Activation of mTOR (mechanistic target of rapamycin) in rheumatic diseases

Intracellular localization and trafficking

The presence of mTOR in several cellular compartments could be key to its ability to sense stress and execute cell-growth signals. mTORC1 was initially detected in the cytoplasm as a stimulator of protein translation by its ability to phosphorylate S6K²¹ and 4E-BP1 (REF. 31). Later, observations showing that mTOR was associated with endosomal compartments brought the trafficking-regulatory small Rab GTPases Rab7 and Rab4A, as well as the Rag family of Ras-related GTP-binding proteins, to the forefront of the cell-signalling field^30,32. mTOR was also reported to translocate to the outer mitochondrial membrane, where it senses changes in the mitochondrial transmembrane potential, ATP depletion and oxidative stress³³. Additionally, mTOR can also be detected inside the nucleus, where it might have a role in the transcription of RNA polymerase I-dependent and RNA polymerase III-dependent genes^34,35. Under conditions of cell stress, mTORC1 is sequestered in an astrin (also known as sperm-associated antigen 5)-dependent manner within stress granules, limiting cell growth and promoting cell survival²⁹.

Consensus is increasing that mTORC1 activation occurs at the surface of the lysosomal membrane in response to changes in amino acid sufficiency²⁷, which is transduced via the Rag family of small GTPases to mediate the translocation of mTORC1 from the cytoplasm to the surface of the lysosome, where mTORC1 is activated by GTP-binding protein Rheb (FIG. 2).

The translocation of mTOR to the lysosomal membrane occurs via endosome traffic that is regulated by Rab7 (REF. 32), Rab5 (REFS 36–38) and Rab4A. Each of these GTPases also regulates autophagy^37,39–42 (FIG. 2). Rab4A forms a positive feedback loop with mTORC1 and negative feedback loop with mTORC2 (REF. 43), and binds to the p85 regulatory subunit of phosphatidylinositide 3-kinase (PI3K), altering its ability to control endocytic recycling^44–47. Notably, Rab4A and Rab5 are overexpressed in T cells of patients with SLE³⁰ and in mouse models of this disease⁴¹; thus, these small GTPases are possible regulators of abnormal T-cell signal transduction in SLE. Furthermore, overexpression of Rab4A, but not Rab5, also precedes the production of antinuclear antibodies (ANA) and SLE onset⁴¹. Importantly, pharmacological inhibition of Rab GTPases by 3-(3-pyridyl)-2-hydroxy-2-phosphono propanoic acid (3-PEHPC) prevents T-cell dysfunction, ANA production, and nephritis in lupus-prone mice⁴¹ (FIG. 2).

The cell-type-specific activation of mTORC1, and its localization to the lysosome during autophagy, are subject to intricate control by metabolic regulatory net-works⁴⁸. Moreover, the traffic of endosomes to lysosomes has long been implicated in interferon production and inflammation, because these structures also carry Toll-like receptors, which are able to sense nucleic acids, particularly in dendritic cells (DCs)⁴⁹.

Metabolic regulatory networks

Signalling pathways that control the proliferation, survival and differentiation of cells in the immune system also regulate the metabolic processes that provide the nutrients required to support these specialized lymphocyte functions⁵⁰. Although mTOR drives proinflammatory lineage specification in the T-cell compartment⁵¹, it might also have anti-inflammatory effects driven by shifting macrophage polarization from a proinflammatory M1 phenotype to an anti-inflammatory M2 phenotype (REFS 52,53). Nevertheless, M2 macrophages can also contribute to inflammation, for example by serving as hosts to cytomegalovirus (CMV), which might explain the potent anti-CMV effects of mTOR inhibitors after organ transplantation⁵⁴.

Outside the immune system, mTOR activation contributes to type I collagen production by dermal fibroblasts and to fibrosis in patients with SSc^11,12. In chondrocytes, genetic inactivation of mTOR prevented the development of OA induced by peroxisome proliferator-activated receptor-γ (PPARγ) deficiency⁵⁵. Furthermore, mTOR blockade with N-acetylcysteine (NAC), an amino acid precursor of glutathione, reduced cognitive attention deficit hyperactivity disorder symptoms in patients with SLE⁵⁶. Given all the evidence for a pathogenetic role for mTOR in rheumatic diseases and associated comorbidities, the metabolic pathways sensed and regulated by this signalling molecule are important to understand.

