The roles of risk model based on the 3-XRCC genes in lung adenocarcinoma progression.

BACKGROUND: The abnormal expression of deoxyribonucleic acid (DNA) repair genes might be the cause of tumor development and resistance of malignant cells to chemotherapeutic drugs. A risk model based on the X-ray repair of cross-complementary (XRCC) genes was constructed to improve the diagnosis and treatment of lung adenocarcinoma (LUAD) patients.
METHODS: The expression levels, diagnostic values, and prognostic values of XRCC genes were identified, and the roles and regulatory mechanisms of the risk model based on the XRCC4/5/6 in LUAD progression was explored via The Cancer Genome Atlas (TCGA) and Oncomine databases.
RESULTS: XRCC1/2/3/4/5/6, XRCC7 (PRKDC), and XRCC9 (FANCG) were overexpressed, and had diagnostic value for LUAD. The XRCC genes were involved in DNA repair, and participated in the regulation of non-homologous end-joining, homologous recombination, etc. The overall survival (OS), tumor (T) stage, and survival status of patients were significantly different between the Cluster1 and Cluster2 groups. XRCC4/5/6 were independent risk factors affecting the prognosis of LUAD patients. The risk score was related to the prognosis, sex, clinical stage, T, lymph node (N), and metastasis (M) stage, as well as the survival status of LUAD patients. The clinical stage and risk score were independent risk factors for poor prognosis in LUAD patients. The risk model was involved in RNA degradation, cell cycle, basal transcription factors, DNA replication etc. The risk scores were significantly correlated with the expression levels of TGFBR1, CD160, TNFSF4, TNFRSF14, IL6R, CXCL16, TNFRSF25, TAPBP, CCL16, and CCL14.
CONCLUSIONS: The risk model based on the XRCC4/5/6 genes could predict the progression of LUAD patients. 2021 Translational Cancer Research. All rights reserved.

Introduction

The deoxyribonucleic acid (DNA) repair system plays a vital role in protecting the human genome from carcinogens. The abnormal expression of DNA repair genes might be the cause of tumor development and resistance of malignant cells to chemotherapeutic drugs (1-6). For example, hydroxycamptothecin (HCPT) could increase the expression of the DNA repair gene, XPF, in bladder cancer and promote apoptosis in T24 and 5637 cells. The increased expression of XPF could reduce the sensitivity of bladder cancer cells, while interfering with the expression of XPF could reduce the resistance of bladder cancer cells to chemotherapy (5). Likewise, interfering with the expression of BRCA1 interacting protein C-terminal helicase 1 (BRIP1), which regulates DNA repair and cell proliferation could induce cell cycle arrest and reduce the proliferation of breast cancer (BC) cells, and promote the invasion of BC cells (6). These examples highlight the important role of the DNA repair system in cancer progression.

The X-ray repair of cross-complementary (XRCC) genes are common components of the DNA repair system and are related to cancer progression. For example, XRCC1 is essential for DNA base excision repair, single strand break repair, and nucleotide excision repair. In ovarian cancer, XRCC1 is positive in 48% of tumor patients, which is related to advanced stage, platinum resistance, disease progression, and so on. The expression level of XRCC1 is an independent risk factor for cancer specificity and progression-free survival. Compared with XRCC1-positive cells, XRCC1-negative cells are sensitive to cisplatin, which is related to DNA double-strand breaks and cell cycle arrest of G2/M (7). XRCC2 overexpression has been found in rectal cancer tissues without preoperative radiotherapy (PRT). Compared with XRCC2-positive patients treated with PRT, XRCC2-negative patients with locally advanced rectal cancer (LARC) have improved overall survival (OS). The level of XRCC2 expression is related to the increase of radiation resistance of LARC, while cancer cells without XRCC2 expression are more sensitive to radiation in vitro, which is related to the arrest and apoptosis of cells in the G2/M phase. When the expression of XRCC2 is interfered with, the repair ability of DNA double strand breaks caused is impaired via radiation (8).

