102]35), were increased in middle-aged
   but not young RIPK1-tKO mice ([103]Fig 3H). Because the opacity of the
   lens is thought to be a direct result of oxidative stress ([104]36),
   oxidative stress may cause the lens opacity observed in RIPK1-tKO mice
   ([105]Fig. 1, C and D). Furthermore, (iv) serum symmetric and
   asymmetric dimethylarginine (SDMA and ADMA), known as risk factors for
   cardiovascular disease ([106]37, [107]38), increased in middle-aged but
   not young RIPK1-tKO mice ([108]Fig. 3H). Inversely, serum carnitine,
   beneficial for cardiovascular diseases ([109]39, [110]40), was reduced
   in middle-aged but not young RIPK1-tKO mice ([111]Fig. 3H).
   Consistently, RIPK1-tKO mice showed significant increases in heart
   /body weight ratios and increased expression of Ifng, Cxcl10, and Cd3e
   in the heart ([112]Fig. 3, J and K), suggesting that RIPK1-KO T cells
   also induced cardiac inflammation. (v) We further observed that serum
   pantothenic acid (vitamin B5), which is significantly decreased in both
   Alzheimer’s disease and Huntington’s disease ([113]41, [114]42), was
   also decreased in middle-aged RIPK1-tKO mice ([115]Fig. 3H). (vi) Serum
   uric acid, which is commonly elevated in subjects with chronic kidney
   disease ([116]43, [117]44), was increased in middle-aged RIPK1-tKO mice
   ([118]Fig. 3I). Consistently, the activity of senescent-associated (SA)
   β-galactosidase was higher in kidneys from RIPK1-tKO mice than from
   control mice ([119]Fig. 3M), indicating that senescence is also induced
   in the kidneys of RIPK1-tKO mice. Together, these data indicate that
   RIPK1 deficiency in T cells causes chronic inflammation and,
   consequently, the premature development of age-related multimorbidity.

RIPK1 inhibits the activation of Akt, mTORC1, and ERK signaling

   To elucidate the molecular mechanism underlying the multiorgan cellular
   senescence induced by RIPK1-KO T cells, we sought to identify the
   signaling pathways regulated by RIPK1 by pathway analysis of genes
   up-regulated in RIPK1-KO CD4 T cells as compared to WT CD4 T cells. The
   expression of cell cycle–related genes was clearly up-regulated in
   RIPK1-KO CD4 T cells ([120]Fig. 4A). Because the mTORC1 pathway is
   critical for cell cycle regulation upon TCR stimulation ([121]7), the
   activation of mTORC1 signals in RIPK1-KO naïve CD4 T cells was
   assessed. RIPK1-KO naïve CD4 T cells ex vivo showed high basal levels
   of mTORC1 activation (phosphorylation of S6 and 4E-BP1) and ERK
   phosphorylation (fig. S4, A and B). We similarly observed that basal
   activation levels of mTORC1 and ERK, but not p38, were also enhanced in
   RIPK1-KO effector CD4 T [T helper (T[H])] cells and that this was
   impaired by treatment with rapamycin, an mTOR inhibitor ([122]Fig. 4B
   and fig. S4C). RIPK1-KO T[H] cells showed increased cell size, enhanced
   cell death, and decreased viable cells, which was also restored by
   rapamycin treatment (fig. S4, D to F). The basal activity of Akt, which
   is upstream of mTORC1 signals, was enhanced in RIPK1-KO T[H] cells,
   whereas Phosphoinositide-dependent kinase 1(PDK1), upstream of Akt, was
   not activated in these cells ([123]Fig. 4B). Next, we assessed whether
   the reduced activation of mTORC1 following rapamycin treatment affected
   the TCR-induced cytokine production and the expression of
   senescence-related genes in RIPK1-KO T[H] cells. Increased expression
   of senescence-related genes and augmented production of IFN-g, IL-4,
   IL-5, and IL-6 in RIPK1-KO Th cells was remarkably inhibited by
   rapamycin-treatment ([124]Fig. 4, C and D). Furthermore, the increased
   expression of senescence-related genes and the increase of mTORC1
   activation in RIPK1-KO T[H] cells were strongly abolished by blocking
   Akt activation with LY294002, an inhibitor of PI3K (fig. S4, G and H).
   These data suggest that RIPK1 controls mTOR signaling through
   regulating Akt activation.

Fig. 4. Elevated activation of Akt, mTORC1, and ERK signaling in RIPK1-KO CD4
T cells.

   [125]Fig. 4.
   [126]Open in a new tab

   (A) Pathway enrichment analysis of gene expression with up-regulation
   in RIPK1-KO CD4 T cells as compared to WT CD4 T cells. Data were
   analyzed with DAVID Bioinformatics Resources 6.8. (B) CD4^+ CD25^−
   CD44^lo CD62L^hi (naïve CD4) T cells from control or RIPK1-tKO mice
   were stimulated with anti-CD3/28 monoclonal antibody (mAb) in the
   presence or absence of rapamycin for 3 days, and cells were further
   cultured in the presence of IL-2 for 3 days. PI^− T[H] cells were
   sorted and subjected to Western blot analysis of the indicated proteins
   (B) and qPCR analysis of the indicated genes (C) or stimulated with
   anti-CD3/28 mAb for 24 hours, and cytokine production was analyzed by
   ELISA (D). (E and F) Naïve CD4 T cells from control or RIPK1-tKO or
   RIPK1^D138N/D138N mice (E) or young or aged mice (F) were stimulated
   with anti-CD3/28 mAb for 3 days, and then the cells were further
   cultured in the presence of IL-2 for 3 days. PI^− T[H] cells were
   sorted and subjected to Western blot analysis of the indicated
   proteins. (G) Oxygen consumption rate (OCR) in T[H] cells from control
   or RIPK1-tKO mice. (H) Basal respiration (left), maximal respiration
   (middle), and adenosine triphosphate (ATP) production (right) are
   shown. (I) OCR in T[H] cells described in (B). (J) Basal respiration
   (left), maximal respiration (middle), and ATP production (right) are
   shown. Error bars represent SEM. Data are representative of at least
   three independent experiments (B to J). *P < 0.05, **P < 0.01,
   ***P < 0.005, and ****P < 0.001 using Student’s t test.

