Targeting signaling pathways in lymphoma: From molecular mechanisms to clinical breakthroughs.

Introduction
Lymphoma, a heterogeneous group of hematologic malignancies originating from the lymphatic system, poses a significant global health burden. The World Health Organization classification delineates over 100 distinct subtypes, broadly categorized into Hodgkin lymphoma (HL) and non-HL (NHL), with the latter comprising a vast spectrum of B-cell, T-cell, and natural killer (NK)-cell neoplasms.[1] This pathological diversity is mirrored by profound molecular complexity, driven by a complex interplay of genetic mutations, chromosomal translocations, epigenetic dysregulation, and alterations from the tumor microenvironment (TME).[2] Consequently, lymphomas exhibit varied clinical behaviors, ranging from indolent diseases that can be managed with a “watch and wait” approach to highly aggressive subtypes requiring immediate and intensive intervention.[3] This inherent heterogeneity presents a formidable challenge in developing universally effective treatments.[4]
For decades, the standard of care for most lymphomas has been combination chemotherapy, often integrated with anti-CD20 monoclonal antibodies for B-cell lymphoma, exemplified by the rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone regimen (R-CHOP).[5] This immunochemotherapy approach has substantially improved survival rates and remains a curative standard for many aggressive B-cell lymphomas. However, its limitations exist.[6] In the realm of T-cell and NK/T-cell lymphomas (NKTCL), which are often characterized by greater biological heterogeneity and inherent chemoresistance, outcomes with conventional chemotherapy are frequently suboptimal.[7] Across both B-cell and T-cell lineages, a significant proportion of patients either fail to respond to initial therapy (refractory disease) or experience relapse.[8] Furthermore, the substantial toxicity associated with cytotoxic agents often precludes their use in elderly or frail patients, a growing demographic in the lymphoma population. These persistent unmet clinical needs have created an urgent need for the development of more effective, better-tolerated, and rationally designed therapeutic strategies that address the distinct biology of each lymphoma subtype.[9]
The quest for such strategies has been revolutionized by the advent of high-throughput multiomics technologies, including next-generation sequencing, transcriptomics, and proteomics. These powerful tools have provided a better understanding of lymphoma pathogenesis, providing insights that transcend traditional histopathology by revealing the intricate molecular landscape.[10] This has led to the development of sophisticated molecular classifications, such as the genetic subtypes and the characterization of the TME in lymphomas.[111213] These new classification systems have illuminated a critical dependency of malignant cells on aberrantly activated intracellular signaling pathways.[14] Key pathways governing cell survival, proliferation, and immune evasion, such as the B-cell receptor (BCR), the intrinsic apoptosis pathway, the phosphatidylinositol 3-kinase (PI3K)/protein kinase B (AKT)/mammalian target of rapamycin (mTOR) network, etc., are now understood to be constitutively active in specific molecularly defined subgroups. This refined understanding has provided the rationale for moving beyond indiscriminate cytotoxic agents toward an era of precision medicine. Therapies designed to specifically inhibit key nodes within these dysregulated cascades have thus emerged as a revolutionary approach, offering the potential for high efficacy in biomarker-selected patient populations.[15]
This review will comprehensively survey the landscape of targeted therapies that have reshaped lymphoma treatment. We will begin by dissecting the molecular architecture and therapeutic vulnerabilities of key oncogenic signaling pathways. For each of these areas, we will cover the journey from mechanism to clinical validation, including strategies to overcome resistance. Following this, we will synthesize how these agents are being integrated into clinical practice, discussing the rise of chemotherapy-free regimens, the critical role of biomarkers, and the promise of genetic subtype-guided combination therapies. Finally, we will address the current challenges and future directions, focusing on mechanisms of acquired resistance, rational combination strategies, and the potential of emerging modalities to pave the path toward precision cures for lymphoma.

Key Signaling Pathways and The Evolution of Targeted Agents in Lymphoma

The BCR signaling pathway

Molecular architecture and rationale for targeting the BCR pathway
The BCR signaling pathway is a key oncogenic driver in many B-cell malignancies, including diffuse large B-cell lymphoma (DLBCL), mantle cell lymphoma (MCL), and chronic lymphocytic leukemia/small lymphocytic lymphoma (CLL/SLL). The pivotal role of the BCR signaling pathway in B-cell biology was first highlighted through the study of X-linked agammaglobulinemia, a primary immunodeficiency caused by mutations in the bruton tyrosine kinase (BTK) gene.[16] In patients with X-linked agammaglobulinemia, these mutations impair B-cell development in the bone marrow and lead to a profound deficiency of serum immunoglobulins. This disease model was instrumental in establishing the BCR signaling pathway as indispensable for B-cell maturation and function.[17] The BCR complex is composed of a membrane surface immunoglobulin for antigen recognition and the signal-transducing heterodimer Igα/Igβ (CD79a/CD79b) [Figure 1]. The cytoplasmic tails of Igα/Igβ contain immunoreceptor tyrosine-based activation motifs (ITAMs), which are essential for initiating the downstream cascade.[18] Upon antigen binding, BCR clustering triggers the recruitment and activation of Src-family kinases (e.g., LYN), which phosphorylate the ITAM tyrosines. These phosphotyrosine sites serve as docking platforms for spleen tyrosine kinase (SYK). Subsequent phosphorylation and activation of SYK by Src-family kinases or through autophosphorylation establishes SYK as a central signaling hub.[19] Crucially, activated SYK then phosphorylates and activates key adaptor proteins, such as B-cell linker protein, which in turn recruits BTK to the plasma membrane. Once recruited, BTK is phosphorylated and fully activated, serving as a critical downstream signal transducer. Activated BTK propagates signals through multiple effector pathways, most notably by activating phospholipase C gamma 2, which leads to the activation of the nuclear factor-kappa B (NF-κB) and mitogenactivated protein kinase pathways. This cascade ultimately controls B-cell proliferation, differentiation, and survival.[20] In B-cell malignancies, this entire pathway is often constitutively activated through several mechanisms, mainly including chronic antigen drive (e.g., in gastric mucosa-associated lymphoid tissue [MALT] lymphoma and splenic marginal zone lymphoma),[21] cell-autonomous signaling (e.g., in follicular lymphoma [FL]),[22] and genetic mutations that hardwire the pathway in an “on” state (e.g., CD79B ITAM or CARD11 mutations in activated B-cell like DLBCL, or the API2-MALT1 fusion in MALT lymphoma).[23] This critical dependency on the BCR-BTK signaling axis renders it a prime therapeutic vulnerability in these diseases.

Clinical development of BCR pathway inhibitors
The validation of the BCR–BTK axis as a therapeutic vulnerability catalyzed the development of BTK inhibitors, which have revolutionized the treatment of B-cell malignancies. The first-generation covalent inhibitor, ibrutinib, was a trailblazer.[24] The RESONATE-2 trial demonstrated the superiority of ibrutinib over chlorambucil in treatment-naïve older patients with CLL/SLL, showing a significant progress in efficacy [Table 1].[25] Similarly, in relapsed/refractory (R/R) MCL, ibrutinib monotherapy also showed durable efficacy [Table 1].[26]
Building on this success, second-generation covalent BTK inhibitors,[27] such as acalabrutinib, zanubrutinib, and orelabrutinib, were developed with improved kinase selectivity to minimize off-target toxicities. The phase III ELEVATE-TN trial confirmed that acalabrutinib, as monotherapy or in combination with anti-CD20 antibody obinutuzumab, demonstrated superior efficacy compared to the immunochemotherapy regimen of obinutuzumab plus chlorambucil for treatment-naïve CLL, with a manageable safety profile [Table 1].[28] Likewise, in the phase III ASPEN trial, which compared zanubrutinib to ibrutinib in patients with Waldenstrom macroglobulinemia (WM), zanubrutinib showed a trend toward a higher rate of complete response rate (CRR) or very good partial response, although the difference did not reach statistical significance for the primary endpoint (28% vs. 19%; P = 0.09) [Table 1].[29] Orelabrutinib, another highly selective second-generation BTK inhibitor, has also demonstrated significant efficacy. A phase II study in Chinese patients with R/R MCL demonstrated high response rates and durable progression-free survival (PFS), leading to its approval in China [Table 1].[30]
The paradigm is now evolving from monotherapy toward rational combination strategies aimed at achieving deeper and more durable remissions. This includes synergistic pairings of BTK inhibitors with anti-CD20 antibodies. In MCL, combining ibrutinib with the anti-CD20 antibody rituximab has been explored to deepen responses, with some studies suggesting a higher overall response rate (ORR) compared to historical data for ibrutinib monotherapy, providing a chemotherapy-free option for R/R MCL.[31] More intensive chemotherapy-free combinations have also been investigated in R/R MCL, demonstrating high rates of complete response (CR) but without a clear survival benefit over less intensive regimens in early studies.[32] In R/R FL, the phase II ROSEWOOD trial demonstrated that the combination of zanubrutinib and obinutuzumab significantly improved the ORR and PFS compared to obinutuzumab monotherapy, establishing a new chemotherapy-free option for this indolent malignancy [Table 1].[33] Pushing the chemotherapy-free boundary into the frontline setting for aggressive lymphomas, the phase II POLARIS trial evaluated a novel triplet combination of orelabrutinib, lenalidomide, and rituximab in treatment-naïve MCL. The regimen established a highly effective, all-oral-based induction therapy for this patient population [Table 1].[34] These combination studies underscore the ongoing effort to optimize BTK inhibitor-based therapy by integrating it with other targeted or immunotherapeutic agents.

