Constitutive Activation of gp130 in T Cells Results in Senescence and Premature Aging

The IL-6 family of cytokines are crucial players in the coordinated activity of both the innate and the adaptive immune system and are involved especially in T and B cell responses (1, 2). Most of the known autoimmune, inflammatory, or myeloma diseases are linked to a dysregulation in the IL-6/JAK/STAT3 signaling axis (3–6). IL-6 binding to its IL-6R (IL-6Rα) forms an IL-6/IL-6R complex that leads to a dimerization of the signal-transducing receptor chain gp130, which activates several intracellular pathways, including the JAK/STAT3, Ras/MAPK, and PI3K/AKT signaling cascades (7–11). Importantly, the activation of STAT3 induces a multitude of target genes that orchestrate important physiological functions, such as the regulation of cell migration, proliferation and differentiation, angiogenesis, and cell survival (12, 13). Increasing evidence indicates that dysregulation of IL-6/gp130 formation and subsequent JAK/STAT3 signaling are not only associated with pathologic immune responses but also lead to chronic lymphoproliferative disorders and cancer (14–16). Consistently, permanent genetically driven IL-6 pathway activation (by gain-of-function mutations in gp130) leads to reactive oxygen species follwed by a strong development of premalignant cells and dysplastic intrahepatic nodules in chronic liver inflammation (17). In addition, senescence is initiated after an external stress imposed on the tissue. Senescent cells arrest in the cell cycle, encounter morphological changes, and develop a senescence-associated secretory phenotype (18, 19). Cytokines, such as IL-6, are a known senescence-associated secretory phenotype component that can promote or sustain senescence, including cytokine signaling pathways, such as JAK/STAT, that contribute to senescence (20–22). Therefore, a mechanistic and systemic understanding of the diverse functions of gp130 could help to better understand the development of several pathological diseases and their progression. In this study, we used a mouse model that reflects a constitutive and cytokine-independent activation of JAK/STAT3 in all T cells. Therefore, the extracellular domain of gp130 was replaced with the dimerizing human c-Jun leucine zipper, leading to forced gp130 dimerization. Lgp130 then was fused to a ZsGreen fluorescent reporter molecule, knocked into the ROSA26 locus, and crossed with CD4-Cre expressing mice (Lgp130^CD4/wt). Activation of aberrant gp130/JAK/STAT3 signaling during the double-positive stage in T cells led in several organs to disease-associated hyperinflammatory signatures that mimic an autoimmune phenotype accompanied by a shortened life span and evidence of senescence. Taken together, this phenotype represents a perfect preclinical model that could be used to assess new candidate therapies in autoimmune diseases.

Lgp130^fl/fl and CD4-Cre mice were originally obtained from one of the authors (S.R.-J., Kiel, Germany) and bred in the Animal Facility of the Heinrich-Heine-University of Düsseldorf, Germany. Both mouse lines were crossed under specific pathogen-free conditions, and experiments were performed with male Lgp130^CD4/wt and Lgp130^fl/wt control mice at the age of 8–12 wk, including one experimental setting until week 18. This study was carried out in accordance with the Declaration of Helsinki and the guidelines from Directive 2010/63/EU of the European Parliament on the protection of animals. The protocol was approved by the local authorities at Heinrich-Heine-University and by the Landesamtes für Natur- Umwelt und Verbraucherschutz (Germany, Nordrhein-Westfalen; Permit No. 81-02.04.2019, A064, O 117/11, O 23/19). Littermate controls (CD4-Cre–negative mice) were used as the control group. Physiological parameters such as total blood cell count, platelet isolation from the peripheral blood, body weight, and organ weight were determined, accompanied by histology, RNA, or protein analysis of spleen, liver, and heart tissue.

Histological analysis from snap-frozen tissue (liver, spleen) and paraffin-embedded heart tissue was performed with H&E staining (catalog no. GHS216; Sigma-Aldrich) and TUNEL (catalog no. 11684795910; Roche). For the analysis of platelet megakaryocytes (MKs) in the spleen and in the bone marrow (BM) of the femur of untreated mice, 5-μm sections of paraffin-embedded tissue were prepared followed by H&E staining. For each mouse/genotype, at least three spleen and BM sections were stained, three to five pictures per section were taken, and the MKs per visual field were counted. For immunohistochemistry, frozen lung tissues were prepared by inflating the lungs through the trachea with optimum cutting temperature compound, followed by tying off the trachea to maintain the fluid in the lung. The tissues were then slowly frozen in a vessel on 2-methyl-1-isopropanol on dry ice. Frozen sections were cut at 8 µm in a cryostat (CM3050S; Leica). The cut sections were treated with 4% formaldehyde for 15 min after blocking with 5% donkey serum for 1 h. Anti-CD45 primary Ab (catalog no. AF114; R&D Systems) was applied overnight at 4°C followed by anti-goat–NL493 secondary Ab (catalog no. NL003; R&D Systems) for 1 h.

