PI3Ks are heterodimeric enzymes composed of a regulatory subunit and a catalytic (p110) subunit. They are classified into different classes, among which class IA PI3Ks (p110α, p110β and p110δ) are primarily activated by receptor tyrosine kinases upon stimulation by growth factors and cytokines, whereas class IB PI3K (p110γ) is activated by G-protein-coupled receptors (GPCRs), such as chemokine receptors (22,23). These general activation mechanisms have been well established in various cell types; in the context of RA, they are mainly supported by studies involving RA-derived FLSs and immune cells.

A key feature of PI3K activation is its potentiation by Ras family GTPases. Once activated, PI3K phosphorylates phosphatidylinositol lipids at the plasma membrane, thereby generating docking sites for proteins containing pleckstrin homology domains, most notably phosphoinositide-dependent kinase 1 (PDK1) and AKT (24). The tumor suppressor PTEN serves as the primary negative regulator of this pathway by dephosphorylating these lipid intermediates (25). AKT functions as the central node of PI3K signaling, and its phosphorylation by PDK1 and mammalian target of rapamycin complex (mTORC)2 results in full activation. Downstream, AKT coordinates cellular responses by phosphorylating and modulating the activity of numerous substrates, including inhibitor of κB kinase (IKK), the pro-apoptotic protein Bad, glycogen synthase kinase-3β (GSK-3β), mTOR and forkhead box O (FoxO) transcription factors (26,27).

There is evidence for constitutive activation of the PI3K/AKT axis in the RA synovial microenvironment (15). One line of evidence comes from the reduced expression of PTEN mRNA in the invasive layer of the RA synovium, a specific sublayer within the synovial lining, as well as in FLSs invading cartilage in a murine arthritis model (28). Loss of this negative regulator supports hyperactivation of the pathway. Consistent with this, levels of phosphorylated-AKT are markedly increased in RA synovial tissue and purified FLSs compared with specimens from patients with osteoarthritis or healthy controls (29). Functionally, pro-inflammatory cytokines, such as TNF-α, potently activate AKT in RA-FLSs in vitro. Inhibition of PI3K by LY294002 or overexpression of PTEN sensitizes these cells to TNF-α-induced apoptosis, demonstrating that the PI3K/AKT pathway provides a survival signal that contributes to the apoptosis-resistant phenotype of RA-FLSs (30,31). Moreover, this pathway mediates resistance to cell death induced by other cytokines, such as TNF-related apoptosis-inducing ligand (32). These findings help explain the dual role of the PI3K/AKT pathway in promoting both synovial hyperplasia and inflammation. Most of these observations are derived from direct studies of RA tissues and FLSs (28–32), thus providing disease-specific evidence. The mechanistic details regarding Ras-mediated PI3K activation are derived mainly from studies non-RA systems (24,33). As direct validation in RA tissues remains limited, these findings should be interpreted as general PI3K/AKT pathway biology rather than RA-specific evidence. Fig. 1 presents a comprehensive schematic diagram of the PI3K/AKT signaling pathway in the pathogenesis of RA.

Once activated, AKT phosphorylates a series of transcription factors and regulatory proteins that modulate gene expression profiles in RA synovial cells. Among these, the NF-κB pathway represents one of the most important downstream effectors. Phosphorylation and activation of IKK lead to the phosphorylation of IκB, resulting in its degradation and the subsequent nuclear translocation of NF-κB (34,35). This, in turn, enhances the transcription of pro-inflammatory genes, including TNF-α, IL-1β, IL-6 and IL-8, as well as MMPs, such as MMP-3 and MMP-9, which have critical roles in synovitis and cartilage degradation (36,37). Moreover, AKT-mediated phosphorylation of FoxO transcription factors leads to their cytoplasmic sequestration and inactivation, thereby suppressing the expression of pro-apoptotic genes such as Bim and FasL (Fig. 2). This mechanism contributes to the apoptosis-resistant phenotype of FLSs in RA (38,39). In addition, mTOR is another major downstream substrate of AKT. Activation of mTORC1 positively regulates mRNA translation and the expression of proteins involved in cell proliferation and metabolism through phosphorylation of ribosomal protein S6 kinase 1 and eukaryotic translation initiation factor 4E-binding protein 1 (40–42). This promotes the synthesis of cyclins, MMPs and inflammatory mediators, thereby enhancing synovial proliferation and inflammation. All of the aforementioned mechanisms have been directly demonstrated in RA-FLSs or RA synovial tissues, representing strong disease-specific evidence.