   Unlike RIPK1-KO T[H] cells, RIPK1 kinase inactive (RIPK1^D138N/D138N)
   T[H] cells showed normal basal activation of mTORC1 and cytokine
   production ([127]Fig. 4E and fig. S4I), suggesting that the kinase
   activity of RIPK1 is dispensable for regulation of T cell senescence.
   Similar to RIPK1-KO T[H] cells, aged WT T[H] cells showed augmented
   basal activation of Akt, mTORC1 signals, and ERK; an increased cell
   size; and enhanced cytokine production, which were all reduced by
   rapamycin treatment ([128]Fig. 4F and fig. S4, J to L).

   Senescent cells have been shown to display metabolic changes such as
   increased glycolysis and mitochondrial oxidative phosphorylation
   (OXPHOS) ([129]45). Glycolysis-induced metabolites modulate
   inflammatory responses by signal modulation, posttranscriptional and
   posttranslational modifications, and epigenetic remodeling ([130]46).
   Because it has been shown that mTORC1 activation is required to
   regulate TCR-induced glycolysis and OXPHOS ([131]7), we next analyzed
   the oxygen consumption rate (OCR), which assesses OXPHOS and
   extracellular acidification (ECAR), a measure of glycolysis. We found
   that RIPK1-KO T[H] cells exhibited elevated levels of OXPHOS and
   glycolysis ([132]Fig. 4, G and H, and fig. S4M) and that these
   increases were strongly inhibited by rapamycin treatment ([133]Fig. 4,
   I and J, and fig. S4N). These data indicate that RIPK1 deficiency in T
   cells causes hyperactivation of mTOR, leading to metabolic alterations
   and, subsequently, inflammatory responses.

Rapamycin treatment restores the function of RIPK1-KO T cells in vivo

   Next, we investigated whether rapamycin treatment could also affect the
   senescent phenotype of RIPK1-KO T cells in vivo. Administration of
   rapamycin to RIPK1-tKO mice daily for 4 days decreased the percentage
   of CD44^hi CD62L^lo effector/memory phenotype CD4 T cells and increased
   the percentage of CD44^lo CD62L^hi naïve CD4 T cells ([134]Fig. 5, A
   and B). The treatment also reduced the expression of senescence-related
   genes and the production of IFN-γ, IL-4, and IL-17A upon TCR
   stimulation by CD4 T cells and T[H] cells from RIPK1-tKO mice compared
   to untreated RIPK1-tKO mice ([135]Fig. 5, C and D, and fig. S5, A and
   B). Consistently, the serum levels of IFN-γ, IL-2, IL-5, TNF-α, and IgG
   were also decreased in these mice ([136]Fig. 5E). These data indicate
   that enhanced activation of mTOR in RIPK1-KO T cells induces the
   senescence of T cells, which results in the elevated cytokine
   production in vivo and that this phenotype can be reversed by rapamycin
   treatment.

Fig. 5. Rapamycin inhibits CD4 T cell senescence in RIPK1-tKO mice.

   [137]Fig. 5.
   [138]Open in a new tab

   (A) The percentage of CD44^hi CD62L^lo cells (B) and CD44^lo CD62L^hi
   cells among CD4 T cells from spleen of control and RIPK1-tKO mice left
   untreated or treated daily with rapamycin for a total of 4 days (n = 3
   per group). (C) qPCR analysis of the indicated genes in CD4 T cells
   from spleen of control and RIPK1-tKO mice left untreated or treated
   daily with rapamycin for a total of 4 days. (D) CD4 T cells from spleen
   of control and RIPK1-tKO mice left untreated or treated daily with
   rapamycin for a total of 4 days were stimulated with anti-CD3/28 mAb
   for 24 hours, and cytokine production was analyzed by ELISA. (E) Serum
   cytokine levels were determined by ECL in control and RIPK1-tKO mice
   left untreated or treated every other day with rapamycin for a total of
   30 days (n = 5 per group). Error bars represent SEM. Data are
   representative of at least three independent experiments (A to E).
   *P < 0.05, **P < 0.01, ***P < 0.005, and ****P < 0.001 using Student’s
   t test. NS, not significant.

RIPK1 inhibits RIPK3- and caspase-8–mediated activation of Akt, mTORC1, and
ERK signaling

   A previous study demonstrating that RIPK1-KO T cells have elevated
   caspase activity ([139]24) prompted us to determine whether increased
   basal activation of mTORC1 leads to augmentation of caspase activity.
   The levels of cleaved caspase-3, caspase-7, and poly(adenosine
   diphosphate–ribose) polymerase (PARP), which is the cleavage target of
   caspase-3, were increased in RIPK1-KO T[H] cells ([140]Fig. 6A).
   Treatment with rapamycin or LY294002 strongly reduced the levels of
   cleaved caspase-3, caspase-7, and PARP in RIPK1-KO T[H] cells
   ([141]Fig. 6A and fig. S6A). Similar results were obtained with aged WT
   T[H] cells ([142]Fig. 6B and fig. S6B), indicating that enhanced basal
   Akt-mTORC1 signaling induces the activation of caspase-3/7 in RIPK1-KO
   and aged CD4 T cells.

Fig. 6. RIPK1 inhibits the capsase-8 pathway.