Overcoming resistance and future directions
Despite the profound efficacy of BTK inhibitors, the emergence of acquired resistance remains a significant clinical hurdle, limiting their long-term efficacy. The mechanisms of resistance are multifaceted and can be broadly categorized into two main types: On-target alterations and activation of bypass signaling pathways.[35] The most well-characterized on-target resistance mechanism is the C481S mutation in the BTK gene, which prevents the covalent binding of first-generation inhibitors. Less commonly, gain-of-function mutations in the downstream effector PLCG2 can also render cells resistant to upstream BTK inhibition.[36] This evolving landscape of resistance necessitates continuous innovation in therapeutic strategies.
One primary strategy to overcome resistance involves developing novel agents that can vertically target the BTK protein itself more effectively. To address the C481S mutation, a new class of noncovalent BTK inhibitors was developed. The pioneering agent in this class, pirtobrutinib, does not rely on the C481 residue for binding and has demonstrated remarkable clinical activity in patients with CLL and MCL, who have progressed on prior covalent BTK inhibitors [Table 1].[37] However, the evolutionary pressure of therapy continues, and resistance to pirtobrutinib can arise from new on-target BTK mutations (e.g., L528W, T474I) within the kinase domain.[38] Looking to the future, an even more advanced approach is the development of BTK-targeting proteolysis-targeting chimeras (PROTACs).[39] These molecules go beyond simple inhibition by inducing the complete degradation of the BTK protein, offering a potential strategy to eliminate both wild-type and mutated BTK and overcome kinase domain-mediated resistance.[40]
A parallel and complementary strategy is to horizontally target bypass pathways that allow malignant cells to circumvent their dependence on BTK. This approach is rooted in the understanding that resistance often involves the upregulation of alternative survival signals. For instance, in ibrutinib-resistant MCL, the identification of MALT1 overexpression or cyclin-dependent kinase 9 activation as escape mechanisms provides a strong rationale for combination therapies targeting these nodes.[41,42] The most clinically validated of these strategies is the dual targeting of BTK and the antiapoptotic protein B-cell lymphoma 2 (BCL-2). The combination of ibrutinib and venetoclax exerts a powerful synergistic effect by simultaneously blocking prosurvival signaling and directly promoting apoptosis, leading to significantly deeper and more durable responses in R/R MCL compared to historical monotherapy data.[43] The future of this approach lies in rationally combining BTK inhibitors with a broader array of agents, including those targeting other cellular vulnerabilities such as nuclear export (e.g., selinexor) or the epigenetic machinery,[444546] to create multipronged attacks that preempt or overcome resistance.

The BCL-2 apoptosis pathway

BCL-2 dysregulation as a therapeutic vulnerability in lymphoma
The evasion of apoptosis is a hallmark of cancer, and in many hematologic malignancies, this is achieved through the dysregulation of the BCL-2 family of proteins. This oncogenic dependency is particularly prominent in B-cell malignancies such as CLL, FL, and MCL, but is also a relevant therapeutic target in subsets of DLBCL. These proteins are the primary regulators of the intrinsic apoptosis pathway. The central event of this pathway is mitochondrial outer membrane permeabilization (MOMP), a point-of-no-return that releases proapoptotic factors such as cytochrome c into the cytoplasm, thereby triggering a caspase cascade that executes cell death [Figure 1].[47484950] The BCL-2 family comprises three functional subgroups: (1) the antiapoptotic proteins (e.g., BCL-2, BCL-xL, and MCL-1), which act as guardians of mitochondrial integrity by sequestering their proapoptotic counterparts;[51] (2) the proapoptotic effector proteins (BCL2 associated X, apoptosis regulator (BAX) and BCL2 antagonist/killer (BAK), which, upon activation, oligomerize to form pores in the mitochondrial outer membrane and execute MOMP;[52] (3) a diverse group of BH3-only proteins (e.g., BCL2-like 11, p53 upregulated modulator of apoptosis, and BCL2-associated death promoter), which function as cellular stress sensors.[53] Upon activation, these sensors neutralize the antiapoptotic guardians, thereby liberating BAX and BAK to induce apoptosis.[545556] In many lymphomas, particularly those of B-cell origin, such as CLL and FL, this homeostatic balance is pathologically tilted toward survival.[57] Malignant cells frequently overexpress antiapoptotic proteins, most notably BCL-2 itself, often as a result of the t(14;18) chromosomal translocation in FL or through other transcriptional mechanisms.[58] This BCL-2 overexpression allows tumor cells to sequester an excess of proapoptotic BH3-only proteins, effectively raising the threshold for apoptosis and conferring resistance to conventional therapies.[59] This state of oncogenic addiction to BCL-2 for survival creates a profound therapeutic vulnerability, providing a compelling rationale for developing agents that can specifically inhibit BCL-2 and restore the intrinsic capacity for cell apoptosis.[60]

Clinical development of BCL-2 inhibitors
The critical dependence of many lymphomas on antiapoptotic proteins has made them compelling therapeutic targets.[57] Early attempts to inhibit BCL-2 with agents such as the antisense oligonucleotide oblimersen or the gossypol derivative AT-101 were largely unsuccessful, constrained by low potency and unfavorable toxicity profiles.[61] A significant advancement came with the development of navitoclax, the first BH3-mimetic drug to enter clinical trials. As a dual inhibitor of both BCL-2 and BCL-xL, navitoclax showed clinical promise but was limited by on-target thrombocytopenia, a consequence of inhibiting BCL-xL, which is essential for platelet survival.[62]
This challenge directly spurred the development of venetoclax, a highly selective BCL-2 inhibitor designed to spare BCL-xL and thereby mitigate platelet toxicity.[63] As a potent BH3-mimetic, venetoclax binds with high affinity to the BH3-binding groove of the BCL-2 protein. This action displaces proapoptotic BH3-only proteins (such as BIM), liberating the effector proteins BAX and BAK to trigger MOMP and subsequent apoptosis.[64] The clinical impact of venetoclax has been profound, particularly in CLL, where it has demonstrated remarkable efficacy both as monotherapy and in combination regimens, even in patients resistant to prior therapies, such as BTK inhibitors [Table 1].[65] A new generation of selective BCL-2 inhibitors is emerging, exemplified by lisaftoclax (APG-2575). This novel BH3-mimetic was first evaluated in a phase I trial, which established its favorable safety profile and showed clinical activity in patients with R/R CLL/SLL and other hematologic malignancies [Table 1].[66] Subsequently, a pooled analysis combining this trial with a phase Ib/II study in a Chinese cohort provided more robust efficacy data.[67] In heavily pretreated CLL patients from this combined dataset, lisaftoclax monotherapy demonstrated an ORR of 73.3%, underscoring its potential as an effective new option in this setting [Table 1].
The success of venetoclax has validated the principle of selectively targeting BCL-2 family members and has broadened its application to other lymphomas. The phase II CAVALLI trial investigated the addition of venetoclax to standard R-CHOP immunochemotherapy as a first-line treatment for DLBCL [Table 1].[68] In this study of 206 patients, the combination regimen yielded a CRR of 69%. With a median follow-up of 32.2 months, the venetoclax plus R-CHOP arm showed a significant improvement in PFS compared to a historical R-CHOP cohort (HR = 0.61). While grade 3/4 hematologic adverse events were more frequent in the combination therapy arm (86%), they were generally manageable and did not increase treatment-related mortality. These findings highlight the potential of this combination to improve outcomes, especially in high-risk, BCL-2-positive DLBCL subgroups.
Meanwhile, research continues on inhibitors targeting other BCL-2 family members. Selective BCL-xL inhibitors are being explored for solid tumors, while selective MCL-1 inhibitors are under investigation for hematologic malignancies such as multiple myeloma and myelodysplastic syndrome, where MCL-1 is a key survival factor.[69] The development of these agents faces challenges, including complex protein structures and potential off-target effects, necessitating sophisticated strategies such as BH3 profiling to identify patients most likely to benefit.[69]