The Cancer Genome Atlas (TCGA) database aims to apply high-throughput genome analysis technology to improve people’s ability to prevent, diagnose, and treat cancer. It has multiple cancer types and groups of data, including gene expression data, microRNA (miRNA) expression data, copy number variation, DNA methylation, and so on (9,10). However, the role of XRCC genes in the progression of lung adenocarcinoma (LUAD) has not been fully elucidated. In recent years, risk models have also been commonly used to assess the prognosis of cancer patients (11,12). In this study, the expression levels, diagnostic value, and prognostic value of XRCC genes in LUAD were evaluated using the Oncomine and TCGA databases, and a risk model was constructed to evaluate the clinical predictive value for the progression of LUAD patients. The following article was presented in accordance with the TRIPOD reporting checklist (available at https://dx.doi.org/10.21037/tcr-21-1431).

Methods

The study was conducted in accordance with the Declaration of Helsinki (as revised in 2013).

Oncomine database

The Oncomine 3.0 (https://www.oncomine) database is used for the study of tumor-related genes, with a wide range of data sources and high reliability (13). The expression of XRCC genes in pan-cancer tissues was analyzed in the Oncomine database. The XRCC genes included the following: XRCC1, XRCC2, XRCC3, XRCC4, XRCC5, XRCC6, FANCG, and PRKDC. The screening criteria were as follows: (I) genes: XRCC1/2/3/4/5/6, FANCG, and PRKDC; (II) analysis type: cancer versus normal analysis; (III) data type: messenger RNA (mRNA); (IV) P<0.05; and (V) fold change: ALL.

Visualization analysis of TCGA data

The gene expression data of HTSeq-FPKM tissue, including 59 cases of lung tissues and 535 cases of LUAD tissues, and the clinical data of 522 cancer patients were downloaded from the official TCGA (https://portal.gdc.cancer.gov/projects/TCGA-LUAD (HTSeq-FPKM) website. Among them, 57 lung tissues and 57 LUAD tissues were derived from the same LUAD patients. The expressions of XRCC1/2/3/4/5/6, FANCG, and PRKDC were identified in lung and LUAD tissues, and the correlation between XRCC genes was analyzed. Principal component analysis (PCA), gene set enrichment analysis (GSEA), and clinical correlation analysis were performed in the 535 cases of LUAD issues.

Consensus clustering and survival analysis

According to the expression levels of XRCC genes, the 535 cases of LUAD tissues in TCGA database were divided into two groups using the “Consensus-ClusterPlus” in R, and PCA was performed (14,15). Kaplan-Meier survival analysis and correlation analysis were performed to evaluate the OS and clinicopathological characteristics (age, sex, clinical stage, T stage, N stage, M stage, and survival status) in both groups.

Construction of the risk model in LUAD

Univariate Cox regression analysis was used to filter the prognostic factors in patients with LUAD. The independent risk factors for poor prognosis of LUAD patients were screened by multivariate Cox regression analysis and the Akaike information criterion (AIC) (16). LUAD patients were divided into high- and low-risk groups according to the gene expression levels. Kaplan-Meier survival analysis evaluated the risk of death in two groups of LUAD patients. The relationship between risk and clinicopathological features (including age, sex, clinical stage, T stage, N stage, M stage) was assessed in patients with LUAD via correlation analysis.

The value of risk model in the prognosis of LUAD

Univariate and multivariate Cox regression analyses were used to assess the effects of the risk model, age, sex, clinical stage, T stage, N stage, and M stage on the prognosis of LUAD patients, and to evaluate the role of the risk model in the prognosis of LUAD patients (17).

Biological processes and signaling mechanisms

The XRCC genes were entered into the String (https://string-db.org) database to conduct Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG), and protein-protein interaction (PPI) analyses. GSEA was used to explore the biological functions and regulatory mechanisms that the influencing factors might be involved in (18-20). The LUAD tissue gene expression data from TCGA database were divided into high- and low-risk groups according to the median value of the risk model score to explore the effects of two groups on each gene. GO [biological process (BP)] and KEGG analyses were carried out using the GSEA software. The screening criteria was as follows: nominal (NOM) P<0.05.