   [143]Fig. 6.
   [144]Open in a new tab

   (A to F) Naïve CD4 T cells from control or RIPK1-tKO or young or aged
   WT mice were stimulated with anti-CD3/28 mAb for 3 days in the presence
   or absence of rapamycin or Z-IETD, and then the cells were further
   cultured in the presence of IL-2 for 3 days. PI^− T[H] cells were
   sorted and subjected to Western blot analysis of the indicated proteins
   (A to C and F) or to qPCR analysis of the indicated genes (D). (E) CD4
   T cells expressing small interfering RNA (siRNA) targeting caspase-8
   (si-Casp-8) or control siRNA (si-Ctrl) were subjected to Western blot
   analysis. Error bars represent SEM. Data are representative of at least
   three independent experiments (A to F). ****P < 0.001 using Student’s t
   test.

   Because RIPK1 blocks caspase-8–dependent apoptosis and RIPK3-dependent
   necroptosis during development ([145]19, [146]20), it is likely that
   caspase-8 and RIPK3 are involved in the increased activation of Akt,
   mTORC1, and ERK signaling in RIPK1-KO T cells. To confirm this, we
   first used Z-IETD, an inhibitor specific for caspase-8. Treatment with
   Z-IETD results in partial reduction of the increased phosphorylation of
   Akt, mTORC1, and ERK and activation of caspase-3/7 in RIPK1-KO T[H]
   cells ([147]Fig. 6C). Z-IETD treatment also partially inhibited the
   expression of senescence-related genes in RIPK1-KO T[H] cells except
   for Il1r2 ([148]Fig. 6D). Similar results were obtained with caspase-8
   knockdown RIPK1-KO T[H] cells ([149]Fig. 6E and fig. S6C), implying
   that RIPK1 inhibits the caspase-8–mediated activation of Akt, mTORC1,
   and ERK. We also analyzed whether the treatment with Z-IETD affects the
   cell death level and the cell division of RIPK1-KO naïve CD4 T cells
   upon anti-CD3/28 stimulation. We found that the increased cell death
   but not the impaired cell division in RIPK1-KO CD4 T cells was partly
   restored by Z-IETD treatment (fig. S6, E and E). These data indicate
   that caspase-8 is critical for the increased cell death and the
   senescent phenotype caused by RIPK1 deficiency in T cells. Furthermore,
   Z-IETD treatment partially decreased the enhanced activation of mTORC1,
   Akt, ERK, and caspase-3/7 and the increased expression of
   senescence-related genes in aged T[H] cells ([150]Fig. 6F and fig.
   S6F). These data indicate that caspase-8 is involved in the induction
   of senescence in RIPK1-KO and normal CD4 T cells.

   We then investigated the involvement of RIPK3 in this process by
   crossing RIPK1-tKO mice with RIPK3-KO mice (RIPK1/3-dKO mice). The
   decreased ratio and number of CD4 T cells in RIPK1-tKO mice were
   partially rescued by RIPK3 deficiency ([151]Fig. 7, A and B). Further,
   the impaired proliferative response of RIPK1-KO naïve CD4 T cells upon
   anti-CD3/28 stimulation was partially restored by RIPK3 deficiency and
   almost fully restored in RIPK1/3-dKO T cells treated with Z-IETD or
   Z-VAD, a pan-caspase inhibitor ([152]Fig. 7C). In addition, the
   enhanced activation of mTORC1, Akt, ERK, and caspase-3/7 was partially
   recovered in RIPK1/3-dKO T[H] cells and almost fully recovered when
   these cells were treated with Z-IETD ([153]Fig. 7D and fig. S7A). The
   enhanced production of IL-4 and IFN-γ and the decreased production of
   IL-2 were partly normalized in RIPK1/3-dKO T[H] cells and more fully
   restored after treatment with Z-IETD ([154]Fig. 7E and fig. S7B).
   Consistently, the increased expression of senescence-related genes in
   RIPK1-KO T[H] cells was partly reduced by RIPK3 deficiency and almost
   fully rescued by RIPK3 deficiency and Z-IETD treatment ([155]Fig. 7F
   and fig. S7C). The expression of RIPK3, but not RIPK1 and caspase-8,
   was up-regulated in CD44^hi CD62L^lo effector/memory phenotype CD4 T
   cells compared to CD44^lo CD62L^hi naïve CD4 T cells, suggesting the
   possible involvement of RIPK3 in the induction of T cell senescence.
   Together, these data suggest that both RIPK3 and caspase-8 are critical
   for the induction of T cell senescence by activating Akt, mTORC1, and
   ERK, which is blocked by RIPK1.

Fig. 7. RIPK1 inhibits the RIPK3- and caspase-8–mediated activation of Akt,
mTORC1, and ERK pathways.

   [156]Fig. 7.
   [157]Open in a new tab

   (A) The percentage and number (B) of CD4 T cells from spleen of
   control, RIPK1-tKO, and RIPK1/3-dKO mice. (C) Naïve CD4^+ T cells from
   control or RIPK1-tKO or RIPK1/3-dKO mice were stimulated with
   anti-CD3/CD28 in the presence or absence of Z-IETD or Z-VAD. Cell
   growth was assessed after 48 hours of stimulation by a WST-8
   proliferation assay. (D to F) Naïve CD4 T cells from control or
   RIPK1-tKO or RIPK1/3-dKO mice were stimulated with anti-CD3/28 mAb for
   3 days in the presence or absence of Z-IETD, and then the cells were
   further cultured in the presence of IL-2 for 3 days. PI^− T[H] cells
   were sorted and subjected to Western blot analysis of the indicated
   proteins (D) or qPCR analysis (F) or stimulated with anti-CD3/28 mAb
   for 24 hours, and cytokine production was analyzed by ELISA (E). (G)
   CD44^lo CD62L^hi naïve CD4 T cells and CD44^hi CD62L^lo effector/memory
   phenotype CD4 T cells were sorted and subjected to Western blot
   analysis of the indicated proteins. Error bars represent SEM. Data are
   representative of at least three independent experiments (A to H).
   *P < 0.05, **P < 0.01, ***P < 0.005, and ****P < 0.001 using Student’s
   t test.