In addition to its role in transcriptional control, the PI3K/AKT pathway can influence gene expression through post-transcriptional and epigenetic mechanisms. AKT activation may stabilize specific mRNAs by inhibiting GSK-3β (Fig. 2), which is considered to mediate mRNA decay (43). In addition, PI3K/AKT signaling can induce epigenetic changes, such as histone acetylation and DNA methylation, thereby remodeling chromatin and altering gene expression in RA-FLSs (Fig. 2). For example, AKT activation has been associated with increased histone H3 acetylation at the promoters of inflammatory genes, thereby facilitating their transcription (44,45). These epigenetic observations are primarily based on in vitro studies of RA-FLSs; however, the direct link between AKT and specific histone modifications in RA synovium in vivo remains less well established and warrants further investigation. Fig. 2 presents a schematic diagram of the transcriptional and epigenetic regulation mediated by the PI3K/AKT signaling pathway.

A critical layer of regulation of PI3K/AKT signaling is provided by ncRNAs. Multiple microRNAs (miRNAs/miRs) have been shown both to regulate and to be regulated by PI3K/AKT signaling. In particular, miR-124a directly targets the p110α catalytic subunit of PI3K (encoded by PIK3CA), thereby inhibiting AKT activation and downstream NF-κB signaling (Fig. 3). This leads to suppression of RA-FLS proliferation and reduced production of inflammatory cytokines (19). By contrast, miR-21 is upregulated in the RA synovium and promotes PI3K/AKT activation by suppressing PTEN (Fig. 3), thereby enhancing FLS survival and inflammation (46,47).

Notably, long ncRNAs (lncRNAs) are key modulators of the PI3K/AKT pathway in RA. For example, the lncRNA THRIL is upregulated in the serum of patients with RA, and knockdown of THRIL in TNF-α-stimulated RA-FLSs reduces the phosphorylation of PI3K and AKT, concomitantly decreases the production of IL-1β and MMP-3 and promotes apoptosis (48) (Fig. 3). These findings suggest that THRIL exerts its pro-inflammatory and anti-apoptotic effects, at least in part, through activation of the PI3K/AKT pathway. Similarly, the lncRNA H19 has been shown to promote AKT phosphorylation by serving as a competing endogenous RNA (ceRNA) for miRNAs that target AKT regulators; for example, H19 acts as a ceRNA to modulate the PI3K/AKT pathway (49,50), thereby promoting inflammatory responses (Fig. 3).

Circular RNAs (circRNAs) also participate in this regulatory network. For example, circRNA_09505 sponges miR-6089 (Fig. 3), thereby relieving its suppression of AKT1 mRNA, which leads to enhanced AKT signaling and aggravated inflammation in RA macrophages (51).

All ncRNA examples discussed in the present review are derived from samples of patients with RA or RA-FLS models, thus representing direct evidence in RA. Fig. 3 presents a schematic diagram of the crosstalk between the PI3K/AKT signaling pathway and ncRNAs.

The interaction between PI3K/AKT and NF-κB is arguably the most understood among these signaling networks. AKT directly phosphorylates and activates IKK, thereby promoting IκB degradation and the nuclear translocation of NF-κB. In this way, it facilitates the increased expression of various pro-inflammatory genes, including TNF-α, IL-1β, IL-6 and MMPs (52,53). This mechanism has been repeatedly validated in RA-FLSs and animal models of RA, providing strong direct evidence. Conversely, NF-κB can further enhance PI3K/AKT signaling by upregulating cytokines such as IL-6, which promote the activation of upstream receptor tyrosine kinases and GPCRs, thereby creating a positive feedback loop that sustains chronic inflammation and drives RA progression (54).