Correlation analysis of LUAD immune cell markers

The relationship between risk model factors and LUAD immune infiltrating cell markers were analyzed in 535 cases of LUAD via correlation analysis. One-to-one correspondence between the risk score and LUAD samples was conducted. The expression level of LUAD immune infiltrating cell markers were explored in the high- and low-risk groups.

Statistical analysis

Cox regression and Kaplan-Meier survival analysis were used to analyze the risk factors associated with OS in patients with LUAD. The univariate and multivariate Cox regression analyses and AIC were used to screen the prognostic factors in patients with LUAD. Correlation analysis was used to analyze the relationship between the risk factors and LUAD immune cell infiltration markers. GraphPrism 5.0 and R (Version 3.6.1) ggplot package were plotted. P<0.05 was regarded as statistically significant.

Results

The expression level of XRCC genes was significantly increased in LUAD tissues

In the Oncomine database, XRCC1, XRCC2, XRCC3, XRCC4, XRCC5, XRCC6, FANCG, and PRKDC were abnormally expressed in pan-cancer tissues, and the expression levels of XRCC genes were mainly increased in pan-cancer tissues (Figure S1). Based on our screening criteria, most of the datasets showed that XRCC genes were predominantly higher in lung cancer tissues. Specifically, the datasets related to the expression of XRCC1, XRCC6, XRCC2, XRCC3, FANCG, XRCC4, XRCC5, and PRKDC were 4 vs. 1, 9 vs. 5, 5 vs. 4, 13 vs. 0, 14 vs. 3, 8 vs. 0, 13 vs. 4, and 18 vs. 2, respectively.

In addition, the expression levels of XRCC1, XRCC6, XRCC3, XRCC2, FANCG, XRCC4, XRCC5, and PRKDC increased in LUAD tissues in the TCGA database, and the difference was statistically significant (). In addition, we sorted the data obtained from the TCGA database and matched the tissues one-to-one to show that the expression levels of XRCC1, XRCC6, XRCC3, XRCC2, FANCG, XRCC4, XRCC5, and PRKDC increased in LUAD tissues ().

Figure 1

High expression of XRCC genes in TCGA LUAD tissues. (A) Heat map showing the expression of XRCC genes in LUAD tissues; (B) scatter plot showing the expression of XRCC genes in LUAD tissues. XRCC, X-ray repair of cross-complementary; TCGA, The Cancer Genome Atlas; LUAD, lung adenocarcinoma. ***, P<0.001.

Figure 2

High expression of XRCC genes in matched LUAD tissues from TCGA database. (A) XRCC1; (B) XRCC6; (C) XRCC2; (D) XRCC3; (E) FANCG; (F) XRCC4; (G) XRCC5; (H) PRKDC. *, P<0.05; ***, P<0.001. XRCC, X-ray repair of cross-complementary; TCGA, The Cancer Genome Atlas; LUAD, lung adenocarcinoma.

Diagnostic value of XRCC genes in LUAD

The diagnostic value of XRCC genes in LUAD was evaluated via receiver operator characteristic (ROC) analysis. The results showed that the area under the curve (AUC) of XRCC1, XRCC6, XRCC2, XRCC3, FANCG, XRCC4, XRCC5, and PRKDC were all between 0.5 and 1, which was statistically significant (). Specifically, the AUCs of XRCC1, XRCC6, XRCC2, XRCC3, FANCG, XRCC4, XRCC5, and PRKDC were 0.6628 (), 0.7785 (), 0.9841 (), 0.7913 (), 0.9943 (), 0.8425 (), 0.8743 (), and 0.8732 (), respectively.

Figure 3

Diagnostic value of XRCC genes in LUAD. (A) XRCC1; (B) XRCC6; (C) XRCC2; (D) XRCC3; (E) FANCG; (F) XRCC4; (G) XRCC5; (H) PRKDC. XRCC, X-ray repair of cross-complementary; LUAD, lung adenocarcinoma; ROC, receiver operator characteristic; AUC, the area under the curve.