Stimulatory and inhibitory signal transducers

The mTOR pathway is largely controlled by upstream checkpoints at three levels: receptor tyrosine kinases and G-protein-coupled receptors, which detect growth factors; the PI3K–PDK1 (phosphoinositide-dependent kinase-1)–AKT (RAC-α serine/threonine-protein kinase) axis, which channels stimulatory signals towards mTORC1 activation; and the key negative regulators PTEN, AMPK (5′-AMP-activated protein kinase catalytic subunit α2), TSC1 and TSC2 (the latter two are also known as hamartin and tuberin, respectively).

Cell-type-specific activation of mTOR

The mTOR pathway has critical roles in the development and function of the innate and adaptive arms of the immune system. An essential role of mTOR in T-cell development was uncovered⁵¹ when mTORC1 was found to be required for the differentiation of T_H1 and T_H17 cells, whereas T_H2-cell development depends on mTORC2 (REF. 51). mTORC1 also inhibits the survival of CD8⁺ memory T cells, an effect that can be reversed by rapamycin treatment^99,100. Both mTORC1 and mTORC2 seem to interfere with the differentiation and function of CD4⁺CD25⁺FoxP3⁺ T_REG cells^51,77,101, but mTORC1 might also support T_REG-cell function¹⁰² by inhibiting the mTORC2 pathway¹⁰³.

Despite its complex role in T_REG-cell function, the largely proinflammatory changes elicited by mTORC1 activation in the adaptive immune system are in apparent contrast with its anti-inflammatory effects on the innate immune system (for example, mTORC1 favours M2 over M1 polarization of macrophages⁵²). The cell-type-specific activation of mTORC1 also varies between rheumatic diseases, as mTORC1 is activated in T cells of patients with SLE³⁰, but not in those of patients with RA¹⁰⁴. These divergent effects of mTOR activation in T cells might be connected to alternative use of glucose between glycolysis and the PPP¹⁰⁵.

mTORC1 activation in SLE

In patients with SLE, mTORC1 is activated^30,70 and mTORC2 is inhibited¹⁰⁶ (FIG. 2). A central role for mTORC1 in abnormal T-cell lineage specification and function in SLE^30,107,108 is consistent with its effects on shaping T-cell development⁵¹.

The involvement of mTOR in SLE was initially suggested by the successful blockade of T-cell hyperactivity and nephritis in rapamycin-treated lupus-prone MRL/lpr mice⁵. Later, rapamycin was shown to block T-cell activation in patients with SLE³⁰, with remarkable therapeutic efficacy^6,109. Interestingly, activation of mTORC1 preceded disease flares by 4 months, and was reduced by therapeutic intervention with rapamycin¹⁰⁹. Thus, measurement of mTORC1 and mTORC2 activity by intracellular staining for phosphorylated 40S ribosomal protein S6 (at Ser235 and Ser236) and phosphorylated AKT (at Ser473), respectively, can be used as biomarkers for pathogenesis, prediction of flares and monitoring of treatment efficacy of mTOR blockade in SLE^70,106,109. Measuring the activity of mTORC1 and mTORC2 in cells of patients with SLE and other autoimmune diseases could be another step in achieving personalized or precision medicine¹¹⁰.

Rapamycin blocks the proinflammatory, necrotic death of DN T cells and the depletion of T_REG cells¹⁰⁹, and mTORC1 inhibition by NAC had similar effects¹¹¹ (FIG 1,FIG 2; TABLE 1). Diminished T_REG-cell frequency has been linked to mTORC1-sensitive methylation of the Foxp3 promoter¹¹², whereas rapamycin treatment reduced the production of IL-4 by DN T cells in vivo¹⁰⁹, accounting for an increased production of anti-DNA autoantibodies by B cells^113,114. The benefits of rapamycin can also be attributed to T_REG-cell expansion¹⁰⁹ via activation of mTORC2, at least in patients with SLE¹⁰⁶. Given the role of mTORC2 signalling in T_REG cells and in regulating the development of T_FH cells^89,90,115, which are expanded in mouse models of lupus¹¹⁶ and in patients with active SLE¹¹⁷, the impact of rapamycin on the proinflammatory DN T-cell subset merits further investigation. The documentation of the first clinical case of fulminant SLE in a patient with tuberous sclerosis also support a fundamental role of mTOR pathway activation in the pathogenesis of both these diseases¹¹⁸: in accordance with a negative regulatory role for the TSC1–TSC2 complex, all lymphocyte subsets of this patient exhibited robust mTORC1 activation¹¹⁸. mTORC1 is also activated in B cells¹¹⁹ and nephritic kidneys of lupus-prone mice¹²⁰ (FIG. 3).

