Genetic drivers of tumor microenvironment and immunotherapy resistance in non-small cell lung cancer: the role of KEAP1, SMARCA4, and PTEN mutations

Abstract

The tumor microenvironment (TME) plays a crucial role in shaping the response to immunotherapy in non-small cell lung cancer (NSCLC). While immune checkpoint inhibitors (ICIs) have revolutionized NSCLC treatment, a significant proportion of patients exhibit primary or acquired resistance. Emerging evidence highlights the role of specific genetic alterations, Kelch-like ECH-associated protein 1 (KEAP1), SWI/SNF-related, matrix-associated, actin-dependent regulator of chromatin subfamily A member 4 (SMARCA4), and phosphatase and tensin homolog (PTEN) mutations, in driving immune evasion and limiting the effectiveness of ICIs. KEAP1 mutations, through constitutive nuclear factor erythroid 2-related factor 2 activation, promote oxidative stress adaptation and metabolic reprogramming, creating an immune-excluded TME that reduces T-cell infiltration. SMARCA4 loss disrupts chromatin remodeling and epigenetic regulation, impairing antigen presentation and fostering a poorly immunogenic tumor phenotype. PTEN inactivation, by dysregulating the PI3K/AKT pathway, enhances tumor proliferation while contributing to an immunosuppressive cytokine milieu. The presence of these mutations is associated with poor outcomes, independent of tumor mutational burden or programmed death-ligand 1 expression, and their co-occurrence, particularly with KRAS and STK11, further compounds resistance. In this review, we explore the biological impact of KEAP1, SMARCA4, and PTEN mutations on the immune landscape of NSCLC, their implications for immunotherapy resistance, and potential strategies to overcome these barriers. As precision oncology advances, identifying therapeutic vulnerabilities in these genetically defined subgroups will be critical to improving patient outcomes and expanding the efficacy of immunotherapy in NSCLC.

Introduction

Lung cancer remains the leading cause of cancer-related deaths worldwide, despite a recent decline in incidence.¹ Non-small cell lung cancer (NSCLC) comprises about 85% of cases, with adenocarcinoma as the most common subtype.² Prognosis remains poor, with a 5-year survival rate of around 26%.³

NSCLC’s genetic heterogeneity is shaped by mutations in genes such as Kelch-like ECH-associated protein 1 (KEAP1), SWI/SNF-related, matrix-associated, actin-dependent regulator of chromatin subfamily A member 4 (SMARCA4), and phosphatase and tensin homolog (PTEN). KEAP1 and NFE2L2 mutations occur in approximately 20% of lung adenocarcinomas (LUAD) and 25%–30% of lung squamous cell carcinomas (LUSC).⁴ SMARCA4 loss is observed in about 10% of NSCLC cases,⁵ while PTEN alterations appear in 15% of LUSC cases but only 3% of LUAD cases.⁶

These mutations contribute to tumor aggressiveness, treatment resistance, and dysregulation of oxidative stress, chromatin remodeling, and cell signaling.⁷ This review focuses on KEAP1, SMARCA4, and PTEN for their impact on NSCLC biology and treatment response. Unlike oncogenic drivers such as EGFR or ALK, these mutations are not directly actionable but significantly shape immune evasion and resistance.

KEAP1 mutations disrupt nuclear factor erythroid 2-related factor 2 (NRF2) signaling, promoting oxidative stress adaptation, metabolic reprogramming, and immune escape, accelerating tumor progression.⁷ SMARCA4 loss impairs the SWI/SNF chromatin remodeling complex, disrupting DNA repair and promoting tumor plasticity.⁸ PTEN inactivation deregulates phosphoinositide 3-kinase / protein kinase B (PI3K/AKT) signaling enhancing proliferation, immune evasion, and poor outcomes, especially in LUSC.⁹

Immune checkpoint inhibitors (ICIs) have transformed NSCLC treatment,¹⁰ but responses vary due to genetic alterations beyond programmed death-ligand 1 (PD-L1) expression.¹¹ Mutations such as STK11 (LKB1), found in about 10% of LUAD, promote an immunosuppressive tumor microenvironment (TME) via neutrophil infiltration and elevated interleukin (IL))-6, suppressing T-cell activity.¹² These frequently co-occur with KEAP1, further reducing ICI efficacy.¹³ Moreover, mutations in KEAP1, SMARCA4, and PTEN disrupt immune-related pathways, fostering resistance.7,9 Paired-biopsy studies implicate these genes in acquired ICI resistance, showing loss at progression, reduced interferon signaling, and diminished CD8+ T cell infiltration.¹⁴

Given their role in tumor progression and therapy resistance, deeper insight into these mutations is essential for advancing personalized treatments and guiding immunotherapy. This review explores their biological and clinical impact, focusing on the TME, immune evasion, and ICI resistance, with the goal of identifying strategies to overcome therapeutic resistance and improve outcomes for patients with NSCLC (see online supplemental file 1 for a full list of abbreviations).

KEAP1 mutations and immunotherapy resistance

Biological role of KEAP1

KEAP1 is a key regulator of the NRF2 pathway, maintaining cellular homeostasis and oxidative stress response.¹⁵ Under normal conditions, KEAP1 acts as a substrate adaptor for a Cullin 3-based E3 ubiquitin ligase (CUL3), promoting NRF2 ubiquitination and degradation, keeping its levels low to prevent unnecessary activation of antioxidant, detoxification, and metabolic genes.^{16 17} As a sensor of oxidative and electrophilic stress, KEAP1 undergoes cysteine modifications on stress exposure, inhibiting NRF2 degradation.¹⁸ This allows NRF2 to accumulate in the nucleus, bind antioxidant response elements, and activate protective genes that restore redox balance.¹⁹ Through this mechanism, KEAP1 safeguards cellular integrity and prevents oxidative stress-related diseases, including cancer.