The biological functions of XRCC genes

In LUAD tissues, we observed significant correlations between the expression levels of the following genes: (I) XRCC1 and XRCC3, and FANCG and XRCC4; (II) XRCC3 and FANCG and XRCC2; (III) FANCG and XRCC5, and XRCC2 and PRKDC; (IV) XRCC5 and XRCC6, and XRCC2 and PRKDC; and (V) XRCC2 and PRKDC (Figure S2A). Using the String database, we found that XRCC genes were involved in biological processes such as DNA repair, DNA recombination, response to radiation, response to X-ray, mitotic recombination, and so on, and were also involved in the regulation of non-homologous end-joining and homologous recombination signaling mechanisms ( and Table S1). In the PPI network, there was a strong functional relationship among the XRCC genes (Figure S2B).

Table 1

The XRCC genes were involved in biological processes

GO: BP	Description	P
GO:0006302	Double-strand break repair	2.94E-11
GO:0006281	DNA repair	3.49E-11
GO:0006310	DNA recombination	3.49E-11
GO:0010212	Response to ionizing radiation	7.62E-10
GO:0009314	Response to radiation	1.68E-09
GO:0006303	Double-strand break repair via nonhomologous end joining	2.33E-09
GO:0009628	Response to abiotic stimulus	4.44E-09
GO:0000723	Telomere maintenance	1.31E-08
GO:0010165	Response to X-ray	3.56E-08
GO:0010332	Response to gamma radiation	2.12E-07
GO:0075713	Establishment of integrated proviral latency	2.84E-07
GO:0006266	DNA ligation	1.33E-06
GO:0006312	Mitotic recombination	2.32E-06
GO:0071475	Cellular hyperosmotic salinity response	3.91E-05
GO:0032481	Positive regulation of type I interferon production	5.69E-05
GO:0000707	Meiotic DNA recombinase assembly	0.00012
GO:0000724	Double-strand break repair via homologous recombination	0.00012
GO:0051351	Positive regulation of ligase activity	0.00012
GO:0042148	Strand invasion	0.00013
GO:0000722	Telomere maintenance via recombination	0.00019
GO:0051103	DNA ligation involved in DNA repair	0.00019
GO:0071481	Cellular response to X-ray	0.00019
GO:0048660	regulation of smooth muscle cell proliferation	0.00021
GO:0006996	Organelle organization	0.00024
GO:0002218	Activation of innate immune response	0.00057
GO:0071480	Cellular response to gamma radiation	0.00069
GO:0007420	Brain development	0.00087
GO:0032205	Negative regulation of telomere maintenance	0.0012
GO:0007131	Reciprocal meiotic recombination	0.0017
GO:0036297	Interstrand cross-link repair	0.0017
GO:0032508	DNA duplex unwinding	0.002
GO:0033044	Regulation of chromosome organization	0.002
GO:0001756	Somitogenesis	0.0027
GO:0043902	Positive regulation of multi-organism process	0.0035
GO:0002244	Hematopoietic progenitor cell differentiation	0.0037
GO:0007399	Nervous system development	0.0054
GO:0080134	Regulation of response to stress	0.0074
GO:0022414	Reproductive process	0.0083
GO:0043085	Positive regulation of catalytic activity	0.0086
GO:0051704	Multi-organism process	0.0086

XRCC, X-ray repair of cross-complementary; GO, Gene Ontology; BP, biological process.

Table 2

The XRCC genes were involved in molecular function

GO: MF	Description	P
GO:0140097	Catalytic activity, acting on DNA	2.35E-07
GO:0003684	Damaged DNA binding	3.37E-07
GO:0008094	DNA-dependent ATPase activity	3.37E-07
GO:0003677	DNA binding	6.46E-05
GO:0000150	Recombinase activity	0.0001
GO:0003690	Double-stranded DNA binding	0.0001
GO:0008022	Protein C-terminus binding	0.00044
GO:0005524	ATP binding	0.00074
GO:0042162	Telomeric DNA binding	0.00074
GO:0003678	DNA helicase activity	0.00078
GO:0008144	Drug binding	0.00088
GO:0003697	Single-stranded DNA binding	0.0019
GO:0016787	Hydrolase activity	0.003

XRCC, X-ray repair of cross-complementary; GO, Gene Ontology; MF, molecular function.