Environment affects the senescent phenotype of CD4 T cells

   Because it has been shown that the SASP factors reinforce and propagate
   senescence ([158]5), we next investigated whether the tissue
   environment of RIPK1-tKO mice converts the WT CD4 T cells into the
   senescent phenotype CD4 T cells by adoptive T cell transfer. We found
   that the percentage of CD44^hi CD62L^lo effector/memory phenotype CD4 T
   cells markedly increased ([159]Fig. 8A), whereas that of CD44^lo
   CD62L^hi naïve CD4 T cells inversely decreased in WT T cells when
   transferred into RIPK1-tKO mice as compared to WT T cells before
   transfer ([160]Fig. 8B). Consistently, the expression of
   senescence-related genes was up-regulated in WT T cells when
   transferred into RIPK1-tKO mice as compared to WT T cells before
   transfer ([161]Fig. 8C). The cell numbers of WT CD4 T cells transferred
   into control or RIPK1-tKO mice were comparable (fig. S8A), suggesting
   that the phenotypic conversion of WT T cells into the senescent
   phenotype CD4 T cells in RIPK1-tKO mice is independent of homeostatic
   proliferation. Conversely, we next determined whether the normal tissue
   environment restores the RIPK1-KO CD4 T cells to the normal CD4 T
   cells. Consequently, the percentage of CD44^hi CD62L^lo effector/memory
   phenotype CD4 T cells markedly decreased ([162]Fig. 8D), whereas that
   of CD44^lo CD62L^hi naïve CD4 T cells inversely increased in RIPK1-KO
   CD4 T cells transferred to WT mice ([163]Fig. 8E). Consistently, the
   expression of senescence-related genes was down-regulated in RIPK1-KO T
   cells transferred into WT mice as compared to RIPK1-KO T cells before
   transfer ([164]Fig. 8F). The cell number of RIPK1-KO CD4 T cells when
   transferred into a WT host was less than that of WT T cells transferred
   into a WT host (fig. S8B). These data indicate the possibility that T
   cell senescence caused by RIPK1 deficiency may not be a T
   cell–intrinsic effect, whereas the cell death caused by RIPK1
   deficiency is a T cell–intrinsic effect.

Fig. 8. Environment affects the senescent phenotype of CD4 T cells.

   [165]Fig. 8.
   [166]Open in a new tab

   (A to C) WT (CD45.1^+) CD4^+ CD25^− T cells were transferred into WT or
   RIPK1-tKO mice, and the percentage of CD44^hi CD62L^lo cells (A) and
   CD44^lo CD62L^hi cells (B) among CD45.1^+ CD4 T cells of WT and
   RIPK1-tKO mice (n = 2 per group) was determined by flow cytometry.
   CD45.1^− CD4 T cells from WT and RIPK1-tKO recipients and CD45.1^+ CD4
   T cells from WT (WT → WT) and RIPK1-tKO (WT → RIPK1-tKO) mice were
   collected and subjected to qPCR analysis (C). (D to F) CD4^+ CD25^− T
   cells from WT and RIPK1-tKO mice were transferred into WT (CD45.1^+)
   mice, and the percentage of CD44^lo CD62L^hi cells (D) and CD44^hi
   CD62L^lo cells (E) among CD4 T cells from WT and RIPK1-tKO mice (n = 2
   per group) was determined by flow cytometry. WT and RIPK1-KO CD4 T
   cells before transfer and WT (WT → WT) and RIPK1-KO (RIPK1-tKO → WT)
   CD4 T cells from CD45.1^+ mice were collected and subjected to qPCR
   analysis (F). (G and H) CD4^+ CD25^− T cells from young and aged mice
   were transferred into young WT (CD45.1^+) mice, and the percentage of
   CD44^lo CD62L^hi cells (G) and CD44^hi CD62L^lo cells (H) among CD4 T
   cells from young and aged mice (n = 2 per group) was determined by flow
   cytometry. (I and J) Naïve CD4 T cells from WT (CD45.1^+) or RIPK1-tKO
   mice were cocultured at a ratio of 1 to 1, or cultured alone, upon
   anti-CD3/28 mAb stimulation for 6 days. PI^− T[H] cells were sorted and
   subjected to qPCR analysis (I) and Western blot analysis (J). Error
   bars represent SEM. Data are representative of at least two independent
   experiments (A to J). *P < 0.05, **P < 0.01, ***P < 0.005, and
   ****P < 0.001 using Student’s t test.

   To understand this mechanism more physiologically, we examined whether
   the normal tissue environment restores the aged CD4 T cells to the
   normal phenotype. The percentage of CD44^hi CD62L^lo
   effector/memory-type CD4 T cells markedly decreased ([167]Fig. 8G),
   whereas that of CD44^lo CD62L^hi naïve CD4 T cells inversely increased
   in aged CD4 T cells transferred into young WT mice ([168]Fig. 8H).
   Consistently, the expression of senescence-related genes was decreased
   in CD4 T cells transferred from aged mice into young WT mice as
   compared to aged CD4 T cells before transfer (fig. S8C). The cell
   numbers of young and aged CD4 T cells transferred into young WT mice
   were comparable (fig. S8D). These data suggest that the senescent
   phenotype of aged CD4 T cells is restored in the normal tissue
   environment.