KEAP1 mutations disrupt its regulatory function, leading to unchecked NRF2 activation and tumorigenesis.²⁰ Loss-of-function (LOF) mutations impair NRF2 binding or ubiquitin ligase activity, causing its nuclear accumulation independent of oxidative stress.²¹ Persistent NRF2 activation enhances oxidative stress resistance, promotes metabolic adaptability, and drives tumor growth.²² This dysregulation fosters aggressive tumor phenotypes with increased proliferation, survival, and chemotherapy resistance.²³ KEAP1 mutations also contribute to immune evasion, reducing ICI efficacy and correlating with poor outcomes, particularly in LUAD.²⁴ Understanding these mechanisms is essential for improving therapeutic strategies in KEAP1-mutated cancers.

KEAP1 mutations and tumor microenvironment

KEAP1 mutations significantly impact the TME by reducing cytotoxic CD8+ T cell (CTL) infiltration, weakening antitumor immunity.²⁵ This creates an immune-cold environment, marked by low immune effector cell presence and diminished immune activity, making KEAP1-mutant tumors less responsive to ICIs.²⁶

KEAP1 mutations in LUAD suppress CD103+ dendritic cell infiltration and function, impairing T cell activation and promoting exhaustion.²⁷ They also stabilize EMSY, which inhibits the type I interferon response required for innate immunity activation.²⁸ As a result, KEAP1-mutant tumors develop an immune-excluded phenotype with reduced tumor-infiltrating lymphocytes and poor response to ICIs.²⁹ CRISPR-Cas9 studies confirm that NRF2 hyperactivation drives this immune evasion.²⁴ Although KEAP1 mutations increase NK cell infiltration, this does not enhance ICI efficacy, reinforcing their role in immune exclusion rather than inflammation.³⁰

The relationship between PD-L1 expression and KEAP1 mutations is complex. In NSCLC models, KEAP1 promotes PD-L1 ubiquitination and degradation, reducing its expression and enhancing CTL activation.³¹ However, in a cohort of 9243 patients conducted in a Chinese population, NFE2L2 mutations in LUSC correlated with increased PD-L1 expression (p≤0.001), while KEAP1 mutations in LUAD showed a similar association (p≤0.01).³² It is important to note that this study did not adjust for potential enrichment of these mutations in tumors with high tumor mutational burden (TMB) or PD-L1 levels, which may limit generalizability of the findings. Despite this, KEAP1-mutant tumors exhibit poor ICI responses, suggesting resistance mechanisms beyond PD-L1.³³ Persistent NRF2 activation promotes PD-L1 expression and proinflammatory cytokines like IL-6 and chemokine (C-X-C motif) ligand 1, fostering an immunosuppressive TME rather than enhancing antitumor immunity.³³ Figure 1 illustrates both the physiological function of KEAP1 and the effects of KEAP1 mutations on the TME.

Figure 1. The schematic illustrates the role of Kelch-like ECH associated protein 1 (KEAP1) in regulating nuclear factor erythroid 2-related factor 2 (NRF2) and its pathological consequences when mutated. Panel A shows normal KEAP1 function, where it degrades NRF2, preventing excessive antioxidant and metabolic activation. Panel B depicts KEAP1 mutations leading to uncontrolled NRF2 activation, driving tumor survival, metabolic reprogramming, and therapy resistance. It also shows the effects on the tumor microenvironment (TME), where increased myeloid-derived suppressor cells (MDSCs) suppress CD8+ T cell infiltration. Interleukin (IL)-6, chemokine (C-X-C motif) ligand 1 (CXCL1), and transforming growth factor-β secretion increases immune suppression. Additionally, KEAP1 mutations impair antigen presentation by dendritic cells (DCs) due to NRF2-mediated suppression of interferon (IFN) signaling, reducing T cell activation, and contributing to immune evasion and immune checkpoint inhibitor (ICI) resistance. Created with BioRender.com. Used with permission under a BioRender Publication License.

Impact of KEAP1 on immunotherapy outcomes

Although often linked to immunotherapy resistance, KEAP1 mutations may primarily function as a poor prognostic marker.²⁴ Multiple studies report significantly worse survival in patients with KEAP1-mutant NSCLC across treatment settings, including chemotherapy and KRAS-targeted therapy,^{34 35} suggesting a role in both tumor aggressiveness and immune evasion.²⁴ Still, evidence supports a predictive component in ICI-treated patients, particularly with co-mutations. A pan-cancer analysis showed shorter overall survival (OS) in KEAP1-mutant tumors despite higher TMB (10 vs 20 months; p<0.001), and in NSCLC, even high-TMB KEAP1-mutant cases responded poorly to pembrolizumab.^{36 37}

Beyond its general prognostic impact, the clonality of KEAP1 mutations appears to further stratify patient outcomes. A retrospective study identified that clonal KEAP1 mutations with loss of heterozygosity (KEAP1 C-LOH) are particularly detrimental, leading to significantly worse survival.³⁸ In one cohort, KEAP1 C-LOH was associated with a progression-free survival (PFS) HR of 2.54 (95% CI 1.42 to 4.46, p=0.002) and an OS HR of 2.47 (95% CI 1.42 to 4.3, p=0.001), findings that were validated in a second cohort (PFS HR 1.63, 95% CI 1.12 to 2.37, p=0.009; OS HR 1.58, 95% CI 1.05 to 2.4, p=0.03).³⁸ In contrast, clonal diploid or subclonal KEAP1 mutations did not significantly affect survival, indicating that KEAP1 C-LOH may represent a particularly aggressive molecular subtype.