Overcoming resistance and future directions
Despite the success of venetoclax, the development of resistance is inevitable and poses a significant clinical challenge. The mechanisms are diverse and can be broadly classified into two categories: (1) alterations intrinsic to the BCL-2 family itself, and (2) activation of parallel prosurvival pathways. The former includes the upregulation of other antiapoptotic proteins, most notably MCL-1 and BCL-xL, which can compensate for BCL-2 inhibition.[70] Additionally, mutations in the BCL2 gene that disrupt drug binding, or loss-of-function alterations in proapoptotic effectors such as BAX, can also confer resistance.[71,72] The latter category often involves the activation of signaling cascades like the PI3K/AKT pathway or aberrations in cellular metabolism.[73]
Consequently, a primary strategy to overcome resistance is to target multiple antiapoptotic BCL-2 family members simultaneously. Given the prominent role of MCL-1 mediating resistance, the combination of venetoclax with selective MCL-1 inhibitors is a highly rational approach currently under active clinical investigation.[74] However, the development of MCL-1 inhibitors has been hampered by concerns about on-target cardiotoxicity.[75] An alternative strategy involves leveraging agents that can indirectly downregulate MCL-1. A phase Ib trial combining polatuzumab vedotin with venetoclax and an anti-CD20 antibody demonstrated high ORR in heavily pretreated patients with R/R FL (76%) and R/R DLBCL (29%).[76] Another major therapeutic avenue is to coinhibit the BCL-2 pathway and key prosurvival signaling cascades. Since venetoclax resistance is frequently linked to the activation of the PI3K/AKT/mTOR pathway, combining venetoclax with inhibitors of this axis is a logical strategy. Preclinical studies have demonstrated that PI3K inhibitors (e.g., duvelisib) or AKT inhibitors (e.g., MK-2206) can synergize with venetoclax to overcome resistance in both sensitive and resistant cell lines.[77,78] Similarly, given the frequent crosstalk between the BCR and BCL-2 pathways, combining venetoclax with BTK inhibitors such as ibrutinib has proven highly effective, yielding high ORR and CRR in diseases such as CLL and MCL.[70,79]
Finally, combining venetoclax with agents that modulate the apoptotic threshold or cellular epigenetics represents a promising frontier. For instance, combining venetoclax with conventional agents such as rituximab can overcome resistance by leveraging immune-mediated killing mechanisms and potentially downregulating BCL-2 expression.[80] Furthermore, epigenetic modulators, such as the DNA hypomethylating agent (HMA) decitabine and various histone deacetylase (HDAC) inhibitors (HDACis), can synergize with venetoclax. These agents are thought to reprime cancer cells for apoptosis by altering the expression of BCL-2 family members, thereby lowering the threshold for venetoclax-induced cell death and restoring sensitivity.[81] These rational combination strategies are key to broadening the applicability of BCL-2 inhibition and achieving more durable clinical outcomes.

The PI3K/AKT/mTOR pathway

Molecular architecture and aberrant activation of the PI3K/AKT/mTOR pathway in lymphoma
The PI3K/AKT/mTOR signaling pathway is a highly conserved network that serves as a central regulator of cell survival, growth, and metabolism. Upon activation by upstream signals such as growth factors, PI3K phosphorylates phosphatidylinositol(4,5)-bisphosphate to generate phosphatidylinositol(3,4,5)-trisphosphate [Figure 1]. This second messenger recruits and activates the serine/threonine kinase AKT, which in turn phosphorylates a multitude of downstream targets, including the mTOR complex. The mTOR complex, which exists as two distinct complexes (mammalian target of rapamycin complex 1 and mammalian target of rapamycin complex 2), orchestrates protein synthesis and metabolic programming by phosphorylating key effectors such as 4E-BP1.[82]
In malignant lymphocytes, this pathway is frequently hyperactivated through various mechanisms. These include (1) upstream oncogenic signals, such as aberrant BCR signaling or overactive receptor tyrosine kinases;[20] (2) loss of endogenous negative regulators, most notably the phosphatase and tensin homolog, which is often deleted or epigenetically silenced;[83] (3) direct activating mutations within pathway components themselves, such as in the PIK3CA gene.[84] The TME also plays a critical role, particularly in diseases such as CLL, where stromal cells provide continuous signals that sustain PI3K/AKT activity and promote cell survival.[85] Aberrant activation of this pathway is a common feature across a spectrum of lymphomas, including B-cell malignancies such as FL, CLL, and DLBCL, as well as in various peripheral T-cell lymphomas (PTCLs). This profound dependency renders the PI3K/AKT/mTOR axis a prime therapeutic target in lymphoma.

Clinical development of PI3K pathway inhibitors
The therapeutic targeting of this pathway has led to the development of a diverse array of inhibitors, including pan-PI3K inhibitors, isoform-selective PI3K inhibitors, dual PI3K/mTOR inhibitors, and specific AKT or mTOR inhibitors.[868788] The clinical journey of these agents has been a lesson in balancing efficacy with on-target and off-target toxicities.
PI3K inhibitors have seen the most extensive clinical development. Early-generation, isoform-selective PI3Kδ inhibitors such as idelalisib and the dual PI3Kδ/γ inhibitor duvelisib demonstrated significant efficacy in R/R B-cell lymphomas.[89] In the DYNAMO trial, duvelisib was used to treat patients with R/R indolent NHL [Table 1].[90] However, their clinical utility was quickly tempered by a challenging toxicity profile. These agents were associated with a high incidence of severe, late-onset, and often life-threatening immune-mediated toxicities, including colitis, pneumonitis, and hepatotoxicity.[91,92] Crucially, this toxicity is not an off-target effect but rather a mechanism-based, on-target consequence of inhibiting PI3Kδ and PI3Kγ, which are essential for the function and survival of regulatory T-cells.[93,94] The suppression of regulatory T cells (Tregs) disrupts immune homeostasis, leading to the observed autoimmune phenomena. This fundamental challenge underscored the need for new agents with an improved therapeutic window.
This challenge directly spurred the development of next-generation PI3Kδ inhibitors specifically designed to mitigate these immune-mediated adverse events. Agents such as umbralisib (which also inhibits CK1ε), linperlisib, and parsaclisib were engineered to have improved selectivity or different pharmacokinetic properties, leading to more favorable safety profiles.[959697] Among these, linperlisib gained approval in China for R/R FL based on a phase II trial that demonstrated high response rates and durable disease control, coupled with a significantly lower incidence of severe immune-mediated toxicities compared to its predecessors [Table 1].[98] In parallel, the pan-PI3K inhibitor copanlisib, which is administered intravenously, has shown durable efficacy in indolent lymphomas with a distinct and potentially more manageable toxicity profile compared to oral agents.[99100101102]
Notably, the role of PI3K inhibitors is expanding beyond B-cell lymphomas into PTCLs, a group of diseases with historically poor outcomes. The rationale is particularly strong for targeting the PI3Kδ and PI3Kγ isoforms, which are highly expressed in T-cells and mediate T-cell receptor signaling and chemokine-induced migration.[89] The dual PI3Kδ/γ inhibitor duvelisib received accelerated approval for R/R PTCL based on the pivotal phase II PRIMO trial. In this study, duvelisib monotherapy demonstrated significant clinical activity, establishing it as a crucial targeted therapy option for this challenging patient population [Table 1].[103] This represents a significant step forward, providing a much-needed targeted therapy option for this challenging patient population, although careful management of toxicities remains paramount.
Targeting nodes downstream of PI3K is another key strategy. mTOR inhibitors, including first-generation allosteric inhibitors (e.g., everolimus) and second-generation ATP-competitive kinase inhibitors, have shown activity, particularly in T-cell lymphomas and HL, often in combination with chemotherapy or other targeted agents.[104105106107] More recently, selective AKT inhibitors like capivasertib are being explored. Preclinical data suggest a powerful synergy when combining AKT inhibition with venetoclax, providing a rational basis for future clinical trials in DLBCL.[108] The overarching goal remains to identify the most effective agent or combination for specific lymphoma subtypes while minimizing toxicity through optimized dosing and rational partnerships.

Overcoming resistance and future directions
Despite progress, the clinical utility of PI3K/AKT/mTOR pathway inhibitors is constrained by acquired resistance and the lack of predictive biomarkers. Resistance mechanisms are complex, involving reactivation of the pathway itself, activation of parallel signaling cascades like the mitogenactivated protein kinase or Janus kinase (JAK)/signal transducer and activator of transcription (STAT) pathways, or metabolic reprogramming.[109,110] For instance, interleukin (IL)-6-mediated STAT3 activation has been identified as a bypass mechanism in PI3K inhibitor-resistant cells, suggesting that cotargeting JAK/STAT could be an effective strategy to overcome resistance.[111] Future efforts are focused on several key areas. First is the development of novel therapeutic modalities, such as PROTAC-based degraders, which aim to eliminate target proteins rather than just inhibit them, potentially offering a more profound and durable effect.[112] A second key area is the pursuit of rational combination therapies. This includes combining pathway inhibitors with each other (e.g., multilevel, low-dose inhibition of AKT and mTOR),[113] or with agents that have complementary mechanisms, such as epigenetic modulators (e.g., enhancer of zeste homolog 2 (EZH2) inhibitor or romidepsin),[114,115] which may also help mitigate toxicity. Finally, and perhaps most critical, is the need for biomarker-driven precision medicine. Leveraging multiomics to identify patients whose tumors are uniquely dependent on this pathway, or to predict which resistance mechanisms might emerge, will be essential to truly optimize the use of these powerful but challenging targeted therapies.