Table 3

The XRCC genes were involved in cellular component

GO: CC	Description	P
GO:1990391	DNA repair complex	6.40E-11
GO:0070419	Nonhomologous end joining complex	2.57E-10
GO:0000784	Nuclear chromosome, telomeric region	8.13E-08
GO:0005654	Nucleoplasm	1.10E-05
GO:0043564	Ku70:Ku80 complex	1.10E-05
GO:0005958	DNA-dependent protein kinase-DNA ligase 4 complex	1.97E-05
GO:0033063	Rad51B-Rad51C-Rad51D-XRCC2 complex	2.58E-05
GO:0005730	Nucleolus	6.28E-05
GO:0005694	Chromosome	6.31E-05
GO:0032991	Protein-containing complex	6.31E-05
GO:0000783	Nuclear telomere cap complex	6.43E-05
GO:0032993	Protein-DNA complex	0.00015
GO:0043232	Intracellular non-membrane-bounded organelle	0.00028
GO:0005657	Replication fork	0.00056

XRCC, X-ray repair of cross-complementary; GO, Gene Ontology; CC, cellular component.

Consensus clustering of XRCC genes identified two clusters of LUAD with different clinical outcomes

With the evolution of clustering from k=2 to 9, k=2 might be the best choice with the least interference in our clustering (). Therefore, we used k=2 for consensus clustering analysis, and defined it as Cluster1 and Cluster2 groups. PCA was performed in the 535 cases of LUAD from the TCGA database, and the results showed that there was a significant difference between the Cluster1 and Cluster2 groups (). Survival analysis showed that the OS of LUAD patients in Cluster1 was better than that of LUAD patients in Cluster2 (). Correlation analysis showed that there was a significant correlation between T stage and survival status of patients in the Cluster1 and Cluster2 groups ().

Figure 4

The overall survival of LUAD patients in the Cluster1 and Cluster2 subgroups. *, P<0.05; **, P<0.01. LUAD, lung adenocarcinoma.

The prognostic value of XRCC genes in patients with LUAD

The value of XRCC genes in the prognosis of LUAD was explored via univariate Cox regression analysis. We found that XRCC4, XRCC5, XRCC6, and PRKDC might be the risk factors affecting the prognosis of LUAD patients (). On this basis, the risk model was constructed under the conditions of multivariate Cox regression analysis and AIC optimization. The results showed that XRCC4, XRCC5, and XRCC6 were independent risk factors affecting the prognosis of patients with LUAD. Kaplan-Meier survival analysis showed that the prognosis of LUAD patients in the high-risk group was worse (). Correlation analysis showed that high- and low-risk were significantly correlated with the gender, clinical stage, T stage, N stage, M stage, and survival status of LUAD patients (). The univariate and multivariate Cox regression analyses showed that the clinical stage and risk score were independent risk factors for poor prognosis in patients with LUAD ().

Figure 5

Prognostic value of XRCC genes in patients with LUAD. (A) Univariate Cox regression analysis; (B,C) risk score was correlated to the clinicopathological features and OS of LUAD patients based on XRCC4, XRCC5, and XRCC6. *, P<0.05; **, P<0.01; ***, P<0.001. XRCC, X-ray repair of cross-complementary; LUAD, lung adenocarcinoma; OS, overall survival.

Figure 6

Univariate and multivariate Cox regression analysis revealed that the clinical stage and risk score were independent risk factors for poor prognosis in patients with LUAD. (A) Univariate Cox regression analysis; (B) multivariate Cox regression analysis. LUAD, lung adenocarcinoma.

The biological functions and signaling pathways involved in the risk model

According to the median risk score, we divided the gene expression data of the 535 cases of LUAD from the TCGA into high- and low-risk groups to explore the influence of genes in two groups. The GSEA results showed that increased risk might involve biological processes such as regulation of DNA replication, mitotic metaphase plate congression, cell cycle DNA replication (Figure S3), as well as signaling systems such as RNA degradation, cell cycle, oocyte meiosis, basal transcription factors, and DNA replication (Figure S4 and ).