The impact of KEAP1 co-mutations was further underscored in a blood-based next-generation sequencing (NGS) analysis from the OAK/POLAR trials, where patients harboring KEAP1 co-mutations had a significantly higher risk of death (HR 3.36; 95% CI 1.96 to 5.78; p<0.0001) compared with wild-type tumors.³⁹ Even single-mutant KEAP1 tumors had worse outcomes (HR 1.71, 95% CI 1.15 to 2.54; p=0.008). This trend was validated in a tissue-based NGS cohort, where KEAP1 co-mutations consistently predicted poorer survival (HR 1.73, 95% CI 1.17 to 2.57; p=0.006).⁴⁰ Notably, KEAP1 single-mutant tumors demonstrated strong ICI resistance (HR 3.34, 95% CI 1.98 to 5.6; p<0.001).⁴¹

In KRAS-mutant LUAD, KEAP1 co-mutations are linked to significantly worse outcomes. In a cohort of 477 patients, PFS was 1.8 vs 4.6 months (HR 2.05; 95% CI 1.63 to 2.59; p<0.0001), and OS was 4.8 vs 18.4 months (HR 2.24; 95% CI 1.74 to 2.88; p<0.0001) for KEAP1-mutant versus wild-type tumors.⁴⁰ In contrast, KEAP1 mutations had no significant survival impact in KRAS wild-type patients.

KEAP1 mutations, particularly co-occurring with STK11, TP53, and KRAS, have been strongly associated with poor ICI responses and worse prognosis in NSCLC.²⁵ A nationwide clinicogenomic database analysis of 2593 patients demonstrated that patients harboring KEAP1/STK11/KRAS mutations had significantly worse OS across all PD-L1 expression levels (PD-L1 0%: HR 2.73, p<0.001; PD-L1 1%–49%: HR 2.64, p<0.001; PD-L1 ≥50%: HR 2.35, p=0.008).⁴²

Similarly, a retrospective study of 330 patients with KRAS-mutant NSCLC found that those with KRAS/KEAP1 co-mutations had a median OS of just 6 months, while OS was not reached in KEAP1 wild-type cases.⁴³ This pattern held in a larger cohort of 536 patients, where co-mutant tumors showed lower overall response rate (ORR) (18% vs 29%, p=0.02), shorter PFS (1.8 vs 4.6 months; HR 2.05; 95% CI 1.63 to 2.59; p=0.0001) and reduced OS (4.8 vs 18.4 months; HR 2.24; 95% CI 1.74 to 2.88; p=0.0001).⁴³

The frequent co-mutation of KEAP1 and STK11 suggests a potential mechanistic link, as these genes are located in close proximity on chromosome 19p13.2 and 19p13.3, which could facilitate co-deletion or co-mutation events; however, this structural association does not necessarily imply a functional or mechanistic interaction.⁴⁴ Bioinformatic analyses confirm that triple-mutant tumors (KRAS+STK11+KEAP1) confer significantly worse survival outcomes compared with KRAS-only mutations, with a median OS of 12.24 months vs 46.8 months (HR 1.81, p<0.001).⁴⁵ Under ICI therapy, this effect remains pronounced, as triple-mutant patients had a median OS of only 4 months, compared with 12 months in KRAS-only tumors (HR 1.63, p=0.046).⁴⁵

Even with chemo-immunotherapy, KEAP1-mutant NSCLC shows significantly worse outcomes.⁴⁶ In a recent cohort, KEAP1-mutant patients had lower ORR (23.3% vs 41.1%, p<0.001), shorter PFS (3.3 vs 6.7 months; HR 1.53, 95% CI 1.28 to 1.84), and reduced OS (8.1 vs 16.9 months; HR 1.71, 95% CI 1.40 to 2.10).⁴⁶ This trend held across both KRAS wild-type (ORR: 25.0% vs 42.2%; OS: 7.4 vs 17.2 months) and KRAS-mutant subgroups (ORR: 21.2% vs 39.3%; OS: 8.1 vs 15.5 months; all p<0.001). Outcomes were poorest in triple-mutant tumors (KEAP1, SMARCA4, STK11), with ORR of 7.7%, PFS of 1.4 months, and OS of 5.0 months vs 43.3%, 6.1 months, and 17.1 months, respectively (p<0.001).⁴⁶

Cytotoxic T-lymphocyte–associated protein 4 (CTLA-4) blockade has emerged as a promising strategy to overcome KEAP1/STK11-driven resistance to PD-L1 inhibitors. The phase III POSEIDON trial demonstrated that patients with NSCLC harboring KEAP1 and/or STK11 mutations had significantly improved OS with durvalumab+tremelimumab+chemotherapy, reaching 15.8 months compared with 7.3 months with chemotherapy alone (HR 0.50; 95% CI 0.29 to 0.87; p=0.01) and 10.5 months with durvalumab+chemotherapy (HR 0.64; 95% CI 0.40 to 1.04; p=0.04).⁴⁷ Similarly, PFS was 6.2 months in the dual ICI group, significantly longer than 2.8 months with chemotherapy alone (HR 0.52; 95% CI 0.28 to 0.95; p=0.03) and 4.8 months with durvalumab+chemotherapy (HR 0.71; 95% CI 0.43 to 1.17; p=0.06). The ORR was also significantly higher in the dual ICI group (42.9%) compared with 28.0% with chemotherapy alone (p=0.02) and 30.2% with durvalumab+chemotherapy (p=0.04).⁴⁷

Preclinical models have validated these findings, showing that KEAP1-deficient and STK11-deficient tumors create a highly immunosuppressive TME, characterized by CD8+ T cell depletion and myeloid-derived suppressor cell (MDSC) enrichment.⁴⁷ Notably, dual immune checkpoint blockade (ICB) therapy effectively engaged CD4+ effector cells and reprogrammed tumor-associated myeloid cells into inducible nitric oxide synthase-expressing tumoricidal phenotypes, an effect not observed with PD-L1 blockade alone. This immune remodeling correlated with significantly prolonged tumor suppression and increased survival in KEAP1/STK11-deficient mouse models treated with dual ICB, compared with PD-L1 monotherapy (p<0.001).