   We further determined whether the coculture environment of WT and
   RIPK1-KO T[H] cells affects the phenotype of each T[H] cells in vitro.
   The expression levels of senescence-related genes were strongly
   inhibited in RIPK1-KO T[H] cells when cocultured with WT T[H] cells as
   compared to RIPK1-KO T[H] cells alone. On the other hand, those of
   senescence-related genes were not altered in WT T[H] cells when
   cocultured with RIPK-KO T[H] cells compared to WT T[H] cells alone
   ([169]Fig. 8I). Consistent with these results, the elevated activation
   of Akt, mTORC1, ERK, and caspase-3/7 was inhibited in RIPK1-KO T[H]
   cells when cocultured with WT T[H] cells as compared to RIPK1-KO T[H]
   cells alone ([170]Fig. 8J). Furthermore, the aberrant production of
   IL-2, IL-4, and IFN-γ was restored in RIPK1-KO T[H] cells when
   cocultured with WT T[H] cells (fig. S8E). It is noted that the cell
   number of RIPK1-KO T[H] cells cocultured with WT T[H] cells was much
   less than of WT T[H] cells cocultured with RIPK1-KO T[H] cells, and it
   was comparable to RIPK1-KO T[H] cells cultured alone (fig. S8F).
   Therefore, the relative cell numbers of CD4 T cells originated from WT
   or RIPK1-KO CD4 T cells may determine the tissue and culture
   environment, which influence inducing senescence. These data indicate
   that the environment cocultured with WT T[H] cells restores the
   senescent phenotype and the enhanced activation of Akt, mTORC1, ERK,
   and caspase-3/7 without altering the increased cell death in RIPK1-KO
   T[H] cells. Collectively, these data indicate that the environmental
   signals are critical to determine the senescent phenotype of CD4 T
   cells by modulating mTOR signaling.

DISCUSSION

   Biallelic mutations in RIPK1 cause combined immunodeficiency and
   inflammatory diseases in humans. Our studies show that RIPK1 deficiency
   in T cells is sufficient for the development of inflammatory diseases
   in mice, ultimately leading to premature death. Similar to RIPK1-tKO
   mice, human RIPK1 deficiency causes lymphopenia and splenomegaly
   ([171]22, [172]23), suggesting that RIPK1 in T cells may play a
   critical role in controlling human inflammatory diseases.

   We further demonstrated that RIPK1 deficiency in T cells results in T
   cell senescence, which is sufficient to cause age-associated features
   including chronic tissue inflammation, neurodegeneration, sarcopenia,
   anemia, and hypoglycemia, indicating the critical role of T cells in
   these age-related pathologies.

   Mechanistically, RIPK1-KO CD4 T cells exhibit elevated basal activation
   of Akt, mTORC1, and ERK, leading to the enhanced expression of
   senescence-related genes and production of proinflammatory cytokines.
   Such basal activation of these genes is restored by a combined
   deficiency of RIPK3 and caspase-8 inhibition, indicating that
   activation of RIP3K/caspase-8 leads to senescence. Aged WT T cells
   exhibit enhanced basal activation of Akt, mTORC1, and ERK signaling
   similar to RIPK-KO T cells, and the treatment with rapamycin or a
   caspase-8 inhibitor corrects the dysfunction of these aged T cells. Our
   findings indicate that the RIPK3 and caspase-8 pathways are involved in
   basal signal activation in aged T cells under physiological conditions,
   suggesting that RIPK1 and RIPK3/caspase-8 may be potentially important
   therapeutic targets for delaying aging and age-associated diseases.

   We showed that host environment affects the senescent phenotype of
   transferred CD4 T cells. The senescent phenotype of aged and RIPK1-KO
   CD4 T cell is restored in the normal tissue environment. Coculture
   experiments consistently revealed that relatively large numbers of
   young CD4 T cells induce the restoration of the enhanced expression of
   senescence-related genes and the increased activation of mTOR in
   RIPK1-KO CD4 T cells. These data suggest the possibility that the
   enhanced activation of mTOR in aged and RIPK1-KO CD4 T cells may not
   merely be T cell–intrinsic effects but rather modulated by the
   environmental signals. Therefore, environmental signals are likely to
   be critical for the modulation of T cell senescence. mTOR activations
   are tuned by such multiple environmental signals as TCR, costimulatory
   molecules, cytokines, growth factors, nutrients, oxygen, and
   metabolites ([173]6). Therefore, there might be two possible
   mechanisms: T cell–intrinsic and environmental signals for the
   induction of senescent CD4 T cells by RIPK1 deficiency. One is that
   enhanced activation of mTOR, which was caused by RIPK1 deficiency and
   mediated by RIPK3/caspase-8, initiates the conversion into the
   senescent T cells. This conversion is induced in a T cell–intrinsic
   manner and amplified by the environmental signals in RIPK1-tKO mice.
   Another possibility is that the environment where massive T cell death
   is induced by RIPK1 deficiency promotes the activation of mTOR, leading
   to T cell senescence. Further investigation is required to elucidate
   the mechanism by which the normal environmental signals restore the
   senescent phenotype of aged and RIPK1-KO CD4 T cells.

   Recent studies reported that a metabolic disorder in mitochondrial
   regulation in T cells induced T cell senescence and age-associated
   inflammation ([174]47–[175]49). Our studies revealed that RIPK1 is a
   key molecule for T cell metabolism by modulating mTOR signaling. A
   recent study reported that RIPK1 is involved in regulating TSC2 (an
   upstream negative regulator of mTORC1) phosphorylation by AMPK, which
   resulted in elevated mTORC1 activity in mouse embryonic fibroblasts
   (MEFs) and a human colon cancer cell line ([176]50). Furthermore,
   another report demonstrated that RIPK1 deficiency increases aspartate
   levels, which enhances the tricarboxylic acid (TCA) flux and adenosine
   triphosphate (ATP) production, leading to defective autophagy induction
   by inhibiting the AMP-activated protein kinase (AMPK)/unc-51 like
   autophagy activating kinase 1 (ULK1) pathway in MEFs ([177]51). In this
   case, RIPK1 regulates aspartate metabolism by inhibiting specificity
   protein 1 (SP1) nuclear expression, probably by altered phosphorylation
   ([178]51). However, our RNA-seq data showing that expression of genes
   related to aspartate catabolism was not down-regulated in RIPK1-KO CD4
   T cells indicate that regulation of mTOR signaling by RIPK1 may be cell
   context dependent.