Different immune populations were identified in single-cell solutions from naive spleen, thymus, and lymph node samples using several Abs (anti-CD3, CD4, CD8, CD25, CD44, and Foxp3; all from Thermo Fisher Scientific). For quantification of cell populations, calibration beads were added to access cell numbers (BD, San Diego, CA). Experiments were analyzed using FlowJo software. Murine platelet activation was performed using fluorophore-labeled Abs for P-selectin expression (Wug.E9-FITC; Emfret Analytics) and the active form of α_IIbβ₃ integrin (JON/A-PE; Emfret Analytics). Heparinized blood was diluted in Tyrode’s buffer and washed twice. Blood samples were mixed with Abs after addition of 2 mM CaCl₂ and stimulated with indicated agonists for 15 min at room temperature. Reaction was stopped by the addition of PBS, and samples were analyzed on a FACSCalibur and LSRFortessa flow cytometer (BD Biosciences). For analysis of glycoprotein surface expression, blood samples were mixed with Abs (GPVI, JAQ1-FITC; GPIb/CD42, Xia.G5-PE; integrin α₅/CD49e, Tap.A12-FITC; Emfret Analytics) and incubated for 15 min at room temperature. For determination of platelet size, the geometric mean of the CD42-positive platelets in their forward scatter profile was used.

RNA purification was performed according to the manufacturer’s instructions (TRIzol; Invitrogen). Gene expression of target genes (Supplemental Table I) was performed by iTaq Universal SYBR Green One-Step Kit (catalog no. 1725151; Bio-Rad). For analysis, the expression levels of all target genes were normalized to the expression of Lgp130^fl/wt (Δ cycle threshold). Gene expression values were calculated based on the ΔΔ^Ct method. Relative quantities (RQs) were determined using the equation: RQ = 2^−ΔΔCt.

In brief, liver, spleen, and heart tissue were lysed in lysis buffer (50 mM Tris-HCl [pH 7.5], 150 mM NaCl, 2 mM EDTA [pH 8.0], 2 mM NaF, 1 mM Na₃VO₄, 1% Nonidet P-40, 1% Triton X-100, and one cOmplete protease inhibitor mixture tablet per 50 ml). After lysis, the protein content was measured by Bicinchoninic acid assay. A total protein amount of 50 µg was then loaded per lane followed by immunoblotting. Abs used for protein detection were as follows: anti–p-STAT3 (catalog no. 9145), anti–total-STAT3 (catalog no. 9139), anti–p-STAT1 (catalog no. 9167), anti–total-STAT1 (catalog no. 9172), anti–p-STAT5 (catalog no. 9359), anti–total-STAT5 (catalog no. 94205), anti–p-Akt (catalog no. 4060), anti–total-Akt (catalog no. 9272), anti–p-p44/42 (catalog no. 9106), anti–total-p44/42 (catalog no. 9102), and anti-GAPDH (catalog no. 2118) were purchased by Cell Signaling. Anti-p53 (catalog no. sc-126) was purchased from Santa Cruz Biotechnology.

Mouse IL-6 (catalog no. 88-7064-88; Invitrogen) and mouse IL-6R α DuoSet ELISA (soluble IL-6R [sIL-6R]; DY1830; R&D Systems) were measured according to the manufacturer’s protocols. Mouse anti-nuclear Abs, mouse anti-dsDNA Abs total Ig (catalog no. 5210 and 5110, respectively), and mouse Rheumatoid Factor (RF) Ig's [total (A+G+M) ELISA Kit; catalog no. 6200] were purchased by Alpha Diagnostic International, and assays were performed according to the manufacturer’s instruction.

Murine blood from retro-orbital plexus was collected in 300 µl heparin solution (20 U/ml in PBS), and total blood cell counts were analyzed by a hematology analyzer (Sysmex KX-N21 [Norderstedt, Germany] and VetScan HM5v2.31 [Lab Technologies Medizintechnik, Vienna, Austria]).