The co-occurrence of KEAP1 and STK11 mutations worsens ICI responses and emerged as an important determinant of resistance to KRAS^G12C inhibitors such as sotorasib and adagrasib. In a cohort of 424 patients, KEAP1 mutations were linked to shorter PFS (2.8 vs 5.4 months, HR 2.26; 95% CI 1.60 to 3.19; p<0.001) and OS (6.3 vs 11.1 months, HR 2.03; 95% CI 1.38 to 2.99; p<0.001).⁴⁸ STK11 mutations also reduced PFS (4.4 vs 5.5 months, HR 1.32; 95% CI 1.00 to 1.73; p=0.01), although OS differences were not significant (HR 1.18; 95% CI 0.85 to 1.64; p=0.17).⁴⁸ The combination of KEAP1 and STK11 mutations conferred even worse outcomes, with KRAS^G12C/KEAP1/STK11-mutant tumors exhibiting PFS of 2.8 vs 5.3 months (HR 2.30; 95% CI 1.60 to 3.30; p<0.001) and OS of 6.3 vs 10.7 months (HR 2.13; 95% CI 1.41 to 3.20; p<0.001). Similar trends appeared in the KRYSTAL-1 trial, where KEAP1 mutations halved PFS (4.1 vs 9.9 months; HR 2.7; 95% CI 1.6 to 4.7; p<0.01) and OS (5.4 vs 19.0 months; HR 3.6; 95% CI 2.0 to 6.6; p<0.01).³⁵ STK11 mutations were likewise associated with inferior PFS (4.2 vs 11.0 months; HR 2.2; 95% CI 1.3 to 3.6; p<0.01) and OS (9.8 months vs not reached; HR 2.6; 95% CI 1.4 to 4.7; p<0.01).

These findings reinforce that KEAP1 and STK11 co-mutations act as major resistance mechanisms against ICIs and KRAS^G12C inhibitors, likely due to their role in tumor metabolic adaptation, immune evasion, and enhanced NRF2 signaling. These insights emphasize the broader immunosuppressive impact of KEAP1 mutations, supporting the need for novel therapeutic approaches to overcome resistance and improve patient outcomes.

In addition to mutations, KEAP1 gene deletions have also been associated with poor clinical outcomes. In a large multicenter study of patients with NSCLC treated with chemoimmunotherapy, KEAP1 deletions were independently predictive of shorter PFS (HR 1.4, p=0.002) and OS (HR 1.5, p=0.003), even after excluding cases with KEAP1 mutations, suggesting that both mutational and copy number loss may contribute to immune resistance and reduced survival.^{35 49}

Table 1 summarizes ongoing clinical trials in which KEAP1 mutations are a central focus, including studies evaluating immunotherapy, targeted agents, and combination strategies.

Table 1. Clinical trials investigating KEAP1-mutant tumors: immunotherapy and targeted therapy approaches.

NCT number	Phase	Study population	Interventions	Primary outcome(s)	Secondary outcome(s)	Status
NCT06130254	I	Solid tumors with KRAS G12C mutation	Adagrasib+olaparib	AEs (NCI CTCAE V.5.0)	–	Recruiting
NCT05275868	I	NSCLC with or without KEAP1/NFE2L2/CUL3 mutation	MGY825 monotherapy	AEs; SAEs; DLTs	PK (AUC, Cmax, Tmax); ORR, PFS, DOR	Recruiting
NCT05996263	–	Advanced NSCLC with PD-L1 TPS ≥50%	Pembrolizumab	PFS	OS	Recruiting
NCT06494540	–	NSQ-mNSCLC with KRAS/STK11/KEAP1/TP53 mutations	Tremelimumab+durvalumab+chemo	rwOS	rwOS; mOS; rwPFS	Recruiting
NCT02417701	II	Stage IV or recurrent NSCLC with KEAP1/NFE2L2/KRAS mutation	Sapanisertib	ORR; DOR	PFS; RPPA feasibility	Completed
NCT05276726	I/II	NSCLC with KRAS G12C and STK11 mutation (KEAP1 WT)	JAB-21822	DLTs; AEs; ORR	PFS; DOR; OS; TTR; DCR	Recruiting
NCT06341660	II	Advanced or recurrent NSCLC with KEAP1 mutation	Carbonizumab+chemo	ORR; safety	–	Recruiting
NCT06335355	II/III	Advanced NSCLC with STK11/KEAP1/KRAS mutations	Adebrelimab+SHR-8068+chemo	ORR (phase I); OS (phase II)	AEs; PFS; DOR; DCR; OS at 3.5 years	Not yet recruiting
NCT06008093	III	Metastatic NSCLC with STK11/KEAP1/KRAS mutations	Durvalumab+tremelimumab+chemo versus pembrolizumab+chemo	OS (overall and genetic subgroups)	PFS; ORR; DOR; TFST; AEs	Recruiting

AEs, adverse events; AUC, area under the curve; Cmax, maximum plasma concentration; DCR, disease control rate; DLTs, dose-limiting toxicities; DOR, duration of response; mOS, median overall survival; NCI CTCAE V.5.0, National Cancer Institute Common Terminology Criteria for Adverse Events, V.5.0; NSCLC, non-small cell lung cancer; ORR, objective response rate; OS, overall survival; PFS, progression-free survival; PK, pharmacokinetics; RPPA, reverse phase protein array; rwOS, real-world overall survival; rwPFS, real-world progression-free survival; SAEs, serious adverse events; TFST, time to first subsequent therapy; Tmax, time to maximum plasma concentration; TTR, time to response; WT, wild-type.

