This study included 15 patients with SMARCA4 mutations. The study retrospectively investigated patients who underwent high-throughput lung cancer sequencing in the Pathology Department of Hunan Provincial People's Hospital (Changsha, China) from January 2021 to November 2024. Among them, there were 9 male patients and 6 female patients, with an age range of 34 to 84 years, and an average age of 60.5 years. The inclusion criteria for the study subjects are as follows: i) Patients who underwent high-throughput lung cancer sequencing in the Pathology Department of Hunan Provincial People's Hospital (Changsha, China) from January 2021 to November 2024; ii) patients diagnosed with lung cancer through histopathology or cytology; iii) patients whose high-throughput sequencing results indicated SMARCA4 gene mutations; iv) patients with complete clinical and pathological data, including sex, age, pathological type, stage, treatment and follow-up information; v) age ≥18 years, with traceable clinical data. The exclusion criteria are as follows: i) Patients with a history of other malignant tumors or who have received other radical treatments for tumors in the past; ii) patients who did not undergo standardized high-throughput sequencing testing or whose sequencing results did not meet the standards; iii) patients with severely missing clinical data that cannot be used for statistical analysis; iv) patients with non-primary lung cancer (such as metastatic cancer). This study focused on analyzing the specific information of SMARCA4 mutations, including nucleotide variations and amino acid sequence differences. This study was approved by the Ethics Committee of Hunan Provincial People's Hospital (approval no. 2024-279).

Genomic DNA was extracted from tumor tissue and subjected to next-generation sequencing using probe capture technology, following the procedures described below. First, the extracted genomic DNA was fragmented, and the fragmented DNA was subjected to end repair to blunt the broken ends. Subsequently, an ‘A’ tail was added to the 3′ end of the repaired DNA fragments. After that, sequencing adapters were ligated to the A-tailed DNA fragments, followed by purification to remove unligated adapters and impurities. Next, PCR amplification enrichment was performed using the Pumaikangduo Gene Detection Kit (Nanjing Shihe Gene Biotechnology Co., Ltd.). The PCR products were purified to construct the genomic DNA pre-library. To obtain the targeted library, the pre-library was hybridized and enriched with a DNA probe (cat. no. CM1195; Nanjing Vazyme Biotechnology Co., Ltd.). The quality and integrity of the processed samples (pre-library and targeted library) were verified using a Bioanalyzer 2100 (Agilent Technologies, Inc.), which was used to assess the fragment size distribution and integrity of the libraries; a high-quality library was defined as having a single, sharp peak without obvious smearing, indicating uniform fragment size and good integrity. The concentration of the Equalbit 1× dsDNA HS working solution was measured using a Qubit 4.0 Fluorometer (Thermo Fisher Scientific, Inc.). Library quantification was conducted using the Library Quantification Kit (cat. no. CM1195; Nanjing Vazyme Biotechnology Co., Ltd.), and the final loading concentration of the targeted library was adjusted to 400 pM. The concentration was converted from mass concentration (measured by Qubit 4.0 Fluorometer) to molar concentration using the formula: Molar concentration (pM)=(Mass concentration (ng/µl) ×10⁶)/(Average library fragment length (bp) ×660 g/mol), where the average fragment length of the targeted library was ~300 bp. High-throughput sequencing was performed on the Illumina NextSeq 550Dx sequencing platform (Illumina, Inc.), using paired-end sequencing with a nucleotide length of 150 bp (150 bp per read for both forward and reverse strands). Data analysis was conducted using the built-in BaseSpace Sequence Hub software (version 5.14.0; Illumina, Inc.), which was used for raw data quality control, read alignment, variant calling and data visualization.

The types of driver genes detected can be divided into two categories: i) Driver mutations that promote tumor proliferation and provide selective growth advantages; and ii) passenger mutations with minimal or no impact on tumor proliferation and metastasis (15). Based on the annotation results, genes were evaluated using mutation frequency-based, functional impact-based, structural genomics-based, and network or pathway-based approaches.