Antiphospholipid antibodies (aPL) trigger considerable pathology and are one of the diagnostic criteria for SLE^121–123. Moreover, aPL can elicit antiphospholipid syndrome (APS) in patients with or without SLE¹²⁴. As shown in a 2014 study, seven of ten patients with APS nephropathy treated with rapamycin (70%) had a functioning allograft 144 months after transplantation, versus only three of 27 patients who were not treated with rapamycin (11%)¹²⁵. Notably, the majority of the patients with APS had also been diagnosed with SLE (20 of 32 patients, 62%)¹²⁵. Unfortunately, the researchers did not document how many of the seven patients who benefited from rapamcyin¹²⁵ also had SLE. mTORC1 activity is not elevated in major PBL subsets of patients with SLE who also have APS, relative to those without APS¹²⁶. Therefore, rapamycin might have benefited renal transplant recipients with underlying SLE¹²⁵.

The mTORC1 inhibitor NAC reversed the depletion of reduced glutathione (GSH) in PBLs of patients with SLE in a double-blind, placebo-controlled, phase I–II clinical trial⁷⁰. Of relevance to its mechanism of action, NAC also reversed the strong activation of mTORC1 in DN T cells, which is consistent with a role of oxidative stress as a regulatory checkpoint. Oxidative stress originates from increased mitochondrial oxygen consumption by complex I of the electron transport chain in T cells from patients with SLE¹²⁷. Along this line, inhibition of mitochondrial oxidative stress with metformin also reduced mTORC1 activity and prevented nephritis in lupus-prone mice¹²⁸. Therefore, these findings support a role for oxidative stress in both activation of mTORC1 and the pathogenesis of lupus nephritis in mice^28,29.

In addition to oxidative stress, traffic to the lysosomal membrane and the detection of amino-acid sufficiency have been implicated in the activation of mTORC1 (REF. 27). A 2015 metabolome analysis unveiled a NAC-responsive accumulation of kynurenine in SLE PBLs⁶⁰ as the strongest metabolic predictor of SLE when comparing patients with matched healthy individuals⁶⁰. Kynurenine was also the strongest predictor of the NAC effect on the mTOR pathway in patients with SLE. In addition to lowering kynurenine levels, NAC greatly augmented the levels of NADPH, an effect thought to occur via sparing of NADPH through enhancement of de novo GSH synthesis^129,130 (FIG. 2). Importantly, kynurenine also triggered mTORC1 activation, particularly in DN T lymphocytes⁶⁰. Therefore, NAC-responsive accumulation of kynurenine is a biomarker of oxidative stress and a trigger of mTORC1 activation — two mechanistically connected metabolic checkpoints in SLE pathogenesis⁶⁰. This comprehensive metabolomic study also revealed a depletion of cysteine in SLE PBLs⁶⁰, further supporting the use of NAC treatment in SLE.

Rheumatoid arthritis

The role of mTOR is far less thoroughly characterized in RA than it is in SLE¹³¹. Unlike CD4 T cells in SLE, CD4 T cells from patients with RA show no significant changes in mTOR activity^104,132. Of note, RA T cells exhibit diminished glycolytic activity, thus depriving them of the energy required to generate oxidative stress or to execute autophagy¹⁰⁴. Despite the lack of changes in mTOR activity in RA T cells, rapamycin and rapalogues have shown efficacy in reducing joint inflammation in animal models of arthritis⁷ and in patients with RA⁸ or JIA⁹. Such clinical benefits might accrue from the inhibition of mTORC1 activation in fibroblast-like synoviocytes (FLS)¹³³ (FIG. 3). In support of this notion, rapamycin decreased the invasive properties of RA FLS¹³³. Additional evidence of the benefits of mTORC1 inhibition in RA includes the ability of IL-17 to induce mTORC1-dependent proliferation of RA FLS¹³⁴, the increase in mTORC1 activity in osteoclasts from patients with RA and in arthritic transgenic mice⁷, and the downregulation of extracellular matrix digestive enzymes and induction of apoptosis in osteoclasts elicited by mTOR inhibition⁷ (FIG. 3). Taken together, these observations suggest a therapeutic benefit from mTOR blockade in RA that might involve the intra-articular cells that mediate erosive joint destruction.