SMARCA4 mutations and immunotherapy resistance

Role of SMARCA4 in chromatin remodeling

SMARCA4 (BRG1) is an ATPase subunit of the SWI/SNF chromatin remodeling complex, which regulates chromatin structure to control gene expression.⁵⁰ This process is essential for transcription, replication, DNA repair, and cell cycle regulation.⁵¹ Using ATP hydrolysis, SMARCA4 repositions, ejects, or restructures nucleosomes, modulating DNA accessibility for transcription factors and regulatory proteins.⁵² As a critical component of SWI/SNF, SMARCA4 influences both local and higher-order chromatin organization by binding super-enhancers and topologically associating domain boundaries, which regulate long-range gene interactions.⁵³ It also interacts with acetylated histones through its bromodomain, particularly H3K14ac, facilitating transcriptional activation.⁵⁴ Under normal conditions, SMARCA4 functions as a tumor suppressor, maintaining genomic stability by remodeling chromatin at DNA break sites to facilitate repair.⁵⁵ This prevents genomic instability, which can drive tumorigenesis. In NSCLC, SMARCA4 is inactivated in 5%–10% of cases due to mutations, deletions, or epigenetic silencing.⁵⁶ Its loss disrupts SWI/SNF chromatin remodeling, leading to widespread gene expression changes that suppress tumor suppressors and activate oncogenic pathways. This promotes unchecked proliferation, apoptosis resistance, and increased tumor aggressiveness.⁵⁷ Additionally, a meta-analysis linked high SMARCA4 expression to poor prognosis, advanced tumor stages, and increased metastasis in LUAD.⁵⁸

SMARCA4 mutations and TME

SMARCA4 mutations shape the TME through epigenetic reprogramming and immune modulation, fostering an immunosuppressive environment that drives resistance to conventional and immune therapies.⁵⁹ SMARCA4 loss disrupts autophagy, leading to reactive oxygen species accumulation, weakening the stress response, accelerating tumor progression, and promoting immune evasion.⁵⁹ In NSCLC, SMARCA4-deficient tumors exhibit reduced immune cell infiltration, with significantly lower levels of dendritic cells and CD4+ T cells, both essential for antitumor immunity.⁶⁰ A recent study demonstrated that SMARCA4 loss triggers epigenetic reprogramming, downregulating innate immune signaling genes such as Stimulator of interferon genes (STING), IL-1β, and type I interferons, reinforcing the immunosuppressive microenvironment.⁶¹ Despite their immune-cold nature, some SMARCA4-deficient tumors contain tertiary lymphoid structures, which may indicate better immunotherapy responses.⁶²

The immunosuppressive nature of SMARCA4-deficient tumors is reinforced by Enhancer of Zeste Homolog 2 (EZH2) upregulation, an epigenetic regulator that limits CTL activity and promotes immune evasion.⁶³ EZH2 inhibition can partially reverse these effects by enhancing CTL function and reducing infiltration of regulatory T cells (Tregs).⁶⁴ Additionally, SMARCA4-deficient tumors exhibit increased FOXP3+ Tregs and neutrophils, further suppressing antitumor immunity.⁶⁵ This dual mechanism, reducing innate immune activation while recruiting immunosuppressive cells, allows these tumors to evade immune surveillance.⁶⁶ Beyond chromatin remodeling, SMARCA4 loss disrupts immune regulation by impairing nuclear factor-kappa B signaling, limiting its ability to regulate immune-related genes and contributing to an immune-cold phenotype.⁶¹ SMARCA4 also influences tumor metabolism, particularly de novo lipogenesis, supporting proliferation and survival under nutrient-deprived conditions.⁶⁷ This metabolic reprogramming, coupled with immune evasion, fosters a therapy-resistant TME.⁶⁸ Figure 2 illustrates both the physiological function of SMARCA4 and the effects of loss of SMARCA4 on the TME.

Figure 2. The schematic illustrates the role of SWI/SNF-related, matrix-associated, actin-dependent regulator of chromatin subfamily A member 4 (SMARCA4) in chromatin remodeling and immune regulation, along with its pathological consequences when lost. Panel A shows normal SMARCA4 function as part of the SWI/SNF complex, where it maintains open chromatin, allowing transcription of key immune-related genes such as STING, interleukin (IL)-1β, and type I interferons (IFN-α/β). Transcriptional regulation involves MYC (MYC proto-oncogene), E2F (E2F family of transcription factors), SOX (SRY-box family of transcription factors), and OCT (octamer-binding transcription factors, such as OCT4). This promotes antitumor immune responses, antigen presentation, and CD8+ T cell activation in the tumor microenvironment (TME). Panel B depicts SMARCA4 loss leading to chromatin compaction and transcriptional repression, silencing immune pathways and driving an immune-cold TME. Reduced STING signaling, type I IFN production, and nuclear factor kappa B (NF-κB) activation impair immune cell recruitment, while increased Enhancer of Zeste Homolog 2 (EZH2) expression promotes an immunosuppressive phenotype. The result is diminished dendritic cell function, decreased CD8+ T cell infiltration, and increased regulatory T cells (Tregs) and neutrophils, leading to immune evasion and resistance to immunotherapy. NK, natural killer. Created with BioRender.com. Used with permission under a BioRender Publication License.

SMARCA4 implications in immunotherapy

SMARCA4 mutations in NSCLC are linked to poor outcomes, particularly in patients treated with ICIs. In a combined analysis of two large cohorts (n=675 and 848), SMARCA4 alterations in LUAD were associated with significantly worse OS (p=0.0348), PFS (p=0.0387), and disease-specific survival (p=0.0147).⁶⁹ In NSCLC specifically, LOF SMARCA4 mutations were associated with an adjusted HR of 1.44 (95% CI 1.06 to 1.96) for OS, which increased to 1.57 (95% CI 1.13 to 2.17) after adjusting for TMB. These findings raise the possibility that, contrary to some preclinical expectations, SMARCA4 mutations may promote immune evasion and resistance to ICI.

However, the role of SMARCA4 in immunotherapy outcomes remains complex and appears to depend on both mutation subtype and broader genomic context. In a retrospective study, ICI therapy was found to improve survival in SMARCA4-altered NSCLC (HR 0.67; p=0.01), despite low PD-L1 expression.⁵⁶ Notably, class 1 SMARCA4 mutations (truncating mutations, fusions, or deletions) correlated with a higher TMB and an improved ORR (p=0.027), although these initial responses did not translate into long-term benefit, as PFS (p=0.74) and OS (p=0.35) were comparable to wild-type tumors. These mixed results suggest that while SMARCA4 alterations may confer transient ICI sensitivity, other factors may limit durable benefit.