Epigenetic landscape

EZH2 inhibitors reversing aberrant histone methylation and transcriptional silencing
Dysregulation of histone methylation is a key oncogenic driver in many lymphomas, with the histone methyltransferase EZH2 being a central player. This mechanism is particularly critical in germinal center (GC) B-cell-derived lymphomas, such as FL and a subset of DLBCL. As the catalytic subunit of the polycomb repressive complex 2, EZH2 mediates the trimethylation of histone H3 at lysine 27 (H3K27me3), leading to transcriptional repression of target genes, including critical tumor suppressors [Figure 1].[116] A landmark discovery in GC B-cell lymphomas was the identification of recurrent, gain-of-function mutations in the catalytic SET domain of EZH2 (e.g., Y641F).[117] These mutations paradoxically collaborate with wild-type EZH2 to drive global H3K27me3 hypermethylation and promote lymphomagenesis.[118,119] In FL, this epigenetic reprogramming allows malignant B-cells to form an aberrant, self-sustaining immunological niche, explaining how an indolent tumor can arise from a highly proliferative environment.[120]
This clear oncogenic dependency spurred the development of EZH2 inhibitors. The first-in-class, selective EZH2 inhibitor, tazemetostat, was developed to specifically target both wild-type and mutant EZH2. It received regulatory approval based on a phase II trial demonstrating durable clinical responses and a favorable safety profile in patients with R/R FL, particularly those harboring EZH2 mutations [Table 1].[121,122] Preclinical studies also revealed that EZH2-mutant DLBCL cells are cytotoxic to tazemetostat, whereas wild-type cells are primarily cytostatic, and highlighted a synergistic effect with BCR pathway inhibitors and glucocorticoids. Furthermore, clinical trials have shown the feasibility of combining tazemetostat with the R-CHOP regimen for the treatment of DLBCL [Table 1].[123]
Although effective, the efficacy of selective EZH2 inhibition is limited in some contexts due to a compensatory mechanism involving its homolog, EZH1.[124] This led to the development of dual EZH1/2 inhibitors like valemetostat. By blocking both enzymes, valemetostat can achieve a more profound reduction of H3K27me3. This has translated into significant clinical activity, particularly in R/R adult T-cell leukemia/lymphoma and other PTCLs, where it recently gained approval in Japan [Table 1].[125126127128] The next frontier involves EZH2-targeting PROTACs, which go beyond inhibition to induce protein degradation, representing a potential strategy to achieve deeper and more lasting therapeutic effects in diseases like Burkitt lymphoma.[129]

DNA HMAs restoring normal DNA methylation patterns
Aberrant DNA methylation is another critical epigenetic lesion in lymphoma, disrupting the delicate balance between methylation writers and erasers. This process is primarily governed by two opposing enzyme families: the DNA methyltransferases (DNMTs) that establish and maintain 5-methylcytosine, and the ten-eleven translocation (TET) enzymes that initiate its removal. In lymphoma, both sides of this equilibrium are frequently dysregulated. On the “writer” side, DNMTs can be pathologically altered [Figure 1].[130] For instance, DNMT1 is often overexpressed in DLBCL, leading to hypermethylation, while loss-of-function mutations in the de novo methyltransferase DNMT3A are a hallmark of PTCLs, particularly the angioimmunoblastic-type (nTFHL-AI).[131,132] Dysregulation of the “eraser” side is even more common. Loss-of-function mutations in the primary eraser, TET2, are among the most frequent genetic alterations in T-cell lymphomas, especially nTFHL-AI, and also occur in B-cell lymphomas like DLBCL [Figure 1].[133] The inactivation of TET2 impairs DNA demethylation, leading to a global hypermethylation phenotype and altered gene expression that disrupts normal lymphocyte differentiation.[134135136137] The function of TET2 can also be indirectly compromised by mutations in the isocitrate dehydrogenase (IDH) enzymes. While wild-type IDH2 converts isocitrate to α-ketoglutarate, a critical cofactor for TET enzymes, recurrent mutations in IDH2 (e.g., R172K), which are particularly characteristic of nTFHL-AI, cause the enzyme to produce a potent oncometabolite, 2-hydroxyglutarate [Figure 1].[138] This 2-hydroxyglutarate competitively inhibits α-ketoglutarate-dependent enzymes, including TET2 and certain histone demethylases.[139] The result is a “TET2-loss-of-function-like” state, characterized by profound DNA and histone hypermethylation.[140] Together, the state of epigenetic chaos that provide a strong rationale for therapeutic intervention with agents that can reset these pathological methylation patterns.
The therapeutic rationale for reversing DNA hypermethylation has led to the clinical investigation of DNA HMAs, such as azacitidine and decitabine. These agents function as cytosine analogs that, upon incorporation into DNA, covalently trap DNMTs, leading to their degradation and the passive loss of methylation patterns. This can reactivate silenced tumor suppressor genes and enhance antitumor immunity.[141] The clinical utility of HMAs has been most extensively explored in PTCLs, a group of diseases where epigenetic dysregulation is a central pathogenic feature. Early studies with single-agent HMAs established their clinical activity. For instance, the guadecitabine monotherapy trial in R/R PTCL reported an ORR of 40%, with notable efficacy in patients with a TFH phenotype and those harboring RHOA G17V mutations, suggesting a potential biomarker-guided approach.[142]
Building on this, HMAs have been integrated into combination regimens. A significant advancement was the integration of oral azacitidine into CHOP-based chemotherapy for newly diagnosed PTCL [Table 1].[143] This phase II trial demonstrated that the regimen was safe and highly active, achieving an impressive CRR of 75% overall, and 88.2% in the TFH subtype. Importantly, this study also identified potential predictive biomarkers, showing that TET2 mutations were significantly associated with achieving a CR and favorable survival, whereas DNMT3A mutations were linked to adverse outcomes. A highly synergistic and chemotherapy-free approach involves combining HMAs with other epigenetic modulators, most notably HDACis. A phase II trial combining oral azacitidine with the HDACi romidepsin in R/R PTCL demonstrated impressive efficacy, with a high ORR and durable responses, again with particular benefit in patients with the TFH phenotype.[144] Furthermore, given their ability to enhance immune signaling, HMAs are being actively explored in combination with immune checkpoint inhibitors (ICIs). In R/R classic HL, where single-agent ICI efficacy is often limited, the combination of decitabine and a programmed cell death protein 1 inhibitor significantly improved response rates, particularly in high-risk patients.[145,146] Similarly, a pilot study of azacitidine plus nivolumab showed promising activity in R/R nTFHL-AI, providing a strong rationale for further investigation of this chemo-free, immuno-epigenetic approach.[147] These studies highlight the evolving role of HMAs, not just as single agents, but as potent synergistic partners capable of reshaping the therapeutic landscape of lymphoma.

HDACis modulating gene expression by promoting histone acetylation
HDACs are critical epigenetic regulators that remove acetyl groups from histones, leading to a condensed chromatin structure and the transcriptional repression of tumor suppressor genes. HDACis block this process, promoting a more open chromatin state, restoring gene expression, and ultimately inducing cell cycle arrest and apoptosis in malignant cells.[148] The clinical potential of this strategy was first demonstrated in 2001 with a report on the activity of depsipeptide (later known as romidepsin) in T-cell lymphoma, paving the way for the development of a new class of anticancer agents.[149]
The most profound impact of HDACis has been in the treatment of T-cell lymphomas, with several agents receiving regulatory approval based on robust clinical data.[150,151] Romidepsin, a potent cyclic peptide HDACi, is approved for both cutaneous T-cell lymphoma and PTCL. In R/R cutaneous T-cell lymphoma, it achieved an ORR of 34–35% as a single agent [Table 1].[152] Its approval in R/R PTCL was based on a pivotal study demonstrating durable responses, establishing it as a key therapeutic option.[153,154] Belinostat, a hydroxamic acid-based pan-HDACi, gained approval for R/R PTCL following the BELIEF trial [Table 1].[155,156] Notably, the response was even higher in the nTFHL-AI subtype, with an ORR of 46% compared to 26% for the overall cohort. Belinostat is also characterized by a favorable safety profile with low hematologic toxicity, making it a suitable option for patients with thrombocytopenia.[157] Chidamide (also known as tucidinostat), a novel benzamide-class, subtype-selective HDACi, was approved in China for R/R PTCL.[158] In a phase 2 trial, it demonstrated an ORR of 28%, comparable to other approved HDACis, but with a remarkably long median overall survival (OS) of 21.4 months, suggesting a significant long-term survival benefit [Table 1].[159] Like belinostat, it showed particular efficacy in the nTFHL-AI subtype. Beyond T-cell lymphomas, in the TRUST trial, the combination of chidamide and R-GemOx in patients with R/R DLBCL yielded high response rates and promising survival outcomes [Table 1].[160]
The success of these drugs has solidified epigenetic modulation via HDAC inhibition as a cornerstone of therapy, particularly for relapsed T-cell malignancies, and has spurred further research into novel combinations, such as with PD-1 inhibitors in NKTCL, to overcome resistance and improve outcomes.[161]