   We have clarified the mechanism of RIPK1 to regulate the activation of
   Akt/mTORC1/ERK in RIPK1-tKO mice and its similarity to aged T cells.
   However, how basal activation of Akt/mTORC1/ERK in normal aged T cells
   is increased as in RIPK1-KO T cells is still unclear. Because a
   caspase-8 inhibitor partly inhibits the activation of Akt, mTORC1, ERK,
   and caspase-3/7, as well as the expression of senescence-related genes
   in aged CD4 T cells, caspase-8 is involved in the induction of T cell
   senescence in a physiological setting. In addition, considering that
   the expression level of RIPK3 but not RIPK1 is increased in aged
   effector/memory phenotype CD4 T cells, it is possible that activation
   of RIPK3 and caspase-8 may be enhanced, although the inhibitory
   function of RIPK1 remains unchanged with aging. A recent study showed
   that the down-regulation of deoxyribonuclease 2 (DNase2) and Three
   prime repair exonuclease 1 (TREX1) expression causes the cytoplasmic
   accumulation of DNA in senescent cells, leading to the SASP by the
   aberrant activation of cyclic GMP-AMP synthase (cGAS)-Stimulator of
   interferon genes (STING) pathway ([179]52). In the similar line,
   another study has shown that DNA is accumulated in the cytoplasm of
   aged CD4 T cells ([180]53). However, in this case, the cytoplasmic DNA
   is sensed by the KU complex but not STING and activates leucine zipper
   and sterile α motif kinase (ZAK), which then activates Akt and mTORC1
   pathways ([181]53). Therefore, it is possible that the activation of
   RIPK3 and caspase-8 may be enhanced upon DNA sensing by the KU complex
   with aging in T cells. Accordingly, RIPK3 and caspase-8 might be
   involved in the regulation of ZAK activation in aged T cells.

   Because RIPK1 interacts with caspase-8 through the FAS-associated death
   domain protein (FADD) adaptor protein in TNF-α signaling, it is likely
   that RIPK1 inhibits caspase-8–mediated activation of Akt, mTORC1, and
   ERK through FADD in T cells. Although previous studies reported that
   caspase-8 is essential for T cell activation in addition to its
   critical roles in apoptosis in both humans and mice ([182]54, [183]55),
   the molecular mechanisms underlying caspase-8–mediated T cell
   activation remain unsettled despite extensive studies. On the basis of
   our analyses, caspase-8 may be involved in the basal activation of Akt,
   mTORC1, and ERK pathways in T cells and that RIPK1 regulates caspase-8
   function.

   Increasing evidence suggests that mTORC1 is a critical regulator for
   aging. Long-term rapamycin treatment or genetic inhibition of mTORC1
   signaling could extend the life span of mice ([184]56, [185]57).
   Furthermore, mTOR inhibition improved the response to the influenza
   vaccine and reduced the percentage of CD4 and CD8 T cells expressing
   Programmed cell death 1 (PD-1), which increases with age, in elderly
   humans ([186]58). It will be an intriguing study to determine whether
   inhibition of mTORC1 signaling specifically in T cells can extend life
   span.

   We found that increased activation of mTORC1 results in the activation
   of caspase-3/7 in RIPK1-KO and aged T cells. Consistent with our
   results, TSC1-KO T cells exhibit higher caspase activity ([187]8).
   Mechanistically, loss of TSC1 or TSC2 causes endoplasmic reticulum
   stress and activates the unfolded protein response, which leads to
   apoptosis in MEFs ([188]59). Although the physiological role of
   noncanonical caspase-3/7 activation mediated by hyperactivation of
   mTORC1 in T cells remains unclear, caspase-3/7 activation could be a
   useful senescence marker for T cells.

   Together, our results demonstrate that RIPK1 deficiency triggers the
   activation of RIPK3/caspase-8 and then mTORC1, which is critically
   involved in inducing T cell senescence, and that senescence-inducing
   signals are modulated by the environmental signals. The observation
   that RIPK1 deficiency in T cells causes age-related diseases and
   premature death in mice provides impetus for further clinical studies
   and the development of new therapeutic strategies targeting RIPK1 and
   RIPK3/caspase-8 for age-related diseases.

METHODS

Animals

   C57BL/6J mice were purchased from Clea Japan Inc. Ripk1^fl/fl and
   Ripk1^D138N/D138N mice were provided by M.P. (University of Cologne).
   Ripk3^−/− mice were provided by Genentech Inc. Foxp3YFP-Cre and CD45.1
   mice were purchased from The Jackson Laboratory. Male and female mice
   of 2 to 26 months old were used. Two to 4, 6 to 10, and 20 to 26 months
   old were used as young, middle-aged, and aged mice, respectively. All
   mice were maintained under specific pathogen–free conditions at RIKEN,
   and all experiments were conducted under protocols approved by the
   RIKEN Yokohama Institute.