Subsequent studies indicate that co-occurring genomic alterations likely drive these differences in outcomes. In one study, patients with SMARCA4-mutant NSCLC who lacked co-occurring STK11 or KEAP1 mutations demonstrated better PFS (p=0.04), a significantly higher ORR (52.63% vs 20.22%, p=0.001), and a greater durable clinical benefit (55.56% vs 32.45%, p=0.045) than SMARCA4 wild-type counterparts, suggesting that SMARCA4 mutations alone may not necessarily predict resistance to ICIs.⁷⁰ However, metabolically defined redox-high LUAD tumors, marked by high reductase-oxidative activity and enriched for KEAP1, STK11, SMARCA4, and NRF2 mutations, exhibit a profoundly immunosuppressive TME, reduced tissue-resident memory CD4+ T cells and impaired CTL function.⁷¹ Patients with redox-high LUAD experienced significantly lower ORR, shorter PFS (3.3 vs 14.6 months, p=0.004), and reduced OS (12.1 vs 31.2 months, p=0.022) following ICI treatment compared with redox-low tumors. Their findings suggest that the presence or absence of specific co-mutations could alter the immune phenotype and therapeutic responsiveness of SMARCA4-mutant NSCLC.

Some of the variability in reported outcomes for SMARC4-mutant NSCLC may be explained by differences in study populations, particularly with respect to KRAS mutation status. In analyses focused specifically on KRAS-mutant NSCLC, co-occurring SMARCA4 mutations have been associated with markedly worse outcomes. One retrospective study found that patients with KRAS-SMARC4 co-mutations had a lower ORR (0% vs 22%), shorter PFS (1.4 vs 4.1 months, p<0.001), and shorter OS (3.0 vs 15.1 months, p<0.001) compared with those with KRAS-mutant/SMARCA4 wild-type tumors.⁷²

A separate multicohort analysis of KRAS-mutant LUAD evaluated the prognostic impact of SMARCA4 mutations across both ICI-treated and non-ICI-treated.⁷³ In ICI-treated cohorts, SMARC4 co-mutations were associated with significantly reduced PFS and OS. Specifically, PFS was shorter in SMARCA4-mutant tumors compared with other KRAS-mutant subgroups with HRs ranging from 2.15 to 3.06 (p=0.0026–0.0091) and OS was similarly reduced (HR 2.46, p=0.0014). In non-ICI-treated cohorts, SMARCA4 mutations were also linked to worse outcomes. Compared with KRAS-only and KRAS-TP53-mutant patients, KRAS-SMARCA4 co-mutations were associated with significantly shorter OS (HR 2.46 vs KRAS-TP53, p=0.0019; HR 2.46 vs KRAS-only, p=0.042). When comparing KRAS-SMARCA4 co-mutation patients with the combined KRAS wild-type group, the effect was even more pronounced: OS HR 11.98 (p=0.0018) and PFS HR 18.7 (p=0.0011). Median survival analyses reinforced this disparity: patients with KRAS-SMARCA4 co-mutations had a median OS of 1.73 months, compared with 2.77 months in those with KRAS-TP53 co-mutations and 6.45 months in those with KRAS-only mutations. Similarly, median PFS was 1.73 months in the KRAS-SMARCA4 group vs 4.22 months in both the KRAS-only and KRAS-TP53 groups. These findings suggest that SMARCA4 mutations contribute to poor survival outcomes across treatment settings, driving both immunotherapy resistance and a worse prognosis in non-ICI-treated patients.

In addition to mutation inactivation, SMARCA4 gene deletions have emerged as a clinically meaningful alteration. SMARCA4-deleted tumors showed significantly lower ORR (29% vs 45%, p=0.0007), shorter PFS (HR 1.6, p<0.0001), and worse OS (HR 1.7, p<0.0001), highlighting the prognostic relevance of both genetic loss and mutational inactivation.⁴⁹

Beyond ICI resistance, SMARCA4 mutations also appear to impact responses to KRAS^G12C inhibitors. A study investigating KEAP1 co-alterations in patients treated with KRAS^G12C inhibitors found that SMARCA4 mutations were associated with markedly worse prognosis.⁴⁸ Patients with SMARCA4 mutations had a significantly shorter PFS (1.6 vs 5.4 months, HR 3.04; 95% CI 1.80 to 5.15; p<0.001) and OS (4.9 vs 11.8 months, HR 3.07; 95% CI 1.69 to 5.60; p<0.01) compared with SMARCA4 wild-type tumors. These outcomes indicate that SMARCA4 mutations influence prognosis beyond immunotherapy resistance, potentially affecting responses to KRAS^G12C inhibitors. Additional studies are needed to understand the broader clinical impact of SMARCA4 mutations, particularly in combination with targeted therapies. Table 2 summarizes ongoing trials focused on SMARCA4-mutant tumors, including immunotherapy and targeted approaches.

Table 2. Clinical trials investigating SMARCA4-altered tumors: targeted and combination therapy approaches.

NCT number	Phase	Study population	Interventions	Primary outcome measures	Secondary outcome measures	Status
NCT05639751	I	Advanced/metastatic solid tumors with SMARCA4 mutation	PRT3789 monotherapy or combined with docetaxel (SMARCA2 degrader)	DLTs, AEs, MTD/RP2D (3 years)	ORR, PFS, CBR, DOR, PK profile (3 years)	Recruiting
NCT06560645	I	Advanced/metastatic solid tumors with SMARCA4 mutation	PRT7732 (oral SMARCA degrader)	DLTs, AEs, MTD, RDE (2 years)	ORR, PK/PD profile (2 years)	Recruiting
NCT06682806	II	Advanced/metastatic solid tumors with SMARCA4 mutation	PRT3789+pembrolizumab	DLTs, AEs (2 years), ORR and DOR (2 years)	ORR, DOR, PFS, OS, CBR (2 years)	Not yet recruiting
NCT06561685	I	Advanced/metastatic solid tumors with SMARCA4 alterations	LY4050784 monotherapy	DLTs, AEs, MTD/RP2D, ORR	ORR, DOR, TTR, DCR, PFS, PK profile (4 years)	Recruiting

AEs, adverse events; CBR, clinical benefit rate; Cmax, maximum plasma concentration; DCR, disease control rate; DLTs, dose-limiting toxicities; DOR, duration of response; MTD, maximum tolerated dose; ORR, objective response rate; OS, overall survival; PD, pharmacodynamics; PFS, progression-free survival; PK, pharmacokinetics; RP2D, recommended phase II dose; SAEs, serious adverse events; TTR, time to response.