The JAK/STAT pathway represents another important inflammatory signaling pathway in RA. Cytokine receptors associated with JAKs can also recruit and activate PI3K through phosphorylated tyrosine residues, thereby leading to downstream AKT activation (55,56). This interaction has been demonstrated in RA-FLSs and synovial tissues. In addition, STAT3, the principal transducer of IL-6 signaling, transcriptionally upregulates genes involved in cell survival and inflammation that are also modulated by AKT, such as Bcl-2 and MMP-9 (57,58). This overlap suggests that combined inhibition of the JAK/STAT and PI3K/AKT pathways may produce synergistic therapeutic effects, as indicated by preclinical and clinical studies of dual-pathway inhibition (59,60).

In RA, the MAPK pathway, including ERK and p38, frequently exhibits cross-activation with PI3K/AKT signaling. For example, AKT antagonizes apoptosis signal-regulating kinase 1 (ASK1), an upstream activator of p38 and JNK, thereby influencing stress-induced apoptosis and inflammation (61,62). However, much of the evidence for the interaction between AKT and ASK1 is derived from non-RA contexts, such as cancer and neurotoxicity models (61,62), and direct validation in RA remains limited. Meanwhile, MAPK activation can also influence PI3K signaling through ribosomal S6 kinase-mediated phosphorylation of AKT or by increasing the expression of growth factors that activate PI3K (63,64). This bidirectional crosstalk further underscores that targeting a single pathway in isolation may be insufficient, and that multi-target therapeutic strategies may be required.

Evidence indicates that Notch signaling modulates PI3K/AKT activity in RA. The Notch intracellular domain (NICD) can transcriptionally upregulate PIK3CA expression or inhibit PTEN, thereby enhancing AKT phosphorylation (65,66). These findings are supported by studies in RA synovial fibroblasts and animal models (65,66). In turn, AKT can stabilize NICD and enhance its transcriptional activity, thereby forming another positive feedback loop that promotes synovial cell survival and inflammatory cytokine production (67). Inhibition of Notch signaling has been shown to reduce AKT activation and ameliorate arthritis in animal models, suggesting that this pathway may represent a viable therapeutic target (68). Notch signaling via the NICD/CSL/Hey/DTX axis also promotes the invasive migration of RA-FLSs, a key feature of synovial hyperplasia and joint destruction (Fig. 4).

The P2X7R/NOD-, LRR- and pyrin domain-containing protein 3 (NLRP3) inflammasome axis also intersects with PI3K/AKT signaling. ATP-mediated activation of P2X7R can stimulate PI3K/AKT signaling, which in turn enhances NLRP3 inflammasome assembly and IL-1β maturation (69,70). These observations are primarily derived from non-RA models, such as brain injury (69) and a colorectal cancer model (70); however, recent studies in RA have begun to explore this link (71,72). Moreover, AKT-mediated activation of NF-κB primes NLRP3 expression, thereby completing a pro-inflammatory circuit that links purinergic signaling with cytokine release and synovitis (73,74). Targeting P2X7R or NLRP3 may therefore indirectly attenuate PI3K/AKT signaling. However, direct evidence for this crosstalk in RA remains preliminary and requires further investigation. Additionally, the PI3K/AKT pathway may contribute to apoptosis resistance and inflammation through its interplay with DNA damage response proteins. For example, DTX3L (also known as ARTD9), which is involved in the DNA damage response, has been shown to promote inflammation in RA-FLSs by increasing STAT1 translocation (75). This crosstalk represents a promising area for future investigation into the mechanisms driving the anti-apoptotic and pro-inflammatory phenotype of RA-FLSs (Fig. 4).

Several small-molecule inhibitors targeting PI3K or AKT have shown promise in preclinical models of RA. For example, the PI3K inhibitor LY294002 and the AKT inhibitor MK-2206 have been shown to induce apoptosis in RA-FLSs, and reduce the production of pro-inflammatory cytokines and MMPs in vitro (76,77). Isoform-specific inhibitors, particularly those targeting PI3Kδ and PI3Kγ, are of particular interest because of their predominant expression in leukocytes and their roles in immune cell activation. Idelalisib, a PI3Kδ inhibitor approved for hematological malignancies, has shown efficacy in reducing disease severity by attenuating B-cell and macrophage activation (78,79). Similarly, duvelisib, a dual PI3Kδ/γ inhibitor, has been reported to suppress synovitis and bone erosion in animal models of RA by impairing neutrophil and macrophage migration into the joints (80).