Challenges and future directions in epigenetic therapy
Despite their success, the clinical utility of epigenetic drugs is often constrained by acquired resistance and the complexity of their biological effects. Therefore, a key future direction lies in the development of rational combination strategies designed to achieve synergistic efficacy and preempt or overcome resistance. These strategies are increasingly being guided by a deep understanding of the intricate crosstalk between epigenetic and other cellular pathways.
One major approach involves creating synergy within the epigenome itself. Combining HMAs with HDACis, for instance, has shown strong preclinical synergy across various T-cell lymphoma models.[162] This combination aims to achieve a more comprehensive epigenetic reprogramming than either agent alone and is currently being evaluated in clinical trials. Another layer of complexity involves overcoming crosstalk-mediated resistance from parallel signaling pathways. In NKTCL, signaling nodes like MELK and JAK3 have been shown to stabilize EZH2, thereby conferring resistance to other therapies; this provides a clear rationale for cotargeting these kinases to restore therapeutic sensitivity.[163,164] Similarly, in MYC-driven B-cell lymphomas, the co-repressive complex formed by MYC, HDAC3, and EZH2 can be disrupted by dual inhibition of HDAC3 and EZH2, leading to the synergistic restoration of tumor suppressor microRNAs.[165] Novel resistance mechanisms, such as the activation of ferroptosis defense pathways in response to EZH2 inhibition in DLBCL, are also uncovering new potential combination partners like ferroptosis inducers.[166]
Perhaps the most promising frontier is the combination of epigenetic agents with immunotherapy. Epigenetic drugs can remodel the TME and enhance tumor cell immunogenicity, for example, by upregulating MHC class I expression on cancer cells, thereby making them more visible to the immune system.[167] This provides a powerful rationale for combining HMAs or EZH2 inhibitors with PD-1/programmed cell death ligand 1 (PD-L1) blockade, as well as T-cell therapy. Such combinations are being actively explored to overcome immune escape and enhance the efficacy of immunotherapy in a range of diseases, including Burkitt lymphoma, classic HL, and FL.[167168169170] Ultimately, the future of epigenetic therapy in lymphoma will depend on our ability to leverage these intelligent, mechanism-based combinations and to identify predictive biomarkers that can guide patient selection toward achieving deeper and more durable responses.

Nuclear export pathway—exportin 1 (XPO1)

The role of XPO1 in oncogenesis and tumor suppressor inactivation in lymphoma
XPO1 (also known as CRM1) protein is a critical regulator of cellular homeostasis, responsible for transporting a wide range of cargo proteins and RNA from the nucleus to the cytoplasm. Its cargo includes numerous tumor suppressor proteins (TSPs) such as TP53, RB1, and CDKN1A, as well as the NF-κB inhibitor, IκB [Figure 1].[171] In many hematologic malignancies, XPO1 is overexpressed, leading to two key oncogenic consequences: (1) the forced nuclear export and subsequent cytoplasmic inactivation of TSPs, preventing them from executing their functions of cell cycle arrest and apoptosis induction; (2) the enhanced degradation of IκB, resulting in constitutive activation of the prosurvival NF-κB pathway. This dual mechanism of inactivating tumor suppressors while activating oncogenic pathways makes XPO1 a prime therapeutic target.[172,173]
Selinexor, a first-in-class, oral, selective inhibitor of nuclear export, functions by covalently binding to the Cys528 residue in the cargo-binding pocket of XPO1.[174] This action blocks the export machinery, leading to the nuclear retention and functional restoration of TSPs and the suppression of oncogenic signaling. This potent mechanism of action has led to its regulatory approval for R/R DLBCL and multiple myeloma.[175]

Clinical development of XPO1 inhibitors
The clinical utility of selinexor has been demonstrated across a range of lymphoma subtypes, including B-cell lymphomas like DLBCL and MCL, as well as T-cell and NK-cell lymphomas, particularly in heavily pretreated populations. The approval for DLBCL was based on the phase II SADAL trial [Table 1]. This study enrolled R/R DLBCL patients who had received at least two prior lines of therapy, a population with a poor prognosis and an expected median OS of less than 6 months. In this challenging setting, single-agent oral selinexor achieved an ORR of 29.1% and a median OS of 9 months.[176] A subsequent detailed survival analysis of the SADAL study further underscored its clinical benefit.[177] At a median follow-up of 14.8 months, the median OS for the entire cohort was 9.0 months, significantly exceeding historical controls. The most striking finding was the profound survival advantage conferred by achieving a response: Patients who attained a partial or complete response had a markedly prolonged median OS of 29.7 months, compared with 4.9 months for non-responders. This demonstrates that for patients who respond, selinexor can provide deep and durable disease control. To move selinexor into earlier lines of therapy, the SELINDA trial evaluated its combination with R-GDP (rituximab, gemcitabine, dexamethasone, and cisplatin) salvage chemotherapy, establishing a recommended phase II dose and demonstrating tolerable safety and promising efficacy in R/R B-cell lymphomas [Table 1].[178] Preclinical studies also suggest that combining selinexor with other targeted agents, such as BET inhibitors, could overcome resistance, particularly in high-risk subtypes like double-hit lymphoma.[179]
Selinexor is also being explored in T-cell and NK-cell lymphomas, diseases with urgent unmet needs. A retrospective study in R/R NKTCL showed that the combination of selinexor with anti-PD-1 therapy could provide meaningful clinical benefit in patients who had failed prior therapies, including those with challenging CNS involvement.[180] Furthermore, a preclinical study in T-cell lymphoblastic lymphoma demonstrated a strong synergistic antitumor effect when selinexor was combined with the HMA decitabine, providing a rationale for clinical investigation of this novel combination in this rare and aggressive disease.[181]

Overcoming toxicities and future directions
A significant challenge with selinexor is its toxicity profile, which, while generally manageable, requires proactive and supportive care to ensure patient adherence and treatment success. The most common adverse events include cytopenia, particularly thrombocytopenia, and constitutional and gastrointestinal side effects such as nausea, anorexia, fatigue, and weight loss. Effective management of these toxicities is crucial. For example, nausea is typically managed with a prophylactic antiemetic regimen, often including a 5-HT3 receptor antagonist (e.g., ondansetron) combined with olanzapine or a neurokinin-1 receptor antagonist for patients at high risk. Anorexia and weight loss require nutritional counseling and appetite stimulants, while cytopenia may necessitate dose interruptions or reductions and careful monitoring. This comprehensive supportive care framework is essential for maintaining patients on therapy and optimizing clinical benefit.[182]
To address these limitations at a pharmacological level, the next-generation SINE compound, eltanexor (KPT-8602), was developed.[183] It was designed with a shorter half-life and reduced blood-brain barrier penetration to have less impact on normal cells, thereby aiming to decrease systemic toxicity while retaining antitumor efficacy. This improved therapeutic window may allow for more frequent or continuous dosing schedules, potentially enhancing its combination potential with other targeted therapies. The development of such better-tolerated agents represents a key direction for the future of XPO1 inhibition in lymphoma, promising to broaden its clinical applicability and improve the patient experience.[184,185]

The JAK/STAT pathway

Molecular architecture and aberrant activation of the JAK/STAT pathway
The JAK-STAT pathway is a crucial mediator of cytokine signaling, essential for normal hematopoiesis and immune cell function. The JAK family comprises four non-receptor tyrosine kinases (JAK1, JAK2, JAK3, and TYK2) that are activated by ligands binding to cytokine receptors. This activation triggers a phosphorylation cascade, leading to the phosphorylation and activation of STAT proteins. Activated STATs then dimerize, translocate to the nucleus, and regulate the expression of target genes involved in cell proliferation, differentiation, and apoptosis, such as BCL2L1 and CCND1 [Figure 1].[186,187]
In various lymphomas, particularly in PTCLs and NKTCL, the JAK/STAT pathway is frequently hyperactivated.[188] This aberrant activation is often driven by gain-of-function mutations in pathway components, such as in the pseudokinase domain of JAK family members (e.g., JAK1 Y658F, JAK2 V617F, and JAK3 A573V) or in the SH2 domain of STAT proteins (e.g., STAT3 Y640F, STAT5B N642H).[189,190] Despite the presence of these activating mutations, the oncogenic function of the pathway remains dependent on the structural scaffold provided by cytokine receptors, making multiple nodes within this axis viable therapeutic targets.[189,190]

Clinical development of JAK/STAT pathway inhibitors
The dependency of certain lymphomas on JAK/STAT signaling has spurred the development of targeted inhibitors. The main JAK/STAT pathway-related drugs currently being developed include JAK inhibitors such as ruxolitinib (JAK1/2 inhibitor), tofacitinib (pan-JAK inhibitor), STAT inhibitors such as lisofylline, and anticytokine neutralizing antibodies such as anti-IL-15 antibody.[189] The JAK1/2 inhibitor ruxolitinib has demonstrated clinical activity in R/R PTCLs, with a phase II study reporting an ORR of 25% and a median PFS of 2.8 months [Table 1].[191] More recently, golidocitinib, a highly selective JAK1 inhibitor, showed greater efficacy in a phase I trial of R/R PTCLs [Table 1].[192]
Combination strategies are also being actively explored. In HL, combining ruxolitinib with the anti-PD-1 antibody nivolumab has shown promise for overcoming resistance to ICIs.[193] Furthermore, preclinical studies in NKTCL have revealed that JAK/STAT signaling can mediate resistance to epigenetic drugs like chidamide, providing a strong rationale for combining JAK inhibitors with HDACis to achieve synergistic antitumor effects.[194] While direct STAT inhibitors are in early development, no agents targeting this node have yet received FDA approval for cancer treatment.[195]