Aspartate aminotransferase (glutamate-oxalacetate transaminase [GOT]), alanine aminotransferase (glutamic-pyruvic transaminase), total bilirubin, total cholesterol, triglycerides, glucose, amylase (AMY), high-density lipoprotein (HDLc), albumin, lactate dehydrogenase, creatine phosphokinase, and alanine phosphatase were measured using the automated biochemical analyzer Spotchem EZ SP-4430 (Arkray, Amstelveen, the Netherlands) and the Spotchem EZ Reagent Single Stripes or Multi Stripes Liver-1, Kenshin-1.

Bile acids and their taurine derivatives in serum were analyzed by ultra-performance liquid chromatography–tandem mass spectrometry (UPLC-MS/MS). The system consists of an ACQUITY UPLC I-Class (Waters, U.K.) coupled to a Waters Xevo-TQS MS/MS equipped with an electrospray ionization source in the negative ion mode. Data were collected in the multiple reaction monitoring mode.

Senescence β-galactosidase activity was measured by an assay kit (catalog no. 23833S; Cell Signaling) according to the manufacturer’s protocol.

Data are provided as arithmetic means ± SEM using GraphPad Prism, Version 8. Statistically significant differences between two groups were determined with a Student t test, including Welch’s correction if indicated. Statistical analysis between several groups was determined using a two-way ANOVA including Tukey correction. Significance was calculated as follows: ^n.s.p > 0.05 (not significant), *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

Lgp130^fl/wt mice carrying the CAG-(loxP)STOP(loxP)-Lgp130-2A-ZsGreen cDNA in the Rosa26 locus were crossed with heterozygous mice expressing the Cre recombinase under the control of the CD4 promoter (Fig. 1A, 1B) (23, 24). Expression of Lgp130 was induced after deletion of the (loxP)STOP(loxP)-cassette, placing Lgp130-2A-ZsGreen under the control of the CAG promoter. Coexpression of ZsGreen was used as marker for recombination efficiency and Lgp130 transgene expression. During thymic development, CD4 is first expressed in the double-positive CD4/CD8 state of all T cells, thus resulting in transgenic expression of Lgp130 in both CD4- and CD8-positive T cells on Cre-mediated recombination. Flow cytometric analysis of thymic T cells from Lgp130^CD4/wt mice showed that CD4- and/or CD8-positive T cells, but not CD4/CD8-negative T cells, expressed ZsGreen (Fig. 1C), whereas control Lgp130^fl/wt mice did not express ZsGreen (Fig. 1C). In the lymph nodes from Lgp130^CD4/wt mice, but not Lgp130^fl/wt mice, peripheral CD4- and CD8-positive T cells, but not peripheral CD19-positive B cells, were found to express ZsGreen (Fig. 1D, 1E, Supplemental Fig. 4). Consequently, ZsGreen fluorescence was detected in thymus, spleen, and lymph nodes but only minimally in the liver of Lgp130^CD4/wt mice (Fig. 1F, Supplemental Fig. 4), which is consistent with the low amount of T cells in the liver (25, 26). Lgp130^CD4/wt mice displayed significantly elevated transcript levels of lgp130 in total spleen in comparison with Lgp130^fl/wt (Fig. 1G). The expression of Lgp130 in T cells goes along with statistically significant phenotypic abnormalities. Median survival of Lgp130^CD4/wt mice was ∼45 d (Fig. 2A). In addition, Lgp130^CD4/wt mice showed indications of embryonic lethality because of a reduced ratio of Lgp130^CD4/wt over Lgp130^fl/wt mice among offspring (Fig. 2B). Consistently, body weight at 6 wk of age was significantly lower (30%) in Lgp130^CD4/wt compared with Lgp130^fl/wt mice (17.5 versus 24.6 g) (Fig. 2C, 2D). Spleen-to-body weight ratio revealed splenomegaly in Lgp130^CD4/wt mice accompanied by elevated heart-, lung-, lymph node-, spleen-, and thymus-to-body weight ratios, whereas the liver-to-body weight ratio was not altered (Fig. 2E, 2F). Surprisingly, the hearts of Lgp130^fl/wt and Lgp130^CD4/wt mice presented no signs of severe hypertrophy or dilation. For this reason, we would conclude that the effect of the increased heart weight might be a response to an increased immune cell infiltration (Supplemental Fig. 3E–3G).