Ankylosing spondylitis

Ankylosing spondylitis (AS) is a chronic inflammatory disease that predominantly affects the axial skeleton, with variable involvement of the peripheral joints and nonarticular structures, such as the heart valves and the eyes¹³⁵. HLA-B27 has been recognized as the most important genetic risk factor for AS, but genome-wide association studies (GWAS) have now identified two AS-associated chromosomal loci independent of HLA-B27 that encode four endoplasmic reticulum aminopeptidases (ERAPs), which are involved in peptide processing before MHC class I presentation^136–138. Although MHC class I peptides are primarily presented to CD8 T cells, IL-12⁺ and IL-23⁺ CD4 T cells (which can be converted to T_H17 cells), have been also implicated as responders to antigen-presenting macrophages and DCs in the inflamed spine of patients with AS¹³⁹.

The upregulation of let-7i miRNA, which contributes to IFNγ production, might also account for T-cell hyper-reactivity in patients with AS¹⁴⁰, as insulin-like growth factor-1 receptor (IGF-1R) is a direct target of let-7i in T cells from these patients¹⁴¹. IGF-1R mediates its effect by activation of the PI3K–AKT–mTOR path-way^140–143. Interestingly, let-7i overexpression increases autophagy, but inhibits apoptosis; thus, let-7i contributes to T-cell dysfunction in patients with AS¹⁴¹. Mo r e o v e r, unbiased proteomics studies found evidence for activation of PI3K, fatty-acid oxidation and insulin signalling in fibroblast-like ligament cells of patients with AS¹⁴⁴. Given all these findings, mTOR-controlled metabolic pathways are likely to shape the repertoire of both adaptive and innate inflammatory cells in AS (FIG. 3).

Psoriatic arthritis

T_H17 cells have pivotal roles in orchestrating the inflammation in the skin and joints of patients with psoriatic arthritis (PsA)¹⁴⁵. Regarding the prominent expansion of T_H17 cells, a paucity of studies have examined its association with mTOR activation in PsA or psoriasis. Importantly, GWAS in patients with psoriasis (20–30% of whom also have PsA) uncovered a link with a single-nucleotide polymorphism in Raptor, a component of mTORC1 (REF. 146). Whereas the functional consequences of this genetic association have not been evaluated, the involvement of mTORC1 in PsA is independently supported by mechanistic studies. As is also observed in patients with RA, the proliferation of FLSs (which can be triggered by IL-22) is mediated by the PI3K–Akt–mTOR cascade in patients with PsA¹⁴⁷ (FIG. 3). Unlike RA T cells, however, PsA T cells exhibit increased signalling through the AKT– mTOR axis¹⁴⁸; this observation may result from increased expression of programmed death-1 (PD-1) in T cells from patients with PsA¹⁴⁸.

The mTOR pathway has also been implicated in keratinocyte hyperproliferation (FIG. 3), as rapamycin was found to arrest human keratinocyte stem cells in the G1 phase¹⁴⁹. Evidence of mTORC1 activation is also provided by the increased phosphorylation of mTOR and S6K in punch biopsy samples of psoriatic skin lesions^150,151.