Overcoming resistance and future directions
Resistance to JAK inhibitors is a growing clinical challenge. Mechanisms include the heterodimerization of JAK kinases (e.g., JAK2 with JAK1 or TYK2), which can reactivate downstream signaling, and protective signals from the bone marrow microenvironment.[186] Potential strategies to overcome resistance include combining JAK inhibitors with agents that promote JAK2 degradation, such as heat shock protein 90 or HDACis.[196] In the context of immunotherapy resistance in PTCL driven by specific JAK3 mutations, the pan-JAK inhibitor tofacitinib has shown efficacy in preclinical models, suggesting a path for precision therapy.[197] Additionally, emerging research indicates that inhibiting the integrin αvβ3 with cilengitide can alleviate abnormal JAK/STAT activation, offering a novel, potentially less toxic therapeutic strategy.[198]
A key future direction will be to conduct clinical trials of JAK/STAT inhibitors specifically in PTCL subtypes known to be highly dependent on this pathway. These include entities such as monomorphic epitheliotropic intestinal T-cell lymphoma, hepatosplenic T-cell lymphoma, and NKTCL, where pathway activation is a defining pathogenic feature.[199,200] Such biomarker-driven studies will be critical for realizing the full potential of JAK/STAT inhibition in precision therapy for PTCLs.[186,197,198]

The cyclic GMP–AMP synthase–stimulator of interferon genes (cGAS–STING) innate immunity pathway

cGAS–STING pathway as a central sensor for innate immunity
The cGAS-STING pathway is a critical component of the innate immune system that functions as a primary sensor for cytosolic double-stranded DNA (dsDNA).[201] The cytosolic dsDNA that activates this pathway can originate from multiple sources [Figure 1]. Endogenous sources within tumor cells include DNA fragments from ruptured micronuclei, which often form due to a deficient DNA damage response, as well as mitochondrial DNA released from damaged mitochondria. Additionally, exogenous dsDNA from pathogens such as viruses and bacteria, or from the debris of neighboring dead cells, can also enter the cytoplasm and trigger cGAS activation. The core mechanism involves cGAS-mediated synthesis of the second messenger cGAMP, which in turn activates the adaptor protein STING.[202] Activated STING, through the TBK1–IRF3 axis, orchestrates the production of type I interferons, ILs, and tumor necrosis factor, which are pivotal for bridging innate and adaptive antitumor immunity by promoting dendritic cell maturation and enhancing cytotoxic T-cell responses.[203204205]

A dichotomous role of the cGAS-STING pathway in lymphoma
While the canonical role of the cGAS–STING pathway is tumor suppressive, its ultimate impact on lymphomagenesis is highly context-dependent, varying significantly across different lymphoma subtypes [Figure 1]. In B-cell lymphomas, such as DLBCL and FL, the pathway functions primarily as a tumor suppressor. Its activation, either intrinsically through genomic instability (e.g., loss of SAMHD1[206]) or extrinsically through therapy (e.g., the bendamustine–rituximab regimen[207] and natural compound sulforaphane[208]), leads to potent antitumor effects, including the induction of programmed cell death and the enhancement of antitumor immunity. In contrast, emerging evidence suggests that in PTCLs, the cGAS-STING pathway can be oncogenic, where its activation is associated with increased proliferation and poor clinical outcome.[209] In this context, chemotherapy-induced DNA damage paradoxically activates this prosurvival pathway, contributing to therapeutic resistance. This discovery reveals a completely opposite therapeutic paradigm for PTCL, where inhibiting cGAS can suppress tumor growth and synergize with chemotherapy. This subtype-specific duality underscores the critical importance of understanding the precise biological role of the pathway before designing therapeutic interventions.
Another mechanism involves the pathway being hijacked and disabled to promote viral oncogenesis, as exemplified by mutations in the RNA helicase DDX3X in Epstein–Barr virus-associated lymphomas, especially NKTCL. In this context, the cGAS-STING pathway represents a critical antiviral defense that must be overcome for the tumor to develop.[210] Pathogenic mutations in DDX3X achieve this by sabotaging the signaling cascade downstream of STING, specifically impairing the activation of the IRF3/7 transcription factors. This targeted crippling of the innate immune response creates a permissive environment for the overexpression of potent EBV oncoproteins such as BNLF2b. The viral oncoprotein then drives the accumulation of pathological R-loops, leading to genomic instability and malignant progression. Thus, in this scenario, the abrogation of the cGAS-STING pathway’s antiviral surveillance function is a prerequisite for tumorigenesis. This highlights a third, distinct mechanism by which dysregulation of this pathway contributes to lymphomagenesis—not by being intrinsically oncogenic, but by having its tumor-suppressive function disabled to permit another oncogenic driver to emerge.

Therapeutic modulation of the cGAS-STING pathway
The dual role of the cGAS-STING pathway necessitates a nuanced therapeutic approach, involving both agonists and inhibitors. STING agonists are being developed with two distinct therapeutic rationales. The canonical approach leverages their ability to potently induce a type I interferon response, aiming to convert immunologically “cold” tumors into “hot” ones and thereby sensitize them to immune checkpoint blockade.[211,212] However, a particularly compelling and distinct application is in directly overcoming therapy resistance by rewiring apoptotic signaling. In hematologic malignancies with TP53 mutations, cancer cells are often resistant to BCL-2 inhibitors (BH3-mimetics) because they lack a TP53-dependent feedforward loop that normally amplifies the apoptotic signal. STING agonists can circumvent this resistance mechanism entirely. Upon activation, the STING pathway can induce the upregulation of proapoptotic BH3-only proteins in a TP53-independent manner. This action effectively restores the missing apoptotic signal, synergizing powerfully with BH3-mimetics to kill cancer cells. This cell-intrinsic, nonimmunological mechanism provides a powerful rationale for combining STING agonists with agents like venetoclax to treat high-risk, TP53-mutant cancers.[213]
Conversely, the discovery of its protumorigenic role in PTCLs has validated the development of cGAS or STING inhibitors as a viable anticancer strategy for specific lymphoma subtypes.[209] The challenge for all these agents lies in achieving targeted delivery to maximize efficacy while minimizing systemic side effects, a problem being addressed by novel delivery systems.[214] The future of STING-based therapy in lymphoma will undoubtedly rely on biomarker-driven patient stratification to determine whether a given patient’s tumor should be treated with an agonist or an inhibitor.

Integration of Targeted Therapies into Clinical Practice

Rise of chemotherapy-free regimens
A major paradigm shift in lymphoma treatment, and a direct triumph of decades of basic and translational research into signaling pathways, is the development of chemotherapy-free regimens. These strategies stem from a deep molecular understanding of how specific lymphomas become addicted to signaling pathways. By replacing or omitting nonspecific cytotoxic agents with rationally designed combinations of targeted drugs, monoclonal antibodies, and immunomodulators, these regimens aim to selectively eradicate malignant cells while sparing normal tissues. This approach has proven particularly valuable for elderly or frail patients and those with R/R disease, with its application expanding across the lymphoma spectrum.
In indolent lymphomas, where the primary goal is often durable disease control with minimal toxicity, this new paradigm is perhaps most advanced. The quintessential example is in CLL/SLL, a disease defined by its profound addiction to the BCR signaling pathway.[215] BTK inhibitors, used either as monotherapy or in combination with anti-CD20 antibodies, provide a highly effective, oral, and chemotherapy-free approach that can induce lasting remissions. This success is a direct clinical manifestation of targeting core molecular dependency of the disease, a principle that is now being extended to other indolent lymphomas.[216] The success in CLL has been mirrored in Waldenström’s macroglobulinemia, which also heavily relies on BCR signaling. The use of BTK inhibitors, either as monotherapy or in combination, has become a central component of its chemotherapy-free management. In other indolent lymphomas, such as marginal zone lymphoma and FL, signaling pathway inhibitors are key components of chemotherapy-free strategies for R/R disease.[217] For instance, PI3K inhibitors and the EZH2 inhibitor (particularly in EZH2-mutant FL) have established efficacy as monotherapies.[218] In aggressive lymphomas, chemotherapy-free strategies are being developed to address the dual challenges of chemo-insensitivity and treatment-related toxicity. In aggressive lymphomas, the central role of the BCR pathway has established BTK inhibitors as the backbone of many chemotherapy-free strategies. For instance, in the R/R setting, multiagent regimens have been developed around this principle. The DR2IVE regimen (dexamethasone, rituximab, lenalidomide, bortezomib, and ibrutinib) is one such example, where the BTK inhibitor is combined with other targeted and immunomodulatory agents to create a multipronged attack on various survival pathways.[219] Similarly, in DLBCL, particularly for elderly or unfit patients with subtypes dependent on chronic active BCR and NF-κB signaling, the BTKi-based iR2 regimen (ibrutinib, rituximab, lenalidomide) has demonstrated significant efficacy, validating the approach of targeting core survival pathways in a chemotherapy-free context.[220] Another phase 2 trial evaluated the ZR2 regimen (zanubrutinib, rituximab, and lenalidomide) in newly diagnosed DLBCL patients aged ≥ 75 years.[221] The results were highly encouraging, with a CRR of 65.0% and a 2-year OS of 82.4%, comparing favorably to historical outcomes with R-miniCHOP. This chemotherapy-free paradigm is also extending to T-cell lymphomas. In R/R PTCLs, therapies targeting the epigenome are particularly prominent. The combination of the oral HMA azacitidine with the HDACi romidepsin, for example, has shown impressive efficacy, providing a crucial option for a patient population with poor outcomes with conventional chemotherapy.[144]
Each of these successful chemotherapy-free regimens is, in essence, a clinical manifestation of our deep understanding of molecular drivers in lymphomas. They illustrate a clear trajectory away from a “one-size-fits-all” cytotoxic approach toward more tailored, biology-driven treatments, marking a new era where insights from signaling pathway research are directly translated into improved patient outcomes.