PTEN mutations and immunotherapy resistance

PTEN activity in human

PTEN, located on chromosome 10, is one of the most frequently inactivated tumor suppressors in cancer.⁶ PTEN loss can result from genetic alterations or transcriptional and post-transcriptional suppression, with even partial reduction driving tumorigenesis.⁷⁴ As a lipid phosphatase, PTEN converts PIP3 to PIP2, suppressing PI3K/AKT signaling to regulate cell proliferation, survival, and metabolism.⁷⁵ However, its prognostic and predictive role remains unclear due to the lack of reliable activity assessment methods.⁷⁶ Emerging evidence suggests PTEN loss affects cancer cell behavior and alters the TME, impacting immune cell composition and function.⁷⁷

PTEN role in tumor TME

Beyond its tumor-intrinsic effects, PTEN shapes the TME by suppressing PI3K/AKT signaling, which regulates cell adhesion, migration, and invasion.⁷⁸ PTEN loss leads to AKT hyperactivation, increased aerobic glycolysis, and enhanced proliferation, survival, and metastasis.⁷⁸ Conversely, PTEN restoration in lung cancer cells suppresses oncogenesis by inhibiting proliferation and inducing cell cycle arrest via SKP2 downregulation.⁷⁹ Given these effects, PTEN loss is increasingly being explored as a predictive biomarker for immunotherapy response.⁸⁰ PTEN also modulates immune responses by negatively regulating immunosuppressive cytokines (IL-10, C-C motif chemokine ligand 2, vascular endothelial growth factor) and IL-2 receptor signaling, which controls Treg activity.⁸¹ Its loss leads to increased Tregs, T helper type 2 cells, M2 macrophages, and MDSCs, while reducing memory T cells and Th1 cells, fostering an immunosuppressive TME.⁸² In B cells, PTEN regulates proliferation, activation, and survival, and its loss is linked to immune dysfunction and autoimmunity.⁸¹ In addition, PTEN-deficient tumors upregulate CXCL1, granulocyte colony-stimulating factor, and IL-23, impairing inflammatory and stromal cell activation and promoting PI3K signaling, driving tumor progression.⁸³ Figure 3 illustrates both the physiological function of PTEN and the effects of loss of PTEN on the TME.

Figure 3. The schematic illustrates the role of phosphatase and tensin homolog (PTEN) in regulating PI3K/AKT signaling and its pathological consequences when lost. Panel A shows normal PTEN function, where it dephosphorylates PIP3 to PIP2, inhibiting PI3K/AKT signaling to regulate cell growth, metabolism, and immune balance. This maintains controlled tumor proliferation, prevents metabolic reprogramming, and supports an active immune environment with CD8+ T cells and memory T cells. Panel B depicts PTEN loss leading to hyperactivated PI3K/AKT signaling, driving uncontrolled tumor growth, metabolic shifts (Warburg effect), and immune evasion. It also shows the effects on the tumor microenvironment (TME), where increased regulatory T cells (Tregs), myeloid-derived suppressor cells (MDSCs), and M2 macrophages suppress antitumor immunity. Elevated interleukin (IL)-10, C-C motif chemokine ligand 2 (CCL-2), vascular endothelial growth factor (VEGF), chemokine (C-X-C motif) ligand 1 (CXCL1), and IL-23 further enhance immune suppression, while reduced CD8+ T cell infiltration and antigen presentation contribute to immunotherapy resistance. Created with BioRender.com. Used with permission under a BioRender Publication License.

Clinical implications of PTEN

Preclinical studies show that PTEN loss accelerates tumor growth, migration, and invasion. PTEN knockout cells exhibited higher metastatic potential and epithelial-mesenchymal transition in nude mice.⁸⁴ In NSCLC, PTEN loss frequently co-occurs with cyclin D1 (CCND1) overexpression, a known oncogene. In a cohort of 30 NSCLC samples, 44.8% harbored PTEN LOF mutations.⁸⁵ In patients with both CCND1 overexpression and hypermethylated PTEN, OS was significantly shorter at 6.1 months, compared with 49.4 months in patients with hypermethylated PTEN alone (p=0.026) and 61.5 months in those with CCND1 overexpression alone (p=0.033). The combined presence of these alterations was associated with worse survival outcomes (HR 2.94; 95% CI 1.18 to 7.31). Moreover, co-inactivation of TP53 and PTEN has been linked to increased genomic instability, lymph node invasion, and worse survival outcomes.⁸⁶

PTEN loss is thought to influence immunotherapy response in NSCLC by reshaping the tumor microenvironment and disrupting immune homeostasis. In cellular models and a cohort of 404 patients, PTEN loss was significantly associated with PD-L1 overexpression, poor OS, and advanced stage.⁸⁷ Conversely, anti-PD-L1 treatment increased PTEN levels and suppressed tumor growth in pancreatic cancer and liver metastases in mouse models.⁸⁸ Clinically, PTEN loss has been linked to ICI resistance even in high PD-L1-expressing tumors, as seen in a metastatic NSCLC case where ICI failure was overcome by mammalian target of rapamycin inhibition.⁸⁹ This paradoxical resistance to programmed cell death protein 1 (PD-1) blockade despite high PD-L1 levels is frequently observed in PTEN-deficient tumors.⁹⁰ In NSCLC, PTEN mutations were found exclusively in non-responders to CTLA-4/PD-1 blockade (6.8%).⁹¹ However, in PTEN-deficient mouse models, combining dasatinib (a Src kinase inhibitor) with CTLA-4 blockade synergistically reduced MDSCs and Tregs while increasing CTL infiltration, suggesting a potential strategy to overcome immune resistance.⁹²