Blood analysis showed that there was an overall reduction in WBC counts with an increase of granulocytes and monocytes and a decrease of lymphocytes in Lgp130^CD4/wt mice as compared with Lgp130^fl/wt control mice (Fig. 3A–D). In addition, RBCs were affected in Lgp130^CD4/wt mice by a significant decrease in total hemoglobin and the hematocrit (Fig. 3E, 3F). In regards to lymphocytes, Lgp130 expression resulted in dramatically reduced numbers of CD4/CD8 double-positive T cells, i.e., 80.6% in Lgp130^fl/wt mice as compared with 9% in Lgp130^CD4/wt mice (Fig. 3G). This was accompanied by a strong increase of CD4/CD8-negative T cells from 2.2 to 49.7% and CD8-positive cells from 6.1 to 25.7% in Lgp130^CD4/wt mice compared with Lgp130^fl/wt mice (Fig. 3H, Supplemental Fig. 6A). CD4/CD8/CD3 triple-negative thymocytes are subdivided into four subsets based on expression of CD44 and CD25, along their developmental stage from CD44⁺CD25⁻ to CD44⁺CD25⁺, CD44⁻CD25⁺, and CD44⁻CD25⁻ (27). In Lgp130^CD4/wt mice, the least mature cell stage (CD44⁺CD25⁻) was clearly overrepresented (Fig. 3I, Supplemental Fig. 6A). Based on the frequencies, we can conclude that Lgp130^CD4/wt mice appear to arrest in the CD4⁻CD8⁻ and CD25⁻CD44⁺ stage during T cell development within the thymus (Supplemental Fig. 1C, 1D). CD8-dominated CD4/CD8 T cell ratios were observed in the spleen and lymph nodes of Lgp130^CD4/wt mice compared with control Lgp130^fl/wt mice (Fig. 3J, Supplemental Fig. 6B). Moreover, we observed significantly increased numbers of regulatory T (T_reg) cells in lymph nodes and a higher tendency in spleens of Lgp130^CD4/wt mice (Fig. 3K, Supplemental Fig. 1E, 6C). Peripheral CD4⁺ T cells showed reduced surface CD44 expression and increased CD69 levels in the spleen. Similar findings were obtained for CD44 on CD8⁺ T cells in the lymph nodes and the spleen, although CD69 expression had only an increased tendency (Supplemental Fig. 1A, 1B). For B cells, we observed an increase of CD19-positive B cells in the lymph nodes, but not in the spleen, of Lgp130^CD4/wt mice compared with Lgp130^fl/wt mice (Fig. 3L, Supplemental Fig. 1F, 6B). Taken together, our data showed that T cell–restricted expression of Lgp130 resulted in increased levels of cytotoxic T and T_reg cells. Importantly, due to the higher expression of CD69, peripheral T cells are in a more activated state compared with wild-type (wt) mice.