Systemic sclerosis

Activation of the mTOR pathway has a central role in the proliferation of fibroblasts (FIG. 3), the cells that mediate skin lesions in the limited cutaneous form of SSc and in other involved organs in progressive SSc. Signalling through the PI3K–AKT–mTOR axis can be initiated by transforming growth factor β (TGFβ) and its proximal effectors, Smad2 and Smad3 (REFS 152,153). Supporting a critical role for mTOR in these diseases, rapamycin abrogates TGFβ-induced fibroblast growth¹⁵⁴. As mentioned previously, mTORC1 activation also contributes to type I collagen production by fibroblasts, the process that leads to fibrosis in SSc skin lesions^11,12 and is inhibited by rapamycin¹³. In a pilot study of patients with progressive SSc, rapamycin was safe, although it had limited efficacy¹⁵. TGFβ also activates mTORC2 in skin fibroblasts, as evidenced by increased phosphorylation of AKT at Ser473¹⁵⁴, and mTORC2 activity also contributes to kidney fibrosis¹⁵⁵.

Osteoarthritis

The pathogenesis of OA originates from excessive mechanical stress and a systemic susceptibility¹⁵⁶. The disease process is characterized by progressive cartilage loss, subchondral bone remodelling, osteophyte formation and synovial inflammation. Notably, cartilage loss might be therapeutically targeted via mTOR blockade. Mechanical stress activates the mTOR pathway in chondrocytes during chondrogenesis and cartilage development¹⁵⁷ (FIG. 3). In turn, treatment with rapamycin blocks the mechanical-stress-induced proliferation of chondrocytes in vitro¹⁵⁷. Thus, blockade of mTORC1 in chondrocytes might not itself be beneficial, but human articular chondrocytes also employ the mTOR pathway to control autophagy, which initiates the release of articular cartilage vesicles (ACVs) into the cartilage matrix²⁰. Interestingly, rapamycin promotes autophagy and consequently increases the release of ACVs by normal chondrocytes, whereas OA chondrocytes fail to increase ACV release in response to mTOR inhibition. These findings indicate that the autophagy defects in OA are downstream of mTOR activation²⁰. Nevertheless, intervening at a critical step in the regulation of autophagy could potentially slow the adverse effects of ageing on cartilage homeostasis, which increase the risk of developing OA¹⁵⁸. Indeed, cartilage-specific inactivation of mTOR enhances autophagy and protects against meniscus-ablation-induced OA; treated animals show a substantial reduction in articular cartilage degradation, apoptosis and synovial fibrosis¹⁵⁹. This substantial effect of rapamycin treatment on OA prevention is remarkable, given that it also extends lifespan in mice¹⁶⁰. Moreover, the inhibition of OA by rapamycin might also improve the quality of life of patients with this disease.

Rapamycin, rapalogues and mTORC1/2 inhibitors

Rapamycin is an allosteric inhibitor of mTOR that acts by forming a high-affinity complex with the 12 kD intracellular protein FKBP12 (peptidylprolyl cis-trans isomerase FKBP1B)¹⁶⁸ (FIG. 1). The resulting complex of rapamycin and FKBP12 blocks mTOR activation in yeast¹⁶⁹ and mammalian cells¹⁶¹. mTOR blockade by rapamycin seems to primarily affect mTORC1 (REFS 170,171), and might cause a secondary activation of mTORC2 (REF. 82).

The duration of treatment with rapamycin is critical for mTORC1 selectivity. Whereas 2 weeks of treatment in mice blocks only mTORC1, leading to detrimental metabolic effects, 6 weeks of treatment leads to a state of metabolic transition, and 20 weeks of treatment improves metabolic profiles and insulin sensitivity by also reducing activity of mTORC2 (REF. 172). In T cells from healthy individuals and patients with SLE, rapamycin treatment in vitro reduced mTORC1 and enhanced mTORC2 activity¹⁰⁶. Although the activity of mTORC2 is reduced in DN T cells from these patients, as evidenced by diminished phosphorylation of AKT at Ser473¹⁰⁶, whether long-term treatment with rapamycin also activates mTORC2 in vivo is currently unknown. Therefore, whether the therapeutic benefit of rapamycin in patients with SLE originates solely from blockade of mTORC1, or whether it also involves activation of mTORC2, is also unclear. Given the role of mTORC2 in control of T_REG-cell survival and function^88–90, studies should be pursued in patients with SLE and other autoimmune diseases to dissect the exact mechanism of action of rapamycin. In the event that mTORC2 is activated by rapamycin in vivo, ATP-competitive, dual mTORC1 and mTORC2 inhibitors (such as torin 1 or sapanisertib) could be even more effective¹⁷³. Importantly, the ATP-competitive dual inhibitor torkinib reduces the expression of cholesterol biosynthesis genes in a 4E–BP1-dependent manner¹⁷⁴. Dual inhibitors are currently in clinical trials in cancer patients.