The tumor tissues used in this study were previously paraffin-embedded and stored after collection. The specific paraffin-embedding process was as follows: Freshly isolated tumor tissues were first fixed to preserve cellular morphology, then dehydrated through a gradient ethanol series, cleared with xylene, embedded in paraffin wax to form tissue blocks, and stored at 4°C until sectioning (16). Prior to sectioning, the paraffin-embedded tissue blocks were sectioned into slices with a thickness of 4 µm. The histological staining procedures were performed as follows: To remove the paraffin and hydrate the slices, paraffin slices were sequentially immersed in xylene (twice, 5 min each), followed by anhydrous ethanol (5 min), 95% ethanol (5 min), and 70% ethanol (5 min). Finally, they were washed with distilled water (3 times, 2 min each). The fresh tumor tissues were fixed using 10% neutral buffered formalin (fixative concentration: 10%, v/v) at room temperature (25°C) for 24 h before paraffin embedding, which was essential to maintain the integrity of tissue structure and cellular components. The sections were then immersed in the hematoxylin staining solution at room temperature (25°C) for 5 to 10 min. Then, they were differentiated using 0.5% hydrochloric acid-ethanol solution (hydrochloric acid concentration: 0.5%, v/v; ethanol concentration: 70%, v/v) for 30 sec, and 1% ammonia solution (concentration: 1%, v/v) was added dropwise to neutralize the residual acid, thereby making the cell nucleus appear blue-purple. The cytoplasm and matrix were stained pink by immersing the slices in eosin staining solution for 1 to 2 min. After dehydrating the slices in 70% ethanol (5 min), 95% ethanol (5 min), and anhydrous ethanol (twice, 5 min each) in order, they were soaked for 5 to 10 min each in xylene (twice) to attain transparency. After applying neutral gum to the slices and sealing them with a cover glass, the slices were dried in an oven set to 37°C for 2 h. The stained sections were placed under an Olympus (Olympus Corporation) optical microscope for observation. The cellular and tissue morphological characteristics of smarca4-deficient lung cancer were examined.

Immunohistochemical staining was performed following standard protocols. To dewax and hydrate, paraffin slices (4 µm) were successively immersed in xylene (twice, 5 min each, at room temperature, 25°C) (xylene, analytical grade; Sinopharm Chemical Reagent Co., Ltd.), anhydrous ethanol (5 min, room temperature) (anhydrous ethanol, 99.7% v/v, analytical grade), 95% ethanol (5 min, room temperature) (95% v/v ethanol, prepared by diluting anhydrous ethanol with distilled water), and 70% ethanol (5 min, room temperature) (70% v/v ethanol, prepared by diluting anhydrous ethanol with distilled water). Finally, they were rinsed with PBS buffer (0.01 M, pH 7.4, containing 0.05% Tween-20; Beyotime Biotechnology) three times for 2 min each. Slices were heated to 95°C in citrate buffer (0.01 M, pH 6.0) using a water bath, held for 15 to 20 min for antigen retrieval (to expose hidden antigen epitopes), allowed to cool naturally to room temperature (~25°C, ~30 min), and rinsed with PBS buffer (0.01 M, pH 7.4, 0.05% Tween-20) three times, for 2 min each. After 10 to 20 min of sealing the slices with a sealing solution [5% BSA (cat. no. A7906; MillaporeSigma) dissolved in PBS buffer, 0.01 M, pH 7.4] at room temperature to prevent non-specific binding, the primary antibody (anti-Ki-67 rabbit monoclonal antibody; cat. no. ab15580; 1:100; Abcam) was applied dropwise (50 µl per slice, covering the entire tissue section) and incubated overnight at 4°C. The slices were washed with PBS buffer (0.01 M, pH 7.4, 0.05% Tween-20) three times, 5 min each, to remove unbound primary antibody. The secondary antibody (goat anti-rabbit IgG H&L HRP; cat. no. ab6721; 1:200; Abcam) was applied dropwise (50 µl per slice) and incubated at 37°C for 30 to 60 min. Next, the DAB solution (Dako; Agilent Technologies, Inc.) was applied dropwise (50 µl per slice), and the slices were observed under a light microscope (Olympus BX53; Olympus Corporation) at ×100 magnification to monitor the color development. After 5 to 10 min (until the positive signal showed a light brown color, avoiding over-coloration), the reaction was stopped with distilled water (room temperature) by rinsing three times, 1 min each, and hematoxylin (cat. no. C0107; Beyotime Biotechnology) was used as a light counterstain for the cell nucleus (stained at room temperature for 30 sec). Slices were dehydrated using 70% ethanol (5 min, room temperature), 95% ethanol (5 min, room temperature) and anhydrous ethanol (twice, 5 min each, room temperature) in order, and xylene (twice, 5 min each, room temperature) was used to make them transparent. Finally, a neutral gum (Sinopharm Chemical Reagent Co., Ltd.) was applied dropwise (3–5 µl per slice) for sealing, and the slices were dried at 37°C in an oven for 2 h before microscopic observation.