In the spleen and liver of Lgp130^CD4/wt mice, we observed an increased STAT1 and STAT3 phosphorylation, with a tendency of higher STAT5 phosphorylation as compared with Lgp130^fl/wt mice, whereas in the heart, STAT1 and STAT5, but not STAT3, phosphorylation was increased (Fig. 4A, Supplemental Fig. 2). As seen before for constitutively active gp130 (23, 28, 29), increased levels of p-ERK and p-AKT were not detected in Lgp130^CD4/wt mice (Supplemental Fig. 2). Next, quantitative mRNA expression analysis revealed a significant increase of SOCS3 and a downregulation of TNF-α, IL-10, and CCL5/RANTES in the spleen of Lgp130^CD4/wt mice (Fig. 4B). IL-6, TNF-α, IL-10, SOCS3, IFN-γ, CCL5/RANTES, and the acute-phase response protein SAA1 were highly upregulated in the liver of Lgp130^CD4/wt mice as compared with Lgp130^fl/wt control mice (Fig. 4B). In the heart of Lgp130^CD4/wt mice, IL-6, TNF-α, IL-10, SOCS3, IFN-γ, and CCL5/RANTES mRNA were significantly elevated (Fig. 4B). In heart and spleen, α smooth muscle actin was significantly downregulated (Fig. 4B). IL-6 protein in the serum of Lgp130^CD4/wt mice was found to be significantly increased compared with Lgp130^fl/wt mice (Fig. 4C). Also, the sIL-6R serum level was upregulated in Lgp130^fl/wt mice (Fig. 4D). Due to this overall proinflammatory phenotype in Lgp130^fl/wt mice, we also found a strong upregulation of lupus-related autoantibodies against anti-nuclear IgG, dsDNA, and rheumatoid arthritis (RA)-related RF in the serum of Lgp130^CD4/wt mice compared with Lgp130^fl/wt control mice (Fig. 4E–G). Moreover, serum from Lgp130^CD4/wt and Lgp130^fl/wt control mice was applied to B cell–deficient spleen tissue to independently assess the degree of autoimmune Abs. As shown in Fig. 4H, almost no Abs from serum of Lgp130^fl/wt control mice bound to B cell–deficient spleen tissue, whereas strong binding was observed for serum from Lgp130^CD4/wt mice. Histological analysis of the liver and heart tissue from Lgp130^CD4/wt mice revealed an increase in the infiltration of immune cells, whereas spleen tissue revealed a dramatic disruption of splenic architecture with loss of splenic marginal zones and the T cell/B cell organization (Fig. 5A). Increased apoptosis was, however, not observed by TUNEL staining of liver, heart, and spleen (Fig. 5B, 5C). Immunohistochemical staining for inflammatory (CD45) cell infiltration in lung tissue was significantly elevated in Lgp130^CD4/wt mice compared with Lgp130^fl/wt mice. This supports the overall polyautoimmune phenotype, but lung-specific changes in the architecture were not further investigated (Supplemental Fig. 3C, 3D). In contrast, expression of cell-cycle genes such as Cyclin A, Cyclin D, and Cyclin E were significantly upregulated in the liver of Lgp130^CD4/wt mice as compared with Lgp130^fl/wt control mice, accompanied by a small induction of the proliferation marker Ki67 (Fig. 5D). Analysis of serum parameters revealed increased GOT, AMY, and alkaline phosphatase (Fig. 5E) and decreased total cholesterol, HDLc, triglyceride, and Glu in Lgp130^CD4/wt mice as compared with Lgp130^fl/wt mice (Fig. 5F). Functional liver parameters such as albumin and total bilirubin were rather reduced in Lgp130^CD4/wt compared with Lgp130^fl/wt mice, whereas glutamic-pyruvic transaminase, LDH, and creatine phosphokinase were slightly increased in Lgp130^CD4/wt mice (Supplemental Fig. 3A). Given that bile acids have been known to regulate cholesterol homeostasis and are recognized as hormones involved in the regulation of various metabolic processes (30), bile acid profiling in the serum was performed next. Lgp130^CD4/wt mice revealed compared with Lgp130^fl/wt controls a significant increase in the overall level of serum bile acids, especially taurine-conjugated and unconjugated bile acids (Supplemental Fig. 3B). Taken together, our data indicate massive liver damage accompanied by cholestasis and an overall metabolic imbalance.