Table 4

The high-risk group was involved in signaling pathways via the GSEA

Name	Size	NES	NOM P value
RNA_degradation	57	2.1544359	0
Cell_cycle	124	2.139343	0
Nucleotide_excision_repair	44	2.1303906	0.001964637
OOCYTE_MEIOSIS	112	2.0731578	0.001996008
Mismatch_repair	23	2.0708838	0
Basal_transcription_factors	35	2.0075533	0
DNA_replication	36	1.98874	0
Proteasome	44	1.9682256	0.001972387
Ubiquitin_mediated_proteolysis	133	1.9554849	0
Protein_export	23	1.9503225	0
pathogenic_escherichia_coli_infection	55	1.9071776	0.005825243
Citrate_cycle_tca_cycle	30	1.8967364	0.004056795
Spliceosome	126	1.890303	0.004032258
Pyrimidine_metabolism	98	1.8573432	0.00204499
Purine_metabolism	157	1.8310792	0.002159827
Cysteine_and_methionine_metabolism	34	1.7912648	0.003861004
P53_signaling_pathway	68	1.733961	0.007843138
One_carbon_pool_by_folate	17	1.7224773	0.016129032
RNA_polymerase	29	1.7176231	0.018367346
Homologous_recombination	28	1.6963832	0.034274194
Biosynthesis_of_unsaturated_fatty_acids	22	1.6282122	0.018072288
Riboflavin_metabolism	15	1.620054	0.036538463
Aminoacyl_trna_biosynthesis	22	1.5583364	0.049701788
Progesterone_mediated_oocyte_maturation	85	1.4684261	0.07370518
Amyotrophic_lateral_sclerosis_als	52	1.4148762	0.047244094
Glyoxylate_and_dicarboxylate_metabolism	16	1.3979565	0.115686275
Glycolysis_gluconeogenesis	62	1.388122	0.082
Huntingtons_disease	177	1.3863257	0.14256199
Terpenoid_backbone_biosynthesis	15	1.378746	0.14141414
Pentose_phosphate_pathway	27	1.3762755	0.1097561
Thyroid_cancer	29	1.3680532	0.091617934
Pancreatic_cancer	70	1.3660588	0.11133201
Adherens_junction	73	1.3573757	0.11576846
Base_excision_repair	33	1.34626	0.17886178
Alzheimers_disease	163	1.3419145	0.15352698
Tgf_beta_signaling_pathway	85	1.3412957	0.11025145
N_glycan_biosynthesis	46	1.3321104	0.1482966
Colorectal_cancer	62	1.3087744	0.14705883
Propanoate_metabolism	31	1.2553445	0.2300195
Glycosylphosphatidylinositol_gpi_anchor_biosynthesis	25	1.2523328	0.22113504
WNT_signaling_pathway	150	1.2449616	0.14
Chronic_myeloid_leukemia	73	1.227265	0.21370968
Small_cell_lung_cancer	84	1.197284	0.23203285
Epithelial_cell_signaling_in_helicobacter_pylori_infection	68	1.1845227	0.20315582
Prostate_cancer	89	1.1837994	0.23883495
Pathways_in_cancer	325	1.1785803	0.20081967
Renal_cell_carcinoma	70	1.164982	0.24055666
Long_term_potentiation	70	1.1516585	0.23943663
Cytosolic_dna_sensing_pathway	54	1.1250238	0.31769723
Ribosome	87	1.1220671	0.43037975
Nicotinate_and_nicotinamide_metabolism	24	1.1175917	0.28849903
Regulation_of_actin_cytoskeleton	212	1.1145742	0.27991885
Parkinsons_disease	125	1.089907	0.39793813
Vasopressin_regulated_water_reabsorption	44	1.0887667	0.33466136
Glutathione_metabolism	47	1.0859108	0.36452243
Lysine_degradation	44	1.0843796	0.34068137
Snare_interactions_in_vesicular_transport	38	1.0831342	0.33714285
Valine_leucine_and_isoleucine_degradation	43	1.0764002	0.38446215
Pyruvate_metabolism	40	1.0609999	0.3767821
Pentose_and_glucuronate_interconversions	28	1.0568165	0.41497976
Selenoamino_acid_metabolism	25	1.0436656	0.38202247
Rig_i_like_receptor_signaling_pathway	70	1.0409458	0.3732535
Regulation_of_autophagy	35	1.0247588	0.4329502
Amino_sugar_and_nucleotide_sugar_metabolism	43	1.022107	0.4027778
Peroxisome	78	1.0060785	0.4322709
Melanoma	71	1.003775	0.44466403
Nod_like_receptor_signaling_pathway	62	1.0017914	0.46601942
Alanine_aspartate_and_glutamate_metabolism	30	0.98454267	0.47233203
Endocytosis	181	0.9828535	0.45691383
Fructose_and_mannose_metabolism	33	0.9741695	0.46626985
Glioma	65	0.9564347	0.49203187