Reagents and antibodies

   Rapamycin, Z-IETD-FMK, Z-VAD-FMK, and LY294002 were obtained from
   Merck. Antibodies (Abs) specific for anti-phospho-S6K1 (#9205; 1:500
   dilution), anti–phospho-S6 (#2211; 1:10,000 dilution),
   anti–phopho-4E-BP1 (#9459; 1:1000 dilution), anti–phospho-Akt (S473)
   (#9271; 1:1000 dilution), anti–phospho-Akt (T308) (#4056; 1:500
   dilution), anti–phospho-ERK1/2 (#9101; 1:1000 dilution), anti-cleaved
   PARP (#9548; 1:1000 dilution), anti–caspase-3 (#9662; 1:1000 dilution),
   anti–caspase-7 (#9492; 1:1000 dilution), anti–phospho-PDK1 (Ser^241)
   (#3061; 1:1000 dilution), anti–phospho-p38 (#9211; 1:1000 dilution),
   anti–β-actin (#4967; 1:10,000 dilution), anti–phospho-S6-PE (#5316;
   1:50 dilution), and anti–phospho-4E-BP1-PE (#7547; 1:50 dilution) were
   obtained from Cell Signaling Technology; anti–caspase-8 (ALX-804-447;
   1:1000 dilution) was obtained from Enzo. Flow cytometric analysis was
   performed on LSRFortessa X-20, and data were analyzed with FlowJo.

Cell preparation

   CD4^+ naïve T cells were purified from spleen and lymph nodes as
   CD4^+/CD25^−/NK1.1^−/CD44^low/CD62L^high cells by sorting using a
   FACSAria (BD Biosciences). T[H] cells were prepared by stimulation of
   naïve CD4^+ T cells with plate-bound anti-CD3ε (2C11; 10 μg/ml) and
   anti-CD28 (PV-1; 10 μg/ml) (anti-CD3/CD28) in the presence or absence
   of rapamycin (100 nM) or LY294002 (10μM) or Z-IETD (50 μM) for 3 days,
   and then the cells were washed and further cultured in the presence of
   IL-2 (10 ng/ml) for 3 days in RPMI 1640 supplemented with 10% fetal
   calf serum, and then Propidium iodide (PI)^− live cells were sorted
   using a FACSAria.

Functional analyses

   Naïve CD4 T cells and T[H] cells were stimulated with immobilized
   anti-CD3ε (10 μg/ml) and anti-CD28 (5 μg/ml). Culture supernatants were
   analyzed by enzyme-linked immunosorbent assay (ELISA) for production of
   IL-2, IL-4, and IFN-γ (BD Biosciences) or electrochemiluminescence
   (ECL) using the V-PLEX Proinflammatory Panel 1 Mouse Kit (K15048D, Meso
   Scale Diagnostics). Sera were analyzed by ECL using the V-PLEX
   Proinflammatory Panel 1 Mouse Kit or U-PLEX mouse IL-17A assay
   (K152UTK; Meso Scale Diagnostics) or ELISA using mouse IgM, IgG, and
   IgE ELISA kits (Bethyl Laboratories); mouse anti-dsDNA ELISA kit
   (AKRDD-061, FUJIFILM); and mouse insulin ELISA kit (Mercodia Uppsala,
   Sweden). Cell growth was assessed using a cell counting kit-8
   (Dojindo).

Real-time quantitative polymerase chain reaction

   After removal of genomic DNA by treatment with DNase (Wako Nippon
   Gene), randomly primed cDNA strands were generated with reverse
   transcriptase III (Invitrogen). RNA expression was quantified by
   real-time polymerase chain reaction (PCR) with gene-specific primers,
   and values were normalized to the expression of Rps18 mRNA.
   Quantitative PCR (qPCR) was performed with the Fast Syber Green Master
   Mix (Applied Biosystems). Data were collected and calculated by using
   the StepOnePlus real-time PCR system (Applied Biosystems).

Western blot analysis

   Cells were lysed in 1% NP-40 lysis buffer [1% NP-40, 50 mM tris, 150 mM
   NaCl, 5 mM EDTA, aprotinin (10 μg/ml), chymostatin (12.5 μg/ml),
   leupeptin (50 μg/ml), pepstatin A (25 μg/ml), 1 mM phenylmethylsulfonyl
   fluoride, and 2 mM Na[3]VO[4]]. The lysates or immunoprecipitates were
   subjected to SDS–polyacrylamide gel electrophoresis and then
   transferred to membranes. Western blots were conducted by reacting the
   membrane with specific Ab and developed with an enhanced
   chemiluminescence assay according to the manufacturer’s recommendations
   (Pierce).

RNA interference

   Double-stranded oligonucleotides corresponding to the target sequences
   were cloned into the pSuper.Retro RNAi plasmid (OligoEngine Inc.). The
   small interfering RNA (siRNA) targeting sequence of caspase-8 is
   5′-GGAAGATCGAGGATTATGAAA-3′. Constructs were transiently transduced
   into Phoenix packaging cells (provided by G. Nolan, Stanford
   University) using Lipofectamine with PLUS reagent (Invitrogen). Naïve
   CD4 T cells were stimulated with plate-bound anti-CD3/CD28 monoclonal
   Abs (mAbs), and cells were transduced by centrifugation at 1640g for 90
   min in retroviral supernatants and polybrene (8 μg/ml) (Sigma-Aldrich)
   on day 1 after stimulation. After 72 hours of stimulation, green
   fluorescent protein–positive cells were sorted with FACSAria.

RNA-seq analysis

   Total RNA was isolated from T cells using the Direct-zol RNA kit (Zymo
   Research) according to the manufacturer’s instructions. The DNA library
   for RNA-seq analysis was constructed with a NEBNext Ultra RNA Library
   Prep Kit for Illumina (New England Biolabs Inc., Ipswich, MA) according
   to the manufacturer’s instruction. The size range of the resulting DNA
   library was estimated on a 2100 Bioanalyzer (Agilent Technologies). The
   DNA library was subjected to the HiSeq 1500 sequencing system
   (Illumina) in a single-end read mode to obtain the sequencing data. The
   sequence reads were mapped to the Mus musculus reference genome
   [National Center for Biotechnology Information (NCBI) version 37] using
   TopHat2 version 2.0.8 and botwie2 version 2.1.0 with default
   parameters, and gene annotation was provided by NCBI. According to the
   mapped data, Cufflinks (version 2.1.1) was used to calculate the FPKM
   (fragments per kilobase per million mapped reads) values. Pathway
   enrichment analysis was done using DAVID Bioinformatics Resources 6.8
   ([189]60). Heatmaps were produced from normalized expression data
   referring to DAVID Bioinformatics Resources 6.8.