Platelets are essential mediators of hemostasis to avoid blood loss by controlling thrombus formation (31). Therefore, blood cell counts of RBCs and platelets, as well as platelet size, were analyzed in whole blood. In Lgp130^CD4/wt mice, RBC count was significantly reduced by ∼20% (Fig. 6A), whereas platelet counts were massively increased from 500 × 10³/µl in Lgp130^fl/wt mice to 5000 × 10³/µl in Lgp130^CD4/wt mice (Fig. 6B). Moreover, a tendency for an increased platelet size was observed (Fig. 6C). In spleen and BM sections, we estimated platelet formation by megakaryopoiesis. The number of MKs was significantly elevated by 30% in spleen and 20% in BM of Lgp130^CD4/wt compared with Lgp130^fl/wt mice (Fig. 6D–F). Platelet activation capacity was analyzed by flow cytometry and Ab binding to activated integrin αIIbβ3 (fibrinogen receptor) and determination of P-selectin exposure as a marker for degranulation. Overexpression of Lg130 in T cells resulted in a significant reduced degranulation measured via P-selectin exposure on the surface of platelets by flow cytometry. Activation of the major ITAM-coupled collagen receptor (GPVI) via collagen-related peptide binding led to a reduced P-selectin degranulation with all tested concentrations. A reduced degranulation was also measured in combination of both autocrine agonists ADP and U46 (thromboxane A₂ analogue) targeting G-protein–coupled receptors on the platelet surface. Only by trend, a reduced degranulation was measured upon PAR4-peptide binding, activating the thrombin-mediated pathway of platelet activation. Interestingly, a mild preactivated state regarding degranulation could be observed in unstimulated platelets (resting control [−]) that stayed unaffected upon stimulation with a mild agonist such as ADP alone (Fig. 6G, Supplemental Fig. 5D). As additional marker, integrin αIIbβ₃ (fibrinogen receptor) activation was measured. Although platelet degranulation upon Lgp130 overexpression in T cells was highly deregulated, integrin activation stays unaffected in unstimulated control samples as with every tested agonist (Fig. 6H, Supplemental Fig. 5C). Moreover, platelets from Lgp130^CD4/wt mice had reduced surface expression of the glycoprotein GPVI and integrin α₅ compared with Lgp130^fl/wt mice, whereas GPIbα and integrin β₃ expression were not significantly changed (Fig. 6I, Supplemental Fig. 5A, 5B). A reduction of GPVI, as receptor for platelet activation and regulator of thromboinflammation (32), implicates an increase in shedding, but because of a basal preactivation, this defect in activation might only be a “secondary” effect. Aging platelets lose sialic acid, the terminal carbohydrate moiety that covers the underlying galactose residues, a ligand for the hepatic Ashwell-Morell receptor (AMR). Desialylated platelets are removed from the circulation via the AMR to regulate thrombopoietin (TPO) production by JAK2-STAT3 (33). To analyze whether platelet clearance in Lgp130^CD4/wt mice is associated with a glycan-dependent mechanism via hepatocytes or Kupffer cells, gene expression of liver samples was investigated. Analysis of liver gene expression revealed enhanced TPO expression resulting in enhanced megakaryopoiesis. As already shown, enhanced JAK-STAT signaling occurs in liver tissue of Lgp130^CD4/wt mice compared with wt mice (Fig. 4A). TPO gene expression is regulated upon different receptor pathways in the liver (34). First, IL-6R and AMR (consisting of asialoglycoprotein receptor [Asgr] 1 and Asgr2 subunits) signaling on hepatocytes induces TPO gene expression. Although overall enhanced TPO is observed, a reduction of the Asgr2 subunit with an overexpression of the IL-6R targets toward an IL-6–driven pathway. However, it was recently discovered that Kupffer cells form a platelet clearance receptor consisting of the Asgr1 and Mgl subunits. Because the Mgl subunit is overexpressed, a Kupffer cell–hepatocyte interaction might be relevant (Fig. 6J). Taken together, the increase in MKs in spleen and BM contributes to the permanently increased number of platelets, resulting in a reactive thrombocytosis in Lgp130^CD4/wt mice compared with Lgp130^fl/wt control mice.

The STAT3/IL-6 axis induces ROS in an IGFBP5-dependent manner (35, 36). As a consequence, ROS-induced DNA damage induces the senescence-induced secretory pathway and p53 stabilization, which resulted in cell-cycle arrest and ultimately in cellular senescence (35–37). IL-6 knockout mice have a reduced p53 level and show fewer signs of cellular senescence (38–40), and IL-6 in complex with the sIL-6R induces premature senescence in human fibroblasts (35). Apart from constitutive activation of STAT3 in Lgp130^CD4/wt mice, these mice showed overall increased IL-6 mRNA levels and an increase of IL-6 and sIL-6R protein in the serum compared with Lgp130^fl/wt control mice (Fig. 4B, 4C). Therefore, we analyzed p53 expression and senescence in Lgp130^CD4/wt mice. By using Western blotting, we observed a strong increase of p53 levels in the heart, liver, and spleen of Lgp130^CD4/wt mice as compared with Lgp130^fl/wt mice (Fig. 7A). In line with this, also the senescence-associated cell-cycle inhibitor p21^WAF was strongly upregulated on the mRNA level in heart, liver, and spleen of Lgp130^CD4/wt mice as compared with Lgp130^fl/wt mice (Fig. 7B). Senescence-associated β-galactosidase staining was performed using spleen and heart tissues. Whereas the spleen seemed to be rather unaffected, a significant increase in senescence-associated β-galactosidase staining of heart sections was found in Lgp130^CD4/wt mice compared with Lgp130^fl/wt mice (Fig. 7C). Taken together, our data indicate an increased senescence phenotype of Lgp130^CD4/wt mice that likely contributes to premature aging and death of the mice.