Predictive biomarkers for signaling pathway-targeted therapies
The ultimate goal of dissecting lymphoma signaling pathways is to enable precision medicine, where therapeutic decisions are guided by predictive biomarkers. Unlike prognostic markers such as MYC/BCL2 dual expression in DLBCL that only stratify risk,[222] predictive biomarkers identify the specific molecular vulnerabilities of a tumor, allowing for the selection of targeted therapies most likely to be effective. This section summarizes some of the key predictive biomarkers discussed in this review, which are pivotal for matching the right drug to the right patient.
Several well-established and emerging biomarkers now directly inform clinical practice. A prime example of this principle is the use of EZH2 mutations in FL. The presence of a gain-of-function EZH2 mutation is a well-established predictive biomarker for sensitivity to the EZH2 inhibitor tazemetostat. Patients with EZH2-mutant tumors derive significantly greater benefit from this agent, rendering the assessment of EZH2 mutation status a key component of treatment decision-making.[223]
Biomarker-driven approaches are also crucial in T-cell lymphomas, where epigenetic dysregulation is a central feature. As highlighted in recent clinical trials, specific genetic alterations can guide the use of epigenetic therapies in PTCLs. For example, the presence of TET2 mutations is associated with a favorable response to HMAs, whereas DNMT3A mutations may predict a poorer outcome, highlighting how genetic biomarkers can guide the use of epigenetic agents.[224] Furthermore, for T-cell and NKTCL dependent on the JAK/STAT pathway, both pathway activation status (e.g., phosphorylated STAT3 levels) and the presence of activating mutations in JAK3 or STAT3 are emerging as powerful predictive biomarkers for response to JAK/STAT inhibitors.[191,225,226]
However, for many targeted agents, the identification of reliable predictive biomarkers remains a critical challenge and an area of active research. For the XPO1 inhibitor selinexor, which is approved for R/R DLBCL, there is an ongoing effort to identify molecular correlates of response. Early studies suggest that the protein expression levels of XPO1 itself, as well as key cargo proteins like TP53 and downstream effectors like MCL-1, may serve as potential protein-level biomarkers to help select patients who are most likely to benefit from this therapeutic approach.[227] The continued discovery and clinical validation of such predictive biomarkers are essential for realizing the full potential of targeted therapies in lymphomas.[228]

Genetic subtype-guided combination therapy by aligning treatment with signaling pathway dependencies
A pivotal advancement in personalizing lymphoma treatment is the use of genetic subtyping to guide the addition of targeted agents to standard immunochemotherapy backbones. This strategy is built upon foundational discoveries that have dissected the molecular heterogeneity of lymphomas, revealing distinct subtypes defined by their addiction to specific signaling pathways. For example, in DLBCL, comprehensive genomic analyses have identified subtypes like MCD (driven by MYD88 and CD79B mutations, indicating chronic active BCR/NF-κB signaling) and BN2 (characterized by BCL6 fusions and NOTCH2 mutations).[229] Similarly, in PTCLs, four major molecular subtypes have been defined, each with unique dependencies, such as the T1 subtype’s reliance on aberrant T-cell receptor/RHOA signaling.[230] These classifications provide a biological roadmap, allowing for the rational selection of targeted agents to inhibit the specific pathways driving each subtype.
The GUIDANCE-01 trial provides a powerful clinical validation of this approach in newly diagnosed, intermediate/high-risk DLBCL [Figure 2A].[231] In this randomized study, patients were stratified into genetic subtypes and received either standard R-CHOP or a subtype-guided regimen of R-CHOP plus a targeted agent (R-CHOP-X)—for instance, adding a BTK or NF-κB pathway inhibitor for BCR-dependent subtypes. The results were striking: the addition of a rationally chosen targeted agent led to significantly higher CRR (88% vs. 66%), 2-year PFS (88% vs. 63%), and 2-year OS (94% vs. 77%) compared to R-CHOP alone. This trial elegantly demonstrated that aligning targeted therapy with the underlying pathway dependency can dramatically improve the efficacy of the standard-of-care backbone.
This biomarker-driven paradigm has also been successfully applied to frontline PTCLs in the GUIDANCE-03 trial [Figure 2B].[232] Acknowledging the diverse epigenetic and signaling aberrations in PTCLs, patients in the experimental arm received a CHOP-X regimen, where the “X” was an epigenetic modulator or targeted agent tailored to their tumor’s specific mutations. For example, patients with histone modifier mutations (CREBBP/EP300) received the HDACi chidamide, whereas others received agents such as the HMA decitabine for TP53 mutations. This personalized strategy resulted in a near-doubling of the CRR (64.6% vs. 33.3%) and a dramatic extension of the median PFS (25.5 vs. 9.0 months) compared to standard CHOP.[232] This study provides compelling evidence that tailoring targeted therapy to the dominant molecular aberrations can overcome the inherent chemoresistance of PTCLs. The same principle is being applied to the R/R setting. The GUIDANCE-06 trial evaluated a genetic subtype-guided approach for R/R DLBCL, adding targeted agents to the R-ICE regimen (rituximab, ifosfamide, carboplatin, and etoposide) salvage chemotherapy backbone (R-ICE-X) [Figure 2C].[233] This strategy yielded impressive outcomes, with a 2-year PFS of 69.3 % and a 2-year OS of 88.3 %, rates that are highly favorable for this patient population.
Collectively, these studies herald a new era of lymphoma therapy. They transform genetic subtyping from a purely prognostic tool into an actionable predictive biomarker, guiding the construction of more effective, personalized combination regimens. This marks a critical step forward, where insights from signaling pathway and epigenetic research are directly used to enhance the efficacy of standard immunochemotherapy for defined patient cohorts.

Current challenges and future directions in lymphoma targeted therapy

Overcoming acquired resistance
While the advent of targeted therapy has significantly improved prognoses for patients with lymphoma, acquired drug resistance remains a primary obstacle limiting long-term therapeutic efficacy.[234,235] Consequently, a central focus of current research is the comprehensive analysis of drug resistance mechanisms and the development of next-generation therapeutic strategies to circumvent them.[235]
Mechanisms of resistance are multifaceted. A primary driver is the acquisition of on-target mutations that alter drug binding.[236] The C481S “gatekeeper” mutation in the BTK gene, for instance, is a well-characterized mechanism of drug resistance to covalent BTK inhibitors, as it disrupts drug binding by altering the active site conformation.[237] Another key mechanism involves the activation of compensatory or bypass signaling pathways. Abnormal upregulation of the PI3K/AKT/mTOR pathway, for example, can circumvent the BTK blockade, thereby promoting cell survival and proliferation independent of the original target.[238] In addition, extrinsic factors, including remodeling of the TME—such as increased infiltration of immunosuppressive cells—and the enrichment of tumor stem cell populations,[239] also contribute to diminished drug sensitivity and therapeutic failure.[240241242243]
To address these resistance mechanisms, several next-generation therapeutic strategies are under investigation. For instance, PROTAC technology offers a new approach by inducing the degradation of the target protein rather than merely inhibiting its enzymatic activity.[244] This mechanism can overcome resistance conferred by mutations that affect inhibitor binding. Notably, BTK-targeting PROTACs have shown efficacy in degrading the C481S mutant protein in preclinical models.[245] Combination therapies are a cornerstone for overcoming resistance. Combining a BTK inhibitor with a BCL-2 inhibitor has shown synergistic effects by simultaneously blocking proliferation signals and promoting apoptosis, significantly improving outcomes in challenging patient populations.[246] Finally, the implementation of liquid biopsies for dynamic monitoring of circulating tumor DNA is emerging as a powerful tool.[247] Detecting resistance mutations in circulating tumor DNA can provide an early warning of impending clinical relapse, enabling preemptive or personalized adjustments to treatment plans and representing a paradigm shift toward precision medicine in lymphoma management.[248]