GEPIA2 (http://gepia2.cancer-pku.cn/) is an online tool for gene expression analysis based on the TCGA and GTEx databases. In the present study, GEPIA2 was used to analyze the differential expression of target genes between tumor and normal tissues. The corresponding cancer dataset was selected, with tumor tissues as the experimental group and normal tissues as the control group, and box plots were used to display the expression distribution. In the survival analysis module, patients were divided into high- and low-expression groups using the median expression value as the cut-off. Survival curves were plotted by the Kaplan-Meier method, and survival differences were evaluated using the Log-rank test. In addition, the correlation analysis module was applied to explore the expression correlation between genes using Pearson correlation coefficient to screen co-expressed genes. P<0.05 was considered to indicate a statistically significant difference.

All experimental data were presented as mean ± standard deviation and each experiment was independently repeated three times to ensure the reliability and reproducibility of the experimental results. Statistical tests were selected based on the experimental design: Unpaired two-tailed t-test was used for comparison between two independent groups, while one-way analysis of variance followed by Tukey's post-hoc test was employed for multiple group comparisons. All statistical analyses were performed using SPSS software (version 26.0; IBM Corp.). P<0.05 was considered to indicate a statistically significant difference.

Between January 2021 and November 2024, 15 patients with NSCLC and SMARCA4 mutations were identified. All the 15 patients in this study were from Hunan Province, China, and were diagnosed with SMARCA4-deficient NSCLC by pathological examination. The likelihood of lung cancer developing in the left and right lungs was equal, and there was no discernible pattern in the location of cases. In general, the tumor was large, measuring >5 cm in diameter. The symptoms of SMARCA4-deficient NSCLC are consistent with those of common NSCLC (such as cough and chest pain), but due to tumor growth characteristics (some reports suggest strong invasiveness), some symptoms may be more pronounced or appear earlier. Any other prominent characteristic co-morbidities had not been found in the 15 patients studied. Additionally, surgery combined with immunotherapy or targeted therapy was the primary treatment (Table I). Microscopic examination showed map-shaped necrosis and cancer cells arranged in a patchy, island-shaped or acinar pattern. Tumor cells appeared as small epithelioid cells; some had abundant cytoplasm, eosinophilic or translucent nuclei, vacuolated nuclei, distinct nucleoli, and visible mitotic figures (Fig. 1A-C). Strong tumor proliferation was observed in ~50% of the regions with Ki67-positive expression (Fig. 1D). The staining results showed that the normal bronchial and alveolar epithelial cells were positive for the epithelial markers CK7, NapsinA and TTF-1 (Fig. 1E-G), while the tumor tissues did not express these proteins. Both the squamous epithelial markers P63 and P40 showed positive basement membrane and negative tumor tissue (Fig. 1H and I). The two neuroendocrine markers, Syn and CD56, showed a pattern of weakly positive and negative expression, respectively, in the tumor tissues (Fig. 1J and K). Tumor cells showed a high level of the mesenchymal marker Vimentin (Fig. 1L). Overall, 15 cases of SMARCA4-deficient non-small cell lung cancer were characterized by large tumor size and high invasiveness, with prominent geographic necrosis and epithelioid cell morphology histologically, and an immunophenotype featuring negative epithelial markers and high expression of the mesenchymal marker Vimentin.

Comparison of the immunohistochemical expression characteristics of EGFR, anaplastic lymphoma kinase (ALK) (D5H3), P53 and programmed cell death 1 ligand 1 (PD-L1) in 15 patients with SMARCA4-deficient tumors showed that the EGFR protein could be either expressed positively (3/15) (Fig. 2A and I) or negatively (12/15) (Fig. 2B and I) in SMARCA4-deficient tumors. However, cases of ALK protein expression (D5H3) positivity in SMARCA4-deficient tumors were lacking (Fig. 2E and I). Additionally, the PD-L1 protein (4/15) had a low likelihood of positivity, as it was only expressed in 4/15 patients (Fig. 2C and I), whereas it was negatively expressed in 11/15 patients (Fig. 2D and I). Furthermore, compared with the wild-type protein expression of P53, the P53 protein showed a higher likelihood of exhibiting complete negative (4/14) or complete positive mutations (11/15), because the P53 protein has three expression patterns, partial positivity indicates the wild type, while complete negativity or complete positivity indicates the mutant type (Fig. 2F-H and J). In SMARCA4-deficient NSCLC, EGFR, ALK, and PD-L1 positivity rates were low, while P53 showed a high frequency of aberrant expression, suggesting limited benefit from conventional targeted therapy and a potential subtype-specific molecular profile.