Rapamycin has shown great potency in blocking the development of SLE in the MRL/lpr⁵ and NZB/W F1 animal models^41,175, in which autoimmunity, T-cell hyper-reactivity, titres of ANA and anti-DNA antibodies, glomerular immunoglobulin deposition, proteinuria, nephritis, and overall survival were all improved. Importantly, rapamycin also attenuates the severity of established nephritis¹⁷⁶. In patients with SLE, rapamycin improved clinical outcomes as measured by the reduction of overall SLE Disease Activity Index (SLEDAI) and BILAG (British Isles Lupus Assessment Group) scores^6,109. More over, a retrospective study showed diminished protein uria and improved renal function in six patients with lupus nephritis treated with rapamycin¹⁷⁷. A prospective open-label clinical trial of rapamycin in 40 patients with SLE is due for completion in 2015, and should provide additional information about the organ systems that respond to mTOR blockade¹⁷⁸.

Beyond its effects on T cells in SLE, rapamycin blocks fibroblast proliferation, collagen production and dermal fibrosis in the skin^11,12 and holds promise for the treatment of patients with SSc¹⁵ (TABLE 1). Interestingly, mTORC2 mediates kidney fibrosis¹⁵⁵, which could explain the limit ed efficacy of mTORC1 blockade by rapamycin in this setting, and suggests that dual ATP-competitive mTOR inhibitors might show improved efficacy (TABLE 1). As noted previously in this Review, rapamycin treatment enhances the survival of human OA chondrocytes in vitro, whereas genetic inactivation of mTOR prevents OA in mice^55,159. These studies suggest that mTOR blockade and stimulation of autophagy in chondrocytes could have therapeutic benefits in OA. Conversely, rapamycin showed limited efficacy in a controlled study of patients with severe psoriasis¹⁷⁹. The newer agent everolimus provided clinical benefit for patients with RA when used in combination with methotrexate, and had an acceptable safety and tolerability profile⁸. Thus, everolimus might offer a new treatment option in patients with RA who have an inadequate response to methotrexate⁸.

PI3K inhibitors

The PI3K–AKT–mTOR pathway is activated in CD4 T cells from mice with lupus-like symptoms induced by graft-versus-host disease¹⁸⁰, or as a consequence of the MRL/lpr genetic background¹⁸¹, and in those collected from patients with SLE¹⁸². Accordingly, pharmacological blockade of PI3Kγ with AS605240 blocks T-cell activation, autoantibody production, renal infiltration and TNF release by macrophages, and ameliorates glomerulonephritis, thus extending lifespan in mouse models of lupus^181,183. Additionally, the broad PI3K inhibitor LY294002 improves the survival of mesenchymal stem cells (MSCs), which exhibit enhanced senescence in patients with SLE¹⁸⁴. PI3Kδ is highly expressed in the synovium and cultured FLS of patients with RA¹⁸⁵, and its expression is selectively induced over other PI3K isoforms by TNF. Moreover, a novel PI3Kδ inhibitor, INK007, blocks TNF-induced AKT activation¹⁸⁵ (TABLE 1).

N-acetylcysteine

The therapeutic use of NAC in SLE has been based on increasing evidence of oxidative stress being involved in the pathogenesis of this disease¹⁸⁶, including effects on abnormal T-cell activation^187,188, the antigenicity of DNA¹⁸⁹ and self proteins¹⁹⁰, and production of ANA¹⁹¹ (TABLE 1). The rationale for reversing oxidative stress with NAC is supported by several observations. Intracellular GSH levels are decreased in PBLs and T cells from patients with SLE^30,192–195, and GSH depletion and oxidative stress contribute to T-cell dysfunction in these patients^189,192,196. Additionally, NAC and GSH-sparing antioxidants improve the clinical outcome of lupus in mice^197–199, and the fact that large dosages (up to 8 g daily) of NAC can be safely administered to humans²⁰⁰ (and that similar dosages of NAC improve fatigue in healthy individuals²⁰¹) further supports the use of NAC in this context. NAC also limits the myelotoxicity of immunosuppressive medications²⁰², which are also commonly used in patients with SLE.