Rational combination strategies
To counteract the complex interplay of resistance mechanisms and tumor heterogeneity, rational combination strategies have become a cornerstone of modern lymphoma therapy. The goal is to achieve synergistic cytotoxicity, prevent or delay the emergence of resistance, and broaden therapeutic efficacy. These strategies primarily involve pairing targeted agents with epigenetic modulators, other targeted agents, immunotherapy, or innovative chemotherapy regimens.
A key strategy involves sensitization through epigenetic reprogramming. Epigenetic agents, such as HMAs and HDACis, can reverse aberrant gene silencing. For instance, in R/R DLBCL, combining the HMA decitabine with standard immunochemotherapy (e.g., R-CHOP) or as a pretreatment for CD19/CD22 chimeric antigen receptor-modified T (CAR-T) therapy has shown promise in resensitizing tumors and improving response rates.[249,250] The underlying rationale is that these agents can restore the expression of tumor suppressor genes and critical surface antigens. Similarly, HDACis can overcome immune escape by upregulating CD20 to enhance the efficacy of anti-CD20 CAR-NK cells or rituximab, and restore tumor necrosis factor-related apoptosis-inducing ligand (TRAIL)-mediated apoptosis pathways.[251,252] In the same way, activation of the JAK/STAT pathway and epigenetic regulation are closely associated with ICI resistance. Combining JAK inhibitors with epigenetic drugs can significantly improve therapeutic efficacy.[253,254]
Another powerful approach is the combination of different targeted agents to induce synthetic lethality. This involves the simultaneous blockade of parallel or compensatory survival pathways. In MCL and double-hit lymphoma, combining the XPO1 inhibitor selinexor with the BCL-2 inhibitor venetoclax has demonstrated strong synergistic effects. Selinexor traps TSPs within the nucleus while venetoclax directly triggers the mitochondrial apoptosis pathway, leading to enhanced cell death.[255,256] Preclinical models in DLBCL also support combining venetoclax with agents like sildenafil, which synergistically downregulate key oncogenes such as MYC and BCL2, providing a rationale for targeting multiple core dependencies.[257]
Integrating targeted therapies with immunotherapy is a rapidly evolving frontier. The central principle is that targeted agents can modulate the TME and enhance tumor cell immunogenicity, thereby boosting the efficacy of immunotherapies like ICIs, CAR-T cells, and bispecific antibodies.[258,259] For example, as previously discussed, the EZH2 inhibitor tazemetostat can upregulate MHC molecules on FL tumor cells, making them more visible to the immune system and potentially overcoming resistance to ICIs.[167] Similarly, combining JAK inhibitors with PD-1 blockade has shown promise in overcoming ICI resistance in HL.[253,254] In the realm of cellular therapy, epigenetic priming with agents like decitabine before CAR-T cell infusion has been shown to improve patient outcomes in R/R DLBCL, likely by resensitizing chemoresistant cells.[249,250] Future strategies will undoubtedly focus on creating synergistic triplets, combining targeted pathway inhibitors (e.g., BTKi, PI3Ki) with ICIs and next-generation cellular therapies to achieve deep and durable responses, transforming the treatment paradigm for even the most aggressive lymphomas.[260,261]

Real-world implementation
Although the molecular and clinical breakthroughs in targeted therapies are transformative, their real-world implementation is significantly constrained by practical challenges, most notably toxicity, cost-effectiveness, and inequitable global access. The successful management of treatment-related toxicities, as discussed for agents such as PI3K and XPO1 inhibitors, is a prerequisite for clinical adoption. However, even when toxicities are manageable, socioeconomic barriers often determine whether a patient can benefit from these innovations.
The high cost of novel agents presents a formidable hurdle. Targeted drugs, despite their significant therapeutic benefits, often enter the market with prohibitively high prices due to expensive research, development, and patent-protected pricing.[262] This economic reality creates a stark divide in global access. In high-income countries, access is largely dependent on the robustness of public or private health insurance systems. In contrast, patients in low- and middle-income countries face a confluence of barriers, including patent laws, limited governmental healthcare budgets, and low individual payment capacity, resulting in extremely low access rates and exacerbating global health disparities.[263]
Beyond direct costs, delays in drug availability further impact patient outcomes. Patented anticancer drugs are significantly less affordable in low- and middle-income countries, and patients in these nations often experience a multiyear lag in accessing new therapies.[264] Although the introduction of biosimilars and generic drugs can help narrow this treatment gap, a multipronged approach is essential for the future. Strategies such as mandatory licensing, drug price caps, differentiated pricing, and expanding universal healthcare coverage will be critical to ensuring that the promise of precision medicine in lymphoma translates into equitable and accessible care for all patients, regardless of their geographic location or economic status.[265]

Leveraging multiomics and artificial intelligence (AI)
The clinical challenge in lymphoma is rooted in the profound heterogeneity of the oncogenic signaling networks that drive tumor survival and resistance. To move beyond a one-gene, one-drug paradigm, the convergence of multiomics technologies and AI is essential for deconstructing this complexity and enabling truly pathway-informed personalized medicine.
Multiomics platforms provide a multilayered readout of signaling pathway activity. Genomics identifies mutations in key pathway components (e.g., MYD88, CD79B); transcriptomics quantifies the downstream output of hyperactive pathways (e.g., NF-κB or MYC target gene signatures); and phosphoproteomics directly measures the activation state of critical kinases (e.g., p-BTK, p-AKT). This integrated approach is crucial for uncovering pathogenic mechanisms. For example, in FL, multiomics identified mutations in epigenetic regulators like CREBBP and KMT2D as early events that lead to sustained BCL6 activity and the dysregulation of downstream pathways essential for GC B-cell biology.[266] Similarly, in PTCLs, this approach has stratified tumors into subtypes defined by distinct TME, which are shaped by underlying cytokine signaling networks and offer different therapeutic vulnerabilities.[267]
The analytical power of AI is required to integrate these data layers and identify functional signaling states that are not apparent from single-analyte analysis. This synergy is revolutionizing diagnostics and prognostics by linking observable features to underlying pathway biology.[268] In digital pathology, deep learning is moving beyond simple classification to correlate morphological phenotypes on whole-slide images with specific oncogenic signaling signatures, providing a visual surrogate for molecular tests.[269] In radiomics, AI can link imaging features from MRI or PET/CT scans to biological processes governed by signaling pathways, such as angiogenesis or proliferation, thereby creating non-invasive prognostic biomarkers.[270,271] In liquid biopsy, the COMOS platform utilizes machine learning to integrate conventional omics, such as methylation, copy number alterations, with novel fragmentomics from cell-free DNA. These fragmentation patterns reflect chromatin accessibility, indirectly reporting on the activity of transcription factors and signaling pathways.[272] Ultimately, this fusion of technologies is set to accelerate the development of pathway-targeted therapies. AI algorithms can mine multiomics datasets to identify novel, druggable nodes within dysregulated signaling networks or predict, based on a tumor’s unique “pathway signature”, which signaling inhibitor a patient is most likely to respond to.[273] Although challenges in clinical integration remain, the combination of multiomics and AI is undeniably shaping a new era of lymphoma treatment, one focused on deciphering and therapeutically targeting the complex and dynamic signaling networks at the heart of the disease.

Conclusions
The journey of targeted therapies in lymphoma represents a landmark achievement in modern oncology, transforming the treatment paradigm from the indiscriminate cytotoxicity of chemotherapy to an era of genetic subtype-guided precision medicine. This evolution was built upon decades of fundamental research that unraveled the intricate signaling pathways that drive lymphomagenesis. The translation of this knowledge into clinically effective agents has delivered unprecedented improvements in patient outcomes, particularly in historically difficult-to-treat subtypes and in patients unsuitable for conventional chemotherapy.
The impact of these therapies extends beyond single-agent efficacy; they have catalyzed an ongoing shift toward personalized treatment strategies. The rise of chemotherapy-free regimens, stemming from a deep understanding of oncogenic signaling pathways, marks a pivotal success in this journey. Furthermore, the integration of molecular subtyping into clinical practice, as powerfully demonstrated by the GUIDANCE trials, is transforming genomics from a prognostic tool into an actionable guide for constructing rational, synergistic combinations. By aligning targeted agents with the specific pathway dependencies of the tumor, we can now enhance the efficacy of standard immunochemotherapy backbones, achieving deeper and more durable responses in molecularly defined cohorts.
Despite these remarkable advances, the path toward curative therapies is paved with challenges. Acquired resistance, driven by on-target mutations and the activation of bypass signaling pathways, remains a formidable obstacle. The future of lymphoma therapy will therefore be defined by our ability to meet these challenges with continuous innovation. This includes the development of next-generation agents like noncovalent inhibitors and PROTAC degraders designed to overcome known resistance mechanisms. It also involves the design of intelligent, mechanism-based combination therapies that can preemptively target multiple survival pathways. Ultimately, the convergence of multiomics and AI technologies will be essential to decipher the full complexity of lymphoma signaling networks, identify novel therapeutic vulnerabilities, and validate predictive biomarkers. By continuing to bridge the gap between molecular mechanisms and clinical breakthroughs, the field is poised to move beyond long-term disease control toward the ultimate goal of achieving precision cures for all patients with lymphoma.

Conflicts of interest
None.