An analysis of 175 NSCLC samples found PTEN deletions in 45.3% of LUSC cases (heterozygous) and 10.3% (homozygous), while LUAD had 27.8% heterozygous and 1% homozygous deletions. In both subtypes, PTEN mRNA levels were significantly lower than in wild-type tumors.⁹³ Using CRISPR-Cas9 to silence PTEN in a LUSC cell line responsive to anti-PD-1 and anti-CTLA-4 therapy, PTEN-deficient cells exhibited reduced treatment response in mouse models.⁹³ Similarly, in a study of 39 patients with metastatic melanoma treated with ICIs (pembrolizumab and nivolumab), tumors with PTEN expression showed significantly greater tumor size reduction than PTEN-deficient tumors (p=0.029).⁹⁴ An analysis of 135 resected stage IIIB/C melanoma metastases revealed significantly lower CTL infiltration in PTEN-deficient tumors compared with PTEN-expressing tumors (p<0.001).⁹⁴

Commonalities and divergences across KEAP1, SMARCA4, and PTEN

Although KEAP1, SMARCA4, and PTEN have often been studied independently, emerging evidence suggests that these alterations represent biologically distinct yet convergent mechanisms of immune resistance in NSCLC, influenced by histological subtype, co-mutation patterns, and smoking exposure.⁹⁵ KEAP1 and SMARCA4 mutations frequently co-occur with KRAS mutations in LUAD, potentially due to shared localization on chromosome 19p13 and overlapping selection pressures during tumor evolution in smokers.^{49 96} KEAP1 loss activates the NRF2 pathway, driving oxidative stress resistance, metabolic reprogramming, and reduced immune activation,²⁰ while SMARCA4 inactivation impairs chromatin remodeling, promotes lineage plasticity, and suppresses antigen presentation.⁵⁰ Both are independently associated with poor prognosis, higher rates of distant metastases, and limited benefit from ICI, even in tumors with high PD-L1 or TMB.^{31 70 96}

In contrast, PTEN mutations are more commonly observed in LUSC, often arising in tumors that lack canonical oncogenic drivers.⁹³ PTEN loss leads to activation of the PI3K/AKT pathway and contributes to immune evasion through altered signaling and reduced T cell infiltration.⁷⁵ These divergent patterns highlight two distinct evolutionary pathways of immune escape in NSCLC: one driven by co-alterations of KRAS, KEAP1, and SMARCA4 in LUAD, and another characterized by PTEN and KEAP1 in squamous histologies, often in the context of high mutational burden and smoking-related genomic damage.^{48 73 97} Recognizing these context-specific mutation profiles is critical for refining prognostic frameworks and developing tailored immunotherapy strategies.

Still, the clinical impact of these alterations may extend beyond their genomic classification. KEAP1, SMARCA4, and PTEN alterations have each been implicated in both primary and acquired resistance to ICI, emphasizing their dynamic roles in tumor-immune interactions.^{14 98} For instance, copy number deletions of KEAP1 and SMARCA4, regardless of coding mutations, have been associated with poor outcomes in patients with NSCLC treated with immunotherapy.⁴⁹ These findings highlight the limitations of relying solely on DNA-based profiling to infer functional loss. Many tumors may harbor mutations or shallow deletions that do not result in complete inactivation, while others may exhibit protein loss through epigenetic or post-transcriptional mechanisms not detectable by sequencing.

In this context, incorporating protein-level assessments can offer critical functional insight.⁹⁹ For example, immunohistochemistry (IHC) for BRG1 (the protein product of SMARCA4) can confirm loss of chromatin remodeling activity, while PTEN IHC may reveal protein absence even in cases without clear truncating mutations.^{55 100} Similarly, expression of downstream NRF2 targets such as NQO1 and GCLC can serve as surrogate markers of KEAP1 pathway activation, providing functional evidence of pathway dysregulation even in the absence of clearly pathogenic mutations.¹⁵ These protein-based readouts provide direct evidence of pathway silencing or activation, enhance biomarker specificity, and may better correlate with immunosuppressive tumor phenotypes and clinical outcomes. As interest grows in biomarker-driven immunotherapy strategies, integrating genomic and proteomic data will be essential to more accurately stratify patients and guide treatment beyond PD-1/PD-L1 blockade, particularly in targeting NRF2-driven metabolism, chromatin remodeling vulnerabilities, or STING pathway suppression.

Conclusion

The treatment landscape of NSCLC is increasingly shaped by tumor-intrinsic mutations that drive both progression and resistance to therapy. Alterations in KEAP1, SMARCA4, and PTEN are key mediators of immunotherapy resistance, each disrupting pathways central to oxidative stress response, chromatin remodeling, or immune regulation. KEAP1 mutations activate NRF2, promoting immune exclusion; SMARCA4 loss impairs chromatin remodeling, dampening anti-tumor immunity; and PTEN inactivation enhances PI3K/AKT signaling, facilitating immune escape. Their co-occurrence with other oncogenic drivers, especially in KRAS-mutant tumors, further exacerbates resistance and worsens outcomes.

Despite their negative prognostic implications, emerging data support tailored strategies, such as dual ICI, metabolic targeting, or KRAS-directed combinations, to overcome resistance. Advancing precision oncology will require deeper understanding of these alterations to refine biomarker-driven therapies and broaden the reach of immunotherapy in NSCLC.

Data availability statement

All data generated or analyzed during this study are included in this published article and its supplementary materials.