All known IL-6–type family members signal via the signal transducing β-subunit gp130 with different biological outcomes and efficacies (41). Several variants of constitutively active gp130 have been described in the literature combined with numerous diseases underlining the importance of the gp130/STAT3 axis. Downstream of gp130, the gain of function of gp130 was early described to develop autoimmune diseases such as RA in mice (42, 43). Recently, it was shown that a B cell–specific constitutively active gp130 variant induced mature lymphoma and plasmacytoma (24). Our study (to our knowledge) and the accompanying study by Heinig et al. (44) define a novel mouse model that mimics T cell–specific gp130/STAT3 signaling in a cytokine-independent manner. Lgp130CD4/wt mice show massive phenotypical abnormalities, including splenomegaly, lymphadenopathy, and an upregulation of hyperinflammatory and autoimmune-like signatures including autoantibodies (45, 46). Furthermore, permanent genetically driven JAK/STAT pathway activation in T cells showed premature aging based on upregulation of cell-cycle markers, accompanied by high levels of the tumor suppressor p53 and increased expression of the cyclin-dependent kinase inhibitor p21^WAF. This accumulation of multiple senescence markers together with chronic inflammation might lead to the observed morphological changes in the organs, lastly resulting in multiorgan failure and the observed aging phenotype and strongly reduced survival. This hypothesis is supported by Kojima et al. (36), who already provided evidence that IL-6, and especially the IL-6/STAT3 axis, induced senescence in human fibroblasts. Our data are consistent with reports describing the complex role of T_reg cells in immunity and aging, which is also in connection with increased numbers of cytotoxic and T_reg cells seen in the T cell–restricted Lgp130 mice (47, 48). The accompanying phenomenon of inflammation and premature aging is summarized as inflammaging, which was first introduced by Franceschi et al. (49) in 2000 as a new branch of aging studies. Inflammaging itself is associated with many aging-associated diseases, such as Alzheimer’s, heart disease (i.e., myocardial infarction, aortic aneurysm), atherosclerosis, cancer, and type II diabetes, and is often linked to and potentially caused by immunosenescence. However, the underlying mechanisms are poorly understood (50–51). Moreover, T cell–restricted expression of Lgp130 and its permanent inflammatory signature showed severe effects on hemostasis. Besides an elevated platelet production and an increase in the number of MKs in spleen and BM, platelet clearance was permanently increased and accompanied by secondary anemia. Consistent with this, platelets seemed to be overreactive that manifested in a reactive thrombocytosis leading to defects in platelets activation and shortened life span in Lgp130^CD4/wt mice. Collectively, our results are consistent with the results described by Kirito et al. (52) in 2002, as well as the indispensable role of STAT3 signaling in the early stage of megakaryopoiesis, presumably in the regulation of expansion of megakaryocytic progenitor cells. Abnormal thrombopoiesis via elevated IL-6 production, also seen in our study, results in thrombocytosis and contributes to a high platelet count, which is potentially associated with thrombosis and death (53, 54). In summary, to our knowledge, this study presents the first causal link between a hyperactivation of the STAT3/IL-6 axis in T cells and inflammaging. Moreover, the Lgp130^CD4/wt mouse model can be used for further and detailed senescence studies, especially in connection with inflammaging. For the treatment of RA, TNF-α inhibitors, such as etanercept, were used until the IL-6R inhibitor tocilizumab outperformed anti–TNF-α agents in effectiveness (55–58). High levels of IL-6 and TNF-α are also associated in the elderly with increased risk of morbidity and mortality (59, 60). For these reasons, a treatment with an anti–IL-6 agent or JAK inhibitors to alleviate the inflammaging phenotype would be of interest as a novel proof-of-principle strategy to develop antiaging interventions. Another important strategy to be considered for further investigations would be to effectively eliminate senescent cells by compounds termed “senolytics.” Previous studies have shown that senolytics target antiapoptotic pathways and eliminate certain types of senescent cell with the potential to improve physiological functions in several tissues (61, 62).

The authors have no financial conflicts of interest.

We thank Martina Spelleken for technical assistance. We are grateful to S.R.-J. (Institute of Biochemistry, Christian-Albrechts-University Kiel, Kiel, Germany) for the provision of the Lgp130^fl/fl mice. We thank SFB 1116 “Master Switches bei kardialer Ischämie” for funding this project.