GSEA, gene set enrichment analysis; NES, normalized enrichment score; NOM, nominal.

The risk model based on XRCC4, XRCC5, and XRCC6 was related to the LUAD immunity

The correlation analysis showed that XRCC4, XRCC5, XRCC6, and their risk model were significantly correlated with the levels of immune factors (). Specifically, the expression level of XRCC4 was positively correlated with the expression levels of TNFSF4, CD80, PDCD1LG2, CXCL8, etc. ( and Table S2), and negatively correlated with the expression levels of CXCL17, IL6R, TAPBP, CXCL16, etc. ( and Table S2). The expression level of XRCC5 was positively correlated with the expression levels of PVR, TGFBR1, CXCL8, XCL1, etc. ( and Table S2), and negatively correlated with the expression levels of TNFRSF14, HLA-DMA, TMEM173, HLA-DPB1, etc. ( and Table S2). The expression level of XRCC6 was positively correlated with the expression levels of CD276, TNFSF13, CXCL16, TNFSF9, etc. ( and Table S2), and negatively correlated with the expression levels of CD160, KLRK1, BTLA, CCL16, etc. ( and Table S2).

Figure 7

XRCC4, XRCC5, and XRCC6 of XRCC genes were correlated to immune markers in LUAD. LUAD, lung adenocarcinoma.

Figure 8

Risk score was correlated to immune markers based on XRCC4, XRCC5, and XRCC6 in LUAD. **, P<0.01; ***, P<0.001. LUAD, lung adenocarcinoma.

The immune factors associated with the intersection of XRCC4, XRCC5, and XRCC6 in both high- and low-risk groups were validated (). Specifically, the expression levels of TGFBR1, CD160, TNFSF4, TNFRSF14, IL6R, CXCL16, TNFRSF25, TAPBP, CCL16, and CCL14 were significantly associated with high- and low-risk scores ().

Discussion

Persistent failure to repair DNA damage might lead to cell cycle arrest, apoptosis, and genomic instability, which leads to the development of many diseases (21). The XRCC genes are important components of the DNA damage repair mechanism and play important biological roles in cancer progression (21-24). At present, numerous studies have confirmed that polymorphisms of DNA damage repair genes such as XRCC1, XRCC3, and XRCC4 were associated with the survival of patients with lung cancer (25-27). However, the role of XRCC genes in the progression of LUAD has not been fully elucidated. In this study, we observed that the expression levels of XRCC1, XRCC6, XRCC3, XRCC2, FANCG, XRCC3, XRCC4, and PRKDC increased in unpaired and paired LUAD tissues. ROC analysis showed that the AUCs of XRCC1, XRCC6, XRCC2, XRCC3, FANCG, XRCC4, XRCC5, and PRKDC were all between 0.5 and 1. Cox regression analysis demonstrated that XRCC4, XRCC5 and XRCC6 were independent risk factors affecting the prognosis of LUAD patients. Kaplan-Meier survival analysis showed that the prognosis of LUAD patients in the high-risk group was worse, and a high-risk score was significantly correlated with the gender, clinical stage, T stage, N stage, M stage, and survival status of LUAD patients. These results indicated that XRCC4, XRCC5, and XRCC6 play an important role in the progression of LUAD and are expected to become biomarkers for the diagnosis and prognosis of LUAD. Muylaert et al. reported that DNA ligase IV/XRCC4 plays a crucial role in the herpesvirus replication cycle. Reducing DNA ligase IV/XRCC4 could inhibit herpes simplex virus type I DNA replication (28). The expression of Ku86 (XRCC5) is significantly increased in serous ovarian cancer (SOC), and down-regulation of Ku86 expression could promote increased γ-H2AX expression, resulting in the inhibition of cell proliferation, cell cycle block in G2 phase, and the increase of G2/G1. X-ray irradiation could also reduce the expression of Ku86 to promote the above biological effects, and increase the expression of γ-H2AX (29). XRCC6 is overexpressed in human osteosarcoma tissues and cells. The high expression of XRCC6 is related to the clinical stage and tumor size of patients with osteosarcoma. The decreased expression of XRCC6 could inhibit the proliferation of osteosarcoma cells through G2/M phase arrest, which might regulate the growth of osteosarcoma through β-catenin/Wnt signaling pathway (30). The XRCC genes were related factors of DNA damage repair, and the risk model based on XRCC4, XRCC5, and XRCC6 could involve mitotic metaphase plate congression, DNA replication, RNA degradation, the cell cycle, oocyte meiosis, basal transcription factors, DNA replication, and so on. This indicates that XRCC4, XRCC5, and XRCC6 are related to cell cycle, DNA damage and DNA replication; however, further confirmation by basic research is needed.