Moreover, mTOR inhibitors such as rapamycin (sirolimus) and everolimus, which target a downstream effector of AKT, have also been evaluated in RA (81,82). Rapamycin has been reported to reduce synovial hyperplasia and the production of inflammatory cytokines in a review, suggesting that it may serve as a potential therapeutic target for RA (83). However, the clinical translation of these agents has been limited by systemic toxicity and off-target effects, highlighting the need for more selective and tissue-specific therapeutic strategies.

The regulatory role of ncRNAs in the PI3K/AKT pathway offers options for therapeutic intervention. For example, miR-124a, which targets PIK3CA, is downregulated in the RA synovium. Restoration of miR-124a expression through synthetic mimics or nanoparticle-based delivery systems has been shown to suppress PI3K/AKT/NF-κB signaling, inhibit FLS proliferation, and reduce inflammatory cytokine production in vitro and in CIA models (84). Similarly, a clinical study showed that miR-21 was markedly positively associated with RA disease activity, indicating its potential value as a biomarker for future RA intervention (85). Targeting miR-21 may regulate the PI3K/AKT signaling pathway and slow RA progression (86). Qu et al (87) reported that overexpression of miR-126 in RA-FLSs inhibited PIK3R2 expression, while promoting proliferation and suppressing apoptosis. This finding suggests that inhibition of miR-126 may downregulate PI3K/AKT signaling and provide therapeutic benefits in RA.

lncRNAs and circRNAs also represent promising therapeutic targets. Silencing of lncRNA H19, which promotes AKT activation, using small interfering RNA (siRNA)- or CRISPR-based approaches has been shown to reduce inflammation and induce apoptosis in RA-FLSs (88). Wang and Liu (89) demonstrated through integrated computational and experimental analyses that regulation of the lncRNA DSCR9/RPLP2/PI3K/AKT axis may represent an important mechanism by which Xinfeng capsule improves the inflammatory response and hypercoagulable state in RA, and that lncRNA DSCR9 may serve as a potential therapeutic target. Liu et al (90) demonstrated, using an RA-FLS-induced human umbilical vein endothelial cell model, that lncRNA HOTAIR, as a potential therapeutic target in RA, can activate the PI3K/AKT pathway through the miR-126-3p/PIK3R2 regulatory axis, thereby promoting angiogenesis in RA. Likewise, circRNA_09505, which sponges miR-6089 and enhances AKT1 expression, promotes the production of TNF-α, IL-6 and IL-12 through a ceRNA mechanism, whereas knockdown of circRNA_09505 was shown to notably alleviate arthritis and inflammation in a CIA mouse model (51). These findings highlight the potential of RNA-based therapeutics to precisely modulate the PI3K/AKT axis, potentially with fewer off-target effects than broad-spectrum kinase inhibitors.

Due to the extensive crosstalk between PI3K/AKT and other signaling pathways, including NF-κB, JAK/STAT and MAPK, combination therapies targeting these pathways may achieve synergistic effects. For example, the co-administration of a PI3Kδ inhibitor and a JAK inhibitor (for example, tofacitinib) has been shown to suppress inflammatory cytokine production and synovial hyperplasia more effectively than either agent alone (91). Likewise, combined inhibition of PI3K and mTOR with agents such as dactolisib (BEZ235) has demonstrated greater efficacy in suppressing inflammatory activity (92).

Natural compounds and repurposed drugs also offer multi-target potential. Curcumin, for example, has been shown to inhibit both the PI3K/AKT and NF-κB pathways, and its combination with methotrexate enhances anti-inflammatory effects in RA models (93,94). Metformin, an AMPK activator, indirectly suppresses PI3K/AKT signaling and has shown beneficial effects in decreasing disease activity in patients with RA (95,96). Wan et al (97) combined in vitro and in vivo experiments to demonstrate that triptolide can downregulate the expression of factors secreted by M1 macrophages, and inhibit the NF-κB, PI3K/AKT and p38 MAPK signaling pathways, thereby ameliorating immune imbalance, joint inflammation and tissue damage in RA. Table I summarizes the therapeutic strategies targeting the PI3K/AKT pathway in RA.