The spectrum of gene mutations in 15 SMARCA4-deficient patients with lung cancer were summarized in Table II. Patients 4 and 12 shared the same mutation type, c.2729C>T (p.T910M). The genetic mutation types of patient 7 and patient 8 are the same; both have experienced a copy number deletion. Furthermore, patient 13 was unique: Three mutations were detected in SMARCA4 of this patient, namely c.3634_3636del (p. E1212del), c.3574C>T (p.R1192C), and c.3733G>A (p.A1245T) (Fig. 3A). SMARCA4 encodes the BRG1 protein, while SMARCB1 encodes the INI1 protein. In clinical diagnosis, immunohistochemical pairing is frequently used to screen for BRG1 and INI1 protein expression to detect SMARCA4-deficient lung cancer. The BRG1 protein encoded by the SMARCA4 was not expressed (the BRG1-stained positive areas in Fig. 3B are normal alveolar epithelium, not tumor regions), whereas the INI1 protein expression was partially positive (6/15) (Fig. 3B). Additionally, IHC was conducted on the tumor tissues of the 15 patients with SMARCA4-deficient lung cancer. Some patients showed negative INI1 expression, as shown in the left part of Fig. 3C, while some showed positive expression, as shown in the right part of Fig. 3C. Only patients 1, 5, 6, 13, 14 and 15 showed positive INI1 protein expression, whereas the remaining nine patients showed negative expression. Table II shows the representative expression of the INI1 protein, along with the corresponding mutation profile. Among the 15 SMARCA4-deficient patients with NSCLC, diverse mutations including missense, copy number loss, and multiple concurrent variants were detected, with loss of BRG1 expression and heterogeneous INI1 expression, supporting the value of IHC pairing of BRG1 and INI1 in the clinical identification of this molecular subtype.

The gene mutation profiles of 15 patients were analyzed. NQO1 and TP53 mutations were detected in nine patients, XRCC1 mutation was detected in seven, DPYD mutation was detected in six, TYMS mutation was detected in five, GSTP1, MTHFR, LRP1B and UGT1A1 mutations were detected in four, and GSTT1, GSTM1, ID3, ERCC1, ERCC2, ARID1A, RAD54L and TERT mutations were detected in three (Fig. 4A). All mutated genes are shown in Table SI. The enrichment analysis conducted on the identified gene set using the Gene Expression Profiling Interactive Analysis (GEPIA)2 database revealed the presence of mutated genes such as ‘EGFR tyrosine kinase inhibitor resistance’, ‘non-small cell lung cancer’, ‘ErbB signaling pathway’, ‘protein kinase regulator activity’, ‘extrinsic apoptotic signaling pathway’ and ‘regulation of ERK1 and ERK2 cascade’. The same pathway demonstrated significant enrichment (Fig. 4B). Additionally, the 15 patients with lung cancer and SMARCA4 deficiency demonstrated partial mutations of specific genes. For example, NQO1 c.559C>T (p.P187S) and XRCC1 c.1196A>G (p.Q399R) were detected 7 times, DPYD c.1627A>G (p.I543V) was detected 5 times, and TYMS c.*450_*455delAAGTTA, GSTP1 c.313A>G (p.I105V) and MTHFR c.665C>T (p.A222V) were detected 4 times. These three gene loci exhibited four different mutations (Table III). The high-frequency mutation genes and sites may control the incidence and progression of SMARCA4-deficient lung cancer. These results identified a distinct mutation spectrum involving DNA repair, metabolic and cancer-related genes in SMARCA4-deficient NSCLC, with several recurrent high-frequency mutation sites that may contribute to tumorigenesis and progression via pathways closely associated with NSCLC and therapy resistance.