It is well known that the progression of cancer is related to factors in the immune microenvironment. For example, C-X-C motif chemokine ligand 8 (CXCL8) is associated with a high tumor burden in LUAD and is negatively correlated with DACH1 expression. High DACH1 expression and low CXCL8 expression has been found to prolong the time of death and tumor recurrence of patients. DACH1 can inhibit the activity of the CXCL8 promoter and reduce the level of CXCL8 expression through transcription at the sites of activating protein-1 (AP-1) and nuclear factor-κB (NF-κB) (31). We found that the expression level of XRCC4 was correlated with the expression levels of TNFSF4, CD80, PDCD1LG2, CXCL8, CXCL17, IL6R, TAPBP, CXCL16, and so on; the expression level of XRCC5 was correlated with the expression levels of PVR, TGFBR1, CXCL8, XCL1, TNFRSF14, HLA-DMA, TMEM173, HLA-DPB1, and so on; and the expression level of XRCC6 was correlated with the expression level of CD276, TNFSF13, CXCL16, TNFSF9, CD160, KLRK1, BTLA, CCL16, and so on. In the high- and low-risk groups, it was found that the expression levels of TGFBR1, CD160, TNFSF4, TNFRSF14, IL6R, CXCL16, TNFRSF25, TAPBP, CCL16, and CCL14 were significantly correlated with a high risk. Meanwhile, Jiang et al. reported that TGFBR1, TNFSF4, and IL6R were associated with lung cancer progression (32-35), which provided some evidence for our research.

The risk model based on TCGA data has good prognostic value. However, clinical tissue samples should be collected to verify the expression of XRCC4/5/6 in LUAD tissues via the RT-PCR and western-blot, and the value of XRCC4/5/6 in the diagnosis and prognosis of LUAD was analyzed. In addition, we need to build cell models in the future to explore the cell growth, migration and singaling mechanisms of XRCC4/5/6 in the progression of LUAD. Generally speaking, the XRCC genes played an important role in the diagnosis and prognosis of LUAD. XRCC4, XRCC5, and XRCC6 were independent risk factors affecting the prognosis of LUAD patients. There were significant differences in prognosis, sex, clinical stage, T stage, N stage, M stage, and survival status of LUAD patients in the high- and low-risk groups. The clinical stage and risk score were independent risk factors for poor prognosis in patients with LUAD. The risk model was involved in mitotic metaphase plate congression, RNA degradation, cell cycle, oocyte meiosis, basal transcription factors, DNA replication, and other processes. XRCC4, XRCC5, XRCC6, and the risk scores were significantly correlated with the expression levels of immune factors of TGFBR1, CD160, TNFSF4, TNFRSF14, IL6R, CXCL16, TNFRSF25, TAPBP, CCL16, and CCL14.

Conclusions

In this study, the risk model based on XRCC4, XRCC5, and XRCC6 could predict the progression of LUAD patients.