Clinical scoring

   Limb grip strength was measured using a grip strength test apparatus
   (Bioseb). For the hindlimb reflex score, mice were suspended from their
   tails, and the hindlimb reflex score was evaluated using an established
   definition: 0, normal; 1, failure to stretch their hindlimbs; 2,
   hindlimb clasping; and 3, hindlimb paralysis. The severity of eye
   opacity was scored using the following definition: 0, normal; 1, one
   teary eye; 2, both teary eyes; 3, one cloudy eye; 4, one teary eye and
   one cloudy eye; and 5, both cloudy eyes.

ITTs and GTTs

   After fasting for 4 hours, 15 μl of Humulin R (100 U/ml) (Eli Lilly and
   Company, Indianapolis, IN) was diluted in 7.2 ml of saline and
   administrated intraperitoneally at 3.6 μl/g of body weight (0.75 mU/g
   of body weight). Blood glucose levels were measured immediately before
   administration (0 min) and at 20, 40, and 60 min using an ACCU-CHEK ST
   Meter (Roche) from the tail vein.

   A glucose solution (20%) was prepared by diluting 50% glucose (Otsuka
   Pharmaceutical Factory, Naruto, Japan) with distilled water. After
   fasting for 16 hours, the 20% glucose solution was intraperitoneally
   administrated at 10 μl/g of body weight (2 mg/g of body weight), and
   blood glucose levels were determined using an ACCU-CHEK ST Meter
   (Roche) immediately before administration (0 min) and at 15, 30, 60,
   and 120 min from the tail vein.

Metabolic assays

   The OCR and ECAR in T[H] cells were measured in an XFp extracellular
   flux analyzer (Seahorse Bioscience) using the XFp Cell Mito Stress Test
   Kit. The parameters used in the assays were the following: 2.5 × 10^4
   cells per well, 1 μM oligomycin, 2 μM carbonyl cyanide
   p-trifluoromethoxyphenylhydrazone, 0.5 μM rotenone/antimycin A, 10 mM
   glucose, and 50 mM 2-deoxyglucose, as indicated.

Liquid chromatography–tandem mass spectrometry analysis

   The liquid chromatography–tandem mass spectrometry analysis of
   metabolites in serum from 10-month-old control or RIPK1-tKO mice was
   performed using a Nexera X2 system (Shimadzu Corp.) equipped with two
   LC-30 AD pumps, a DGU-20A[5R] degasser, an SIL-30 AC autosampler, a
   CTO-20 AC column oven, and a CBM-20A control module, coupled with an
   LCMS-8060 triple quadrupole mass spectrometer (Shimadzu Corp.). A
   pentafluorophenylpropyl column (Discovery HS F5; 150 mm by 2.1 mm, 3
   μm; Sigma-Aldrich) was used for the separation of metabolites. The
   mobile phase was composed of A, 0.1% (v/v) formic acid in water, and B,
   0.1% (v/v) formic acid in acetonitrile. The flow rate, column
   temperature, and injection volume were set as 0.25 ml/min, 40°C, and 3
   μl, respectively. The gradient program for mobile phase B was as
   follows: 0 min, 0%; 2 min, 0%; 5 min, 25%; 11 min, 35%; 15 min, 95%; 20
   min, 95%; 20.1 min, 0%; and 25 min, 0%. The mass spectrometer was
   equipped with an electrospray ionization source.

Histological analysis

   Lungs, salivary glands, tibialis muscles, and visceral adipose tissue
   were fixed for at least 24 hours with 4% paraformaldehyde and then
   embedded in paraffin. Sections were stained with hematoxylin (Sakura
   Fintek Japan) and eosin (Sakura Fintek Japan) and examined by light
   microscopy (BZ-X810, KEYENCE).

   Sciatic nerves were removed, cut, and fixed with 2% glutaraldehyde and
   2% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4). Samples were
   washed twice in phosphate buffer and postfixed in osmium tetroxide,
   then dehydrated in an ascending ethanol gradient, and then embedded in
   epoxy resin. Semithin sections (0.5 μm) were cut using an
   ultramicrotome and stained with 1% toluidine blue sodium borate stain.

SA β-galactosidase assay

   For the β-galactosidase quantitative assay, kidney was lysed with the
   T-PER Tissue Protein. Extraction reagent (Thermo Fisher Scientific,
   78510) was used, lysates were centrifuged, and the supernatant was
   collected. Lysates were mixed with 50 μl of Pierce β-galactosidase
   assay reagent (Thermo Fisher Scientific, 75705) and incubated for 30
   min, and then the absorbance was measured at 405 nm.

Adoptive transfer experiment

   A total of 3 × 10^6 CD45.1 WT CD4^+ CD25^− T cells were transferred
   into WT or RIPK1-tKO mice, and transferred cells were collected 7 days
   after transfer and subjected to flow cytometric and qPCR analysis. A
   total of 3 × 10^6 WT or RIPK1-KO or aged CD4^+ CD25^− T cells were
   transferred into CD45.1 WT mice, and transferred cells were collected 4
   days after transfer and subjected to flow cytometric and qPCR analysis.

Quantification and statistical analysis

   Statistical significance was determined by a two-tailed unpaired
   Student’s t test using KaleidaGraph (Synergy Software). Data are
   presented as the mean ± SEM values. P < 0.05 was considered
   statistically significant.

Acknowledgments