NAC is inexpensive and widely accessible to patients in health-food stores, but is currently unavailable as an oral medication by prescription. Therefore, we initiated a randomized, double-blind, placebo-controlled phase I–II study to evaluate the safety and tolerability of NAC, as well as its metabolic, immunological and therapeutic effects, in 36 patients with SLE. The results of this study indicate that NAC is safe and reduces disease activity and fatigue over 3 months⁷⁰. The therapeutic action of NAC is mediated by the reversal of GSH depletion and blockade of mTORC1 (REF. 70), which is a sensor of oxidative stress^{28–30,33,79}. The clinical efficacy of NAC in patients with SLE is hypothesized to occur through a newly identified molecular mechanism: disruption of the mitochondrial hyperpolarization (MHP)–oxidative stress–mTOR axis in T cells¹⁸⁶. DN T cells exhibit the greatest mTORC1 activation^70,109 and are predisposed to proinflammatory necrotic cell death^70,109; consequently, along with the blockade of their mTORC1 activity, proliferation of these cells is reversed by NAC treatment⁷⁰. Thus, elimination of DN T cells can contribute to the efficacy of NAC.

The use of NAC instead of rapamycin to block mTOR might have additional benefits. Rapamycin induces hyperlipidaemia, at least in renal transplant recipients²⁰³, and infections are common adverse events in these patients, with pneumonia and sepsis being most frequent²⁰⁴. These unwanted consequences of rapamycin are of concern, because infections and cardiovascular disease are the leading causes of mortality in patients with SLE²⁰⁵. Thus, given that oxidative stress contributes to cardiovascular disease²⁰⁶, patients might have additional benefit from treatment with NAC instead of rapamycin, at least those with end-stage renal disease²⁰⁷. Future studies should focus on whether GSH depletion and activation of mTORC1 predict responsiveness to treatment by NAC and mTOR inhibitors, and whether selective or non-selective mTORC1, mTORC2 or dual mTORC1/mTORC2 inhibitors deliver differential clinical benefits (TABLE 1).

Targets under development

Metformin activates AMPK and, thus, inhibits mTORC1 (REF. 208; TABLE 1). Metformin also reduces mTORC1 activity by moderating mitochondrial oxidative stress, which has been implicated in the prevention of nephritis in lupus-prone mice¹²⁸, and inhibits inflammatory arthritis in the K/B × N model of lupus²⁰⁹. Beyond inhibiting mTORC1-dependent inflammation, metformin might also reduce metabolic comorbidities such as hyperlipidaemia and atherosclerosis in patients with SLE or RA²¹⁰.

Fingolimod, a S1P–receptor antagonist with effects upstream of mTORC1 (REFS 211,212), has been introduced for the treatment of multiple sclerosis²¹³. Fingolimod sequesters lymphocytes in the secondary lymphoid organs by inducing S1P receptor internalization and degradation²¹³, and blocks the expansion of DN T cells and prevents nephritis in MRL/lpr mice^214,215. Given their potentially synergistic mechanism of action, fingolimod, metformin, NAC and rapamycin might be combined for the treatment of these autoimmune diseases.

Among the pharmacological interventions currently in preclinical development for the treatment of rheumatic diseases, targeting of calcium/calmodulin-dependent protein kinase type IV (CaMK-IV) seems to be promising (TABLE 1). Overexpression of CaMK-IV in T cells from lupus-prone mice²¹⁶ might act upstream of mTORC1; but given that rapamycin treatment in vivo restores T-cell activation-induced Ca2+ fluxing⁶, CaMK-IV might also act downstream of mTORC1. Similarly to mTORC1 blockade, pharmacological inhibition of CaMK-IV with KN-93 expanded TREG cells²¹⁷, decreased the expression of the co-stimulatory molecules CD86 and CD80 on B cells, decreased ANA production and suppressed nephritis and skin disease in MRL/lpr mice²¹⁸. KN-93 also inhibited CaMK-II in vitro, and blocked the proliferation of FLSs from patients with RA²¹⁹. Another CaMK-II inhibitor, rimacalib, which has anti-inflammatory effects on macrophages²²⁰, has entered clinical trials for the treatment of RA (